Image transmitter

Abstract

A transmitter is capable of compressing an incoming image with a relatively small delay time and transmit the compressed image data to a receiver. Image data contains at least i pixel values of pixels arranged in line along a single direction, each pixel value being expressed in n bits. A transmitter for compressing such image data and transmitting the image data to a receiver via a transmission path includes a blocking section, a data compression section, and a data sending section. The blocking section takes every p pixel values among the i pixel values in the image data to form a data block, and sequentially outputs a plurality of the data blocks each including the p pixel values. The data compression section reduces an amount of data from each data block outputted from the blocking section and thereby outputs a compressed block. The data sending section sends the compressed block outputted from the data compression section onto the transmission path.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to an image transmitter, and more specifically to an image transmitter for compressing incoming image data and transmitting the compressed image data to a receiver via a transmission path.

2. Description of the Background Art

Examples of conventional image compression schemes are the MPEG (Motion Picture Experts Group) scheme and the DVC (Digital Video Cassette) scheme. According to these image compression schemes, incoming image data is subjected to DCT (Discrete Cosine Transform) and variable-length coding on a macro block-by-macro block basis, whereby a high compression rate for the image data may be realized. An implementation example of such an image compression scheme is a moving picture encoder which is disclosed in Japanese Patent Laid-Open Publication No. 7-280911.

However, after the aforementioned moving picture encoder receives one line of pixels arranged along a horizontal direction within an image to be processed, the moving picture encoder may also receive a next line of pixels. As a result, before all of the pixels which compose one macro block are received, the moving picture encoder may receive a number of pixels which are unrelated to that macro block. The receipt of such unnecessary pixels causes a delay time associated with compression processing in conventional moving picture encoders.

SUMMARY OF THE INVENTION

Therefore, an object of the present invention is to provide a transmitter which is capable of compressing an incoming image for transmission to a receiver with a relatively small delay time.

The present invention has the following features to attain the object above. According to one aspect of the invention, there is provided an image transmitter for compressing image data and transmitting the image data to a receiver via a transmission path, wherein the image data at least contains i pixel values of pixels arranged in line along a single direction, each pixel value being expressed in n bits, the image transmitter comprising: a blocking section for taking every p pixel values among the i pixel values in the image data to form a data block, and sequentially outputting a plurality of the data blocks each including the p pixel values; a data compression section for reducing an amount of data from each data block outputted from the blocking section to output a compressed block; and a data sending section for sending the compressed block outputted from the data compression section onto the transmission path, wherein i, n, and p are predetermined natural numbers.

These and other objects, features, aspects and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram illustrating a structure of an image transmitter Tx according to one embodiment of the present invention;

FIGS. 2A and 2B are diagrams illustrating an image MG represented by image data TD which is received by a blocking section 1 shown in FIG. 1;

FIG. 3 is a diagram illustrating a format of the image data TD which is received by the blocking section 1 shown in FIG. 1;

FIG. 4 is a diagram illustrating a format of data block DB_r which is outputted from the blocking section 1 shown in FIG. 1;

FIG. 5 is a block diagram illustrating a first implementation (a “data compression section 2a”) of a data compression section 2 shown in FIG. 1;

FIG. 6 is a block diagram illustrating the detailed structure of a DPCM encoding section 22 shown in FIG. 5;

FIG. 7 is a block diagram illustrating the detailed structure of a near-instantaneous compression section 23 shown in FIG. 5;

FIG. 8 is a diagram illustrating a format of differential data DD_p×(r−1)+2 to DD_p×r which are outputted from the a format of the DPCM encoding section 22 shown in FIG. 5;

FIG. 9 is a diagram illustrating bit patterns BP₁ to BP_n of maximum differential data MDD_v to be selected by a differential data selection section 2321 shown in FIG. 7;

FIG. 10 is a table illustrating the relationship between bit patterns BP₁ to BP_n shown in FIG. 9 and level values LV₁ to LV_t+1;

FIG. 11 is a diagram illustrating a format of compressed differential data CDD_p×(r−1)+2 to CDD_p×r which are outputted from a data reduction section 233 shown in FIG. 7;

FIG. 12 is a diagram illustrating an exemplary relationship between the level values LV₁ to LV_t+1 shown in FIG. 10 and the bits to be deleted from the differential data DD_p×(r−1)+2 to DD_p×r;

FIG. 13 is a diagram illustrating another exemplary relationship between the level values LV₁ to LV_t+1 shown in FIG. 10 and the bits to be deleted from the differential data DD_p×(r−1)+2 to DD_p×r;

FIG. 14 is a diagram illustrating a format of data packet DP_r which is outputted from a packet assembling section 24 shown in FIG. 5;

FIG. 15 is a block diagram illustrating a first implementation (a “receiver Rxa”) of a receiver Rx shown in FIG. 1;

FIG. 16 is a block diagram illustrating the detailed structure of a decompression/decoding section 7 shown in FIG. 15;

FIG. 17 is a block diagram illustrating the detailed structure of a DPCM decoding section 72 shown in FIG. 15;

FIGS. 18A and 18B are diagrams illustrating the processing to be performed by a near-instantaneous decompression section 71 shown in FIG. 16 responsive to the level value LV₁;

FIGS. 19A and 19B are diagrams illustrating the processing to be performed by the near-instantaneous decompression section 71 shown in FIG. 16 responsive to the level value LV₂;

FIG. 20 is a diagram illustrating a format of reproduced image data RTD to be reproduced by an image data reproduction section 8 shown in FIG. 15;

FIG. 21 is a diagram illustrating the generation of padded data block PDB performed by the blocking section 1;

FIG. 22 is a block diagram illustrating a variant (a “data compression section 2b”) of the data compression section 2a shown in FIG. 5;

FIG. 23 is a block diagram illustrating the detailed structure of a DPCM encoding section 25 shown in FIG. 22;

FIG. 24 is a block diagram illustrating the structure of a variant (“receiver Rxb”) of the receiver Rxa shown in FIG. 15;

FIG. 25 is a diagram illustrating an exemplary process performed by a missing block recovery section 9 shown in FIG. 24;

FIG. 26 is a block diagram illustrating a second implementation (a “data compression section 2c”) of the data compression section 2 shown in FIG. 1;

FIG. 27 is a diagram illustrating a format of coefficients CF_p×(r−1)+1 to CF_p×r which are outputted from an orthogonal transform section 26 shown in FIG. 26;

FIG. 28 is a diagram illustrating a portion of the process performed by a data reduction section 27 shown in FIG. 26;

FIG. 29 is a diagram illustrating the rest of the process performed by the data reduction section 27 shown in FIG. 26;

FIG. 30 is a block diagram illustrating a second implementation (a “receiver Rxc”) of the receiver Rx shown in FIG. 1;

FIG. 31 is a block diagram illustrating the overall configuration of a driving assistant system TS₁ incorporating transmitters Tx and receiver Rx as shown in FIG. 1;

FIG. 32 is a diagram illustrating a process performed by the driving assistant system TS₁ shown in FIG. 31;

FIG. 33 is a schematic diagram illustrating technological effects associated with the driving assistant system TS₁ shown in FIG. 31; and

FIG. 34 is a block diagram illustrating the overall structure of a remote control system TS₂ incorporating a transmitter Tx and a receiver Rx as shown in FIG. 1.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

FIG. 1 is a block diagram illustrating an overall structure of an image transmitter Tx according to one embodiment of the present invention. As shown in FIG. 1, the image transmitter Tx is constructed so as to be capable of data communication with a receiver Rx via a transmission path N, and includes a blocking section 1, a data compression section 2, and a data sending section 3. The transmission path N may be wired or wireless.

The blocking section 1 receives image data TD to be processed. As shown in FIG. 2A, the image data TD typically represents one frame of image MG. In FIG. 2A, the image MG is composed of (i×j) pixels PE. Herein, “i” and “j” are predetermined natural numbers, which in the present embodiment are assumed to be 640 and 480, respectively. More specifically, the image MG has a width equal to i pixels PE along a width (horizontal) direction HD. The image MG has a length equal to j pixels PE along a longitudinal (vertical) direction VD, the longitudinal direction VD running perpendicular to the width direction HD. Although only one pixel PE (a rectangular region which is shown hatched) is labeled as “PE” in FIG. 2A for conciseness, it will be appreciated that each rectangular region in the image MG represents a pixel PE. FIG. 2A also shows that the value of each pixel PE is expressed in an n-bit binary format. Herein, “n” is a predetermined natural number, which in the present embodiment is assumed to be 8. In the following description, the value of each pixel PE will be referred to as a “pixel value XV”.

As shown in FIG. 2B, a pixel PE in the image MG which is located at a k^th position along the width direction HD and at an m^th position along the longitudinal direction VD will conveniently be expressed as a “pixel PE_j×(m−1)+k”, whose value will be expressed as a “pixel value XV_j×(m−1)+k”. Herein, “k” is a natural number such that 1≦k≦i; and “m” is a natural number such that 1≦m≦j. The suffix “i×(m−1)+k” for the reference numeral “PF” represents an order by which the blocking section 1 receives the respective pixels PE. For example, under the above assumption where the image MG is composed of 640×480 pixels PE, a pixel PE which is located at the first position along the width direction HD and at the first position along the longitudinal direction VD will be represented as a “pixel PE₁”. Similarly, a pixel PE which is located at the 100^th position along the width direction HD and at the 121^st position along the longitudinal direction VD will be represented as a “pixel PE₇₆₉₀₀”. Similarly, the last pixel PE, or a pixel which is located at the 640^th position along the width direction HD and at the 480^th position along the longitudinal direction VD, will be represented as a “pixel PE₃₀₇₂₀₀”.

As shown in FIG. 3, the image data TD which is received by the blocking section 1 has a format such that the image data TD is composed of (i×j) pixel values XV_i×(m−1)+k (m=1,2, . . . j, k=1,2, . . . i), that is, the image data TD is a collection of pixel values XV₁ to XV_i×j. More specifically, in the image data TD, the pixel value XV₁ is located at the beginning, immediately followed by a pixel value XV₂. In turn, the pixel value XV₂ is followed by pixel values XV₃ to XV_i, which compose the rest of the first line along the width direction HD. Then comes a second line of pixel values XV_i+1, followed by pixel values XV_i+2 to XV_2×i. Similarly, third, fourth, . . . , and j^th lines of pixel values XV_2×i+1 to XV_3×i, pixel values XV_3×i+1 to XV_4×i, . . . and pixel values XV_i×(j−1)+1 to XV_i×j respectively follow.

In addition to the pixel values XV₁ to XV_i×j, the image data TD may contain any other additional information to be used for certain purposes. However, such additional information is not essential for the present embodiment of the invention, and therefore is omitted from the illustration and descriptions.

In accordance with the aforementioned format of the image data TD, the blocking section 1 receives the pixel values XV₁ to XV_i×j in this order. Among the incoming pixel values XV₁ to XV_i×j, the blocking section 1 takes every p (where p is a predetermined number) pixel values to generate q data blocks DB, each of which is composed of p pixel values. Herein, “p” is a divisor of i, and in the present embodiment is assumed to be 8; and “q” is equal to {(i×j)/p}. Under the above assumptions where i=640, j=480, and p=8, q is 38400. For conciseness, a data block DB which is the r^th generated data block will be denoted as a “data block DB_r” with a suffix r. Herein, “r” is a natural number such that 1≦r≦q. As shown in FIG. 4, such a data block DB_r (r=1,2, . . . q) will be a collection of p pixel values XV_p×(r−1)+1 to XV_p×r. Under the above assumptions, the first generated data block DB₁ is a collection of pixel values XV₁ to XV₈, given that r=1. Similarly, the second data block DB₂ is composed of pixel values XV₉ to XV₁₆. The last generated data block DB₃₈₄₀₀ is composed of pixel values XV307192 to XV₃₀₇₂₀₀. Such data blocks DB_r are outputted from the blocking section 1 to the data compression section 2 in the order in which they are generated (see FIG. 1).

As described in connection with the background art, according to the typical conventional image compression schemes such as the MPEG scheme or the DVC scheme, a number of unnecessary pixels are likely to have been received before all of the pixels composing a macro block are received, giving rise to an unwanted delay time. In contrast, according to the present embodiment of the invention, the blocking section 1 generates a data block DB_r which is composed of p pixel values XV₁ to XV_i×j in the order they are received. In other words, the blocking section 1 becomes ready to construct a data block DB_r as soon as it receives p contiguous pixel values XV along the width direction HD, which can then be quickly passed to the next stage, i.e., the data compression section 2. As a result, the delay time can be reduced as compared to the conventional image compression schemes.

The data compression section 2 performs a first or second compression process (described later) for each incoming data block DB_r. Thus, the data compression section 2 generates a compressed block CB_r (r=1,2, . . . q) having a fixed length composed of predetermined s bits. Each compressed block CB_r thus generated is outputted from the data compression section 2 to the data sending section 3 (see FIG. 1).

Next, with reference to FIGS. 5 to 7, a first exemplary implementation of the data compression section 2 shown in FIG. 1 will be described. In the following description, the first implementation of the data compression section 2 will be referred to as a “data compression section 2a”. As shown in FIG. 5, in order to realize the aforementioned first compression process, the data compression section 2a includes a splitter section 21, a DPCM encoding section 22, a near-instantaneous compression section 23, and a packet assembling section 24. As specifically shown in FIG. 6, the DPCM encoding section 22 of FIG. 5 includes a delay section 221 and a subtraction section 222. As specifically shown in FIG. 7, the near-instantaneous compression section 23 shown in FIG. 5 includes a buffer section 231, a level determination section 232, and a data reduction section 233.

Next, the first compression process which is performed by the data compression section 2a having the above-described structure will be specifically described. Data blocks DB_r which are sent from the blocking section 1 (described above) are sequentially received by the splitter section 21 in the data compression section 2a. Each data block DB_r includes an array of p pixel values XV_p×(r−1)+1 to XV_p×r (see FIG. 4). As shown in FIG. 6, as each data block DB_r is received, the splitter section 21 outputs the pixel value XV_p×(r−1)+1, which is located at the beginning of the received data block DB_r, to both the packet assembling section 24 and the delay section 221. Moreover, the splitter section 21 outputs each of the pixel values XV_p×(r−1)+2 to XV_p×r−1 to both the delay section 221 and the subtraction section 222. Furthermore, the splitter section 21 outputs a p^th pixel value XV_p×r to the subtraction section 222.

As shown in FIG. 6, the DPCM encoding section 22 receives pixel values XV_p×(r−1)+1 to XV_p×r. The DPCM encoding section 22 performs a differential pulse code modulation (Differential Pulse Code Modulation) to encode each of the received pixel values XV_p×(r−1)+2 to XV_p×r, thereby generating differential data DD_p×(r−1)+2 to DD_p×r, which are then outputted to the near-instantaneous compression section 23.

More specifically, the delay section 221 sequentially receives pixel values XV_p×(r−1)+1 to XV_p×r−1. The delay section 221 applies a delay amount DL₁ to the respective received pixel values XV_p×(r−1)+1 to XV_p×r−1, and outputs the resultant pixel values as delayed pixel values LXV_p×(r−1)+1 to LXV_p×r−1 to the subtraction section 222. Now, the delay amount DL₁ will be specifically described. The subtraction section 222 receives (as described later) pixel values XV_p×(r−1)+2 to XV_p×r from the splitter section 21. The delay amount DL₁ is prescribed at a value which ensures that the delayed pixel values LXV_p×(r−1)+1 to LXV_p×r−1 will be received by the subtraction section 222 substantially concurrently with the pixel values XV_p×(r−1)+2 to XV_p×r. In the present embodiment of the invention, the delay amount DL₁ is prescribed to be equal to one clock which defines the operation timing of the DPCM encoding section 22.

The subtraction section 222 receives the pixel values XV_p×(r−1)+2 to XV_p×r from the splitter section 21. The subtraction section 222 also receives the delayed pixel values LXV_p×(r−1)+1 to LXV_p×r−1 from the delay section 221. Note that the aforementioned delay amount DL₁ ensures that the delayed pixel value LXV_p×(r−1)+1 from the delay section 221 and the pixel value XV_p×(r−1)+2 from the splitter section 21 are received by the subtraction section 222 substantially simultaneously. The subtraction section 222 subtracts the currently-received delayed pixel value LXV_p×(r−1)+1 from the currently-received pixel value XV_p×(r−1)+2 to generate a differential data DD_p×(r−1)+2 representing a difference value therebetween. In other words, the subtraction section 222 calculates a difference value between the received pixel value XV_p×(r−1)+2 and a preceding pixel value XV_{p×(r−1)−1} in the image data TD (see FIG. 2B). This difference value may take a positive or negative value. Accordingly, as shown in FIG. 8, a sign bit SB_p×(r−1)+2, which is a one-bit expression of the sign (i.e., positive or negative) of the difference value, is added to the differential data DD_p×(r−1)+2 as its most significant bit (hereinafter referred to as the “MSB”). In the present embodiment of the invention, it is assumed that the sign bit SB_p×(r−1)+2 is “0” when the difference value is positive, and “1” when the difference value is negative. The sign bit SB_p×(r−1)+2 is followed by an n-bit expression of the absolute value of the difference value AV_p×(r−1)+2 (=|pixel value XV_p×(r−1)+2−pixel value XV_p×(r−1)+1|). Thus, the differential data DD_p×(r−1)+2 is composed of (n+1) bits.

Moreover, the subtraction section 222 subtracts the delayed pixel value LXV_p×(r−1)+2 from the concurrently-received pixel value XV_p×(r−1)+3 to generate a differential data DD_p×(r−1)+3. Thereafter, the subtraction section 222 repeats similar processes until it generates a differential data DD_p×r from the pixel value XV_p×r and the delayed pixel value LXV_p×r−1. As can be seen from FIG. 8, the differential data DD_p×(r−1)+3 to DD_p×r have a format which is similar to that of the differential data DD_p×(r−1)+2. As shown in FIG. 6, the subtraction section 222 sequentially outputs the generated differential data DD_p×(r−1)+2 to DD_p×r to the near-instantaneous compression section 23.

As described above, q data blocks DB_r are generated for one frame of image MG, and the DPCM encoding section 22 generates differential data DD_p×(r−1)+2 to DD_p×r for each data block DB_r. Therefore, the near-instantaneous compression section 23 sequentially receives q sets of differential data DD_p×(r−1)+2 to DD_p×r. The near-instantaneous compression section 23 generates compressed data CD_p×(r−1)+2 to CD_p×r from each set of received differential data DD_p×(r−1)+2 to DD_p×r, respectively, in accordance with a near-instantaneous compression scheme.

More specifically, the buffer section 231 shown in FIG. 7, which is constructed so as to be capable of storing (p−1)×(n+1) bits of data, stores the differential data DD_p×(r−1)+2 to DD_p×r from the subtraction section 222. The buffer section 231 outputs the stored differential data DD_p×(r−1)+2 to DD_p×r to both the level determination section 232 and the data reduction section 233.

The level determination section 232 generates one level value LV_r for every set of differential data DD_p×(r−1)+2 to DD_p×r received from the buffer section 231. The data reduction section 233 deletes predetermined t bits from each of the received differential data DD_p×(r−1)+2 to DD_p×r, as will be described in more detail later. Herein, “t” is a natural number such that 1≦t<(n+1), and in the present embodiment is assumed to be 5. Stated differently, the data reduction section 233 leaves u bits intact among the (n+1) bits which compose each of the differential data DD_p×(r−1)+2 to DD_p×r. Herein, “u” is equal to (n+1−t), which under the above assumptions is 4. The level value LV_r specifies the positions of the u bits to be left intact among the (n+1) bits which compose each of the differential data DD_p×(r−1)+2 to DD_p×r.

In order to derive the aforementioned level value LV_r, the level determination section 232 shown in FIG. 7 includes a differential data selection section 2321 and a level selection section 2322. The differential data selection section 2321 receives the differential data DD_p×(r−1)+2 to DD_p×r (see FIG. 8) from the buffer section 231. The differential data selection section 2321 selects one of the differential data DD_p×(r−1)+2 to DD_p×r having the greatest absolute values AV_p×(r−1)+2 to AV_p×r−1, and outputs that differential data to the level selection section 2322. In the following description, the differential data which is thus selected by the differential data selection section 2321 will be referred to as the “maximum differential data MDD_v”. Herein, “v” is a natural number in the range from {p×(r−1)+2} to p×r.

The level selection section 2322 receives the maximum differential data MDD_v from the differential data selection section 2321. The level selection section 2322 determines the sign (i.e., positive or negative) of the received maximum differential data MDD_v based on the value of its sign bit SB_v.

If the current maximum differential data MDD_v has a positive value (i.e., the sign bit SB_v is “0”), then the level selection section 2322 operates in the following manner. Herein, in the case where the sign bit SB_v is “0”, the maximum differential data MDD_v has one of n bit patterns BP₁ to BP_n as shown in FIG. 9. As shown, the bit pattern BP₁ is a bit pattern in which the first instance of “1” appears next to the sign bit SB_v, i.e., at the second bit from the MSB. Similarly, the bit patterns BP₂ to BP_n are bit patterns in which the first instance of “1” appears at the third to (n+1)^th bits from the MSB, respectively. In FIG. 9, any bit shown by the symbol “−” may take either “0” or “1”. In the following description, in the positive maximum differential data MDD_v, the bit position at which the first instance of “1” appears (i.e., one of the second to the (n+1)^th bits), as counted from the MSB, will be referred to as a “reference bit position” RBL₁ to RBL_n.

In the present embodiment of the invention, as shown in FIG. 10, (t+1) level values LV₁ to LV_t+1 are previously assigned to the bit patterns BP₁ to BP_n in the following manner. As described above, “t” represents the number of bits to be deleted from each of the differential data DD_p×(r−1)+2 to DD_p×r, which in the present embodiment is assumed to be 5. Specifically, level values LV₁ to LV_t are assigned to the bit patterns BP₁ to BP_t, respectively. The same level value LV_t+1 is assigned to all of the bit patterns BP_t+1 to BP_n. In other words, the respective level values LV₁ to LV_t are assigned to the reference bit positions RBL₁ to RBL_t, whereas the same level value LV_t+1 is assigned to the reference bit positions RBL_t+1 to RBL_n.

In the following description, as shown in FIG. 10, the bit pattern BP_w is one of the bit patterns BP₁ to BP_n. Similarly, the reference bit position RBL_w is one of the reference bit positions RBL₁ to RBL_n. In other words, w is a natural number in the range from 1 to n. The level value LV_y is one of the level values LV₁ to LV_t+1, and is expressed in z bits; y is a natural number in the range from 1 to (t+1); and z is the number of digits which are required when converting (t+1) into a binary expression. Preferably, the value of z is as small as possible. For example, when t=5, z is most preferably 3.

The level selection section 2322 detects the bit position at which the first instance of “1” appears, the check being begun at the MSB of the current maximum differential data MDD_v. In other words, the level selection section 2322 determines one of the reference bit positions RBL₁ to RBL_n which corresponds to the current maximum differential data MDD_v. Next, the level selection section 2322 selects one of the level values LV₁ to LV_t+1 which is assigned to the currently-determined one of the reference bit positions RBL₁ to RBL_n, and outputs this level value to both the data reduction section 233 and the packet assembling section 24. Among the level values LV₁ to LV_t+1, the currently-outputted level value is referred to as the “level value LV_y”, as defined earlier.

As described above, the data reduction section 233 receives the differential data DD_p×(r−1)+2 to DD_p×r from the buffer section 231. The data reduction section 223 also receives the level value LV_y from the level determination section 232. Based on the current level value LV_y, the data reduction section 233 deletes t bits from each of the current differential data DD_p×(r−1)+2 to DD_p×r. As a result, as shown in FIG. 11, the data reduction section 233 generates compressed differential data CDD_p×(r−1)+2 to CDD_p×r, each of which is composed of u bits.

According to the present embodiment of the invention, as shown in FIG. 12, the bits to be deleted from the current differential data DD_p×(r−1)+2 to DD_p×r are predetermined for each of the aforementioned level values LV₁ to LV_t+1. For the sake of illustration, FIG. 12 shows an exemplary case in which n=8, t=5, and u=4. Specifically, regardless of which one of the level values LV₁ to LV_t+1 is currently received, the data reduction section 233 leaves intact (i.e., without deleting) the sign bits SB_p×(r−1)+2 to SB_p×r of all differential data DD_p×(r−1)+2 to DD_p×r. When the level value LV₁ is received, the data reduction section 233 leaves intact the aforementioned reference bit position RBL₁ and the following (u−2) bits (i.e., the (u−2) bits immediately on the lower side) contained in each of the differential data DD_p×(r−1)+2 to DD_p×r, while deleting the other t bits. Similarly, when one of the level values LV₂ to LV_t is received, the data reduction section 233 leaves intact the aforementioned reference bit positions RBL₂ to RBL_t and the following (u−2) bits in each of the differential data DD_p×(r−1)+2 to DD_p×r. When the level value LV_t+1 is received, the data reduction section 233 leaves intact the lower (u−1) bits in each of the differential data DD_p×(r−1)+2 to DD_p×r, while deleting the second to (u−2)^th bits as counted from the MSB.

As mentioned earlier, it is assumed that the sign bits SB_p×(r−1)+2 to SB_p×r are “0”. Under this assumption, the data reduction section 233 may alternatively operate in the manner shown in FIG. 13, so that when one of the level values LV₁ to LV_t is received, the data reduction section 233 simply leaves intact the reference bit positions RBL₂ to RBL_t as well as its preceding bit and the following (u−2) bits in each of the differential data DD_p×(r−1)+2 to DD_p×r. When the level value LV_t+1 is received, the lower u bits are left intact in each of the differential data DD_p×(r−1)+2 to DD_p×r. Note that the operation scheme of FIG. 13 is directed to the same exemplary case which FIG. 12 is drawn to, i.e., n=8, t=5, and u=4.

Thus, the data reduction section 233 deletes t bits from each of the currently-received differential data DD_p×(r−1)+2 to DD_p×r to generate compressed differential data CDD_p×(r−1)+2 to CDD_p×r, each of which is composed of u bits. The respective generated compressed differential data CDD_p×(r−1)+2 to CDD_p×r are outputted to the packet assembling section 24.

The above description illustrates the case where the sign bits SB_p×(r−1)+2 to SB_p×r are “0”. On the other hand, in the case where the sign bits SB_p×(r−1)+2 to SB_p×r are “1”, the level selection section 2322 detects the reference bit position at which the first instance of “0” appears, as counted from the MSB of the currently-received maximum differential data MDD_v. Furthermore, the level selection section 2322 selects one of the level values LV₁ to LV_t+1 which is assigned to the currently-determined reference bit position, and outputs this level value to both the data reduction section 233 and the packet assembling section 24.

As mentioned earlier, the packet assembling section 24 receives the pixel value XV_p×(r−1)+1 from the splitter section 21 as well as the level value LV_y from the level determination section 232. The packet assembling section 24 also receives the compressed differential data CDD_p×(r−1)+2 to CDD_p×r from the data reduction section 233. Based on these received data, the packet assembling section 24 assembles a data packet DP_r as shown in FIG. 14. As shown in FIG. 14, the assembled data packet DP_r contains the pixel value XV_p×(r−1)+1, the level value LV_y, and the compressed differential data CDD_p×(r−1)+2 to CDD_p×r, and has a fixed length of {(n+z+u×(n−1)} bits. The packet assembling section 24 outputs the assembled data packet DP_r to the data sending section 3 as the aforementioned compressed block CB_r (see FIG. 1).

As shown in FIG. 1, the data sending section 3 includes a buffer section 31 and a sending control section 32. The buffer section 31 stores the fixed-length data packet DP_r. Since typical conventional image compression schemes such as the MPEG scheme or the DVC scheme perform a variable-length coding, a large-capacity transmission buffer is inevitably required. The use of such a large-capacity transmission buffer for buffering variable-length encoded data may introduce a substantial delay time in the conventional image compression schemes. In contrast, according to the present embodiment of the invention, the buffer section 31 only needs to store the fixed-length data packet DP_r, so that the delay time associated with the buffering which takes place in the buffer section 31 is minimized, and the delay time becomes substantially constant for every data packet DP_r. Thus, the use of fixed-length data packet DP_r also serves to reduce the delay time relative to that which is associated with the conventional image compression schemes.

After the data packet DP_r is stored in the buffer section 31, the sending control section 32 receives the data packet DP_r from the buffer section 31 and sends it onto the transmission path N. The data packet DP_r, as an example of compressed data CD_r, is transmitted through the transmission path N and then received by the receiver Rx, as shown in FIG. 1.

The receiver Rx subjects the received data packet DP_r to predetermined processing. Hereinafter, with reference to FIG. 15 to FIG. 17, a first implementation of the receiver Rx shown in FIG. 1 will be described. In the following description, the first implementation of the receiver Rx will be referred to as a “receiver Rxa”. As shown in FIG. 15, in order to realize reproduction processing for the image data TD, the receiver Rxa includes a data receiving section 5, a packet deassembling (disassembling) section 6, a decompression/decoding section 7, and an image data reproduction section 8. The data receiving section 5 includes a buffer section 51 and a reception control section 52. As shown in FIG. 16, the decompression/decoding section 7 includes a near-instantaneous decompression section 71 and a DPCM decoding section 72. As shown in FIG. 17, the DPCM decoding section 72 includes a delay section 721 and an adder section 722.

Next, the reproduction processing for the image data TD which is performed by the receiver Rxa will be described in detail. The data packets DP_r from the transmission path N are sequentially received by the receiving section 5. In the receiving section 5, the buffer section 51 stores the data packet DP_r. The buffer section 51, which is only required to store the fixed-length data packets DP_r, as is the case with the buffer section 31, contributes to the minimization of the delay time. After the buffering, the reception control section 52 receives the data packet DP_r from the buffer section 51, and outputs the data packet DP_r to the packet deassembling section 6.

As described above, the data packet DP_r contains the pixel value XV_p×(r−1)+1, the level value LV_y, and the compressed differential data CDD_p×(r−1)+2 to CDD_p×r (see FIG. 14). For every received data packet DP_r, the packet deassembling section 6 performs a deassembling process, and, as shown in FIG. 15, outputs the pixel value XV_p×(r−1)+1 which is located at the beginning of the received data packet DP_r to the image data reproduction section 8, and outputs the pixel value XV_p×(r−1)+1, the level value LV_y, and the compressed differential data CDD_p×(r−1)+2 to CDD_p×r to the decompression/decoding section 7. To describe the outputting to the decompression/decoding section 7 more specifically, as shown in FIG. 16, the packet deassembling section 6 sequentially outputs the level value LV_y and the compressed differential data CDD_p×(r−1)+2 to CDD_p×r to the near-instantaneous decompression section 71. Furthermore, as shown in FIG. 17, the packet deassembling section 6 outputs the pixel value XV_p×(r−1)+1 to the delay section 721.

As described earlier, the decompression/decoding section 7 receives the pixel value XV_p×(r−1)+1, the level value LV_y, and the compressed differential data CDD_p×(r−1)+2 to CDD_p×r, as shown in FIG. 15. The decompression/decoding section 7 employs the received pixel value XV_p×(r−1)+1 and the level value LV_y to perform a decompression/decoding process for the compressed differential data CDD_p×(r−1)+2 to CDD_p×r, thereby generating decoded pixel values DXV_p×(r−1)+2 to DXV_p×r which can be regarded as substantially the same as the aforementioned pixel values XV_p×(r−1)+2 to XV_p×r.

More specifically, the near-instantaneous decompression section 71 in the decompression/decoding section 7 receives the level value LV_y and the compressed differential data CDD_p×(r−1)+2 to CDD_p×r. The near-instantaneous decompression section 71 performs a near-instantaneous decompression to decompress the compressed differential data CDD_p×(r−1)+2 to CDD_p×r based on the received level value LV_y. Thus, the near-instantaneous decompression section 71 generates decompressed differential data DDD_p×(r−1)+2 to DDD_p×r, which are sequentially outputted to the adder section 722.

To describe the above process more specifically, the near-instantaneous decompression section 71 recognizes the reference bit position RBL_w in the received level value LV_y. Specifically, as can be seen from FIG. 10, one of the reference bit position RBL₁ to the reference bit position RBL_t will be recognized in the case where y is in the range from 1 to t. In the case where y=t+1, the near-instantaneous decompression section 71 will recognize one of the reference bit positions RBL_t+1 to RBL_n.

Once the reference bit position RBL_w is determined, the near-instantaneous decompression section 71 knows the bit positions which were deleted from the differential data DD_p×(r−1)+2 to DD_p×r in the transmitter Tx. More specifically, as can be seen from FIG. 12 or FIG. 13, when the level value LV₁ is received, the near-instantaneous decompression section 71 recognizes that t bits have been deleted from the differential data DD_p×(r−1)+2 to DD_p×r while leaving intact the MSB (i.e., sign bits SB_p×(r−1)+2 to SB_p×r), the reference bit position RBL₁, and the following (u−2) bits. Similarly, when the level values LV₂ to LV_t are received, the near-instantaneous decompression section 71 recognizes that t bits have been deleted from the differential data DD_p×(r−1)+2 to DD_p×r while leaving intact the respective MSBs, the aforementioned respective reference bit positions RBL₂ to RBL_t, and the following (u−2) bits. When the level value LV_t+1 is received, the near-instantaneous decompression section 71 recognizes that the second to (u−2)^th bits as counted from the MSB have been deleted from each of the differential data DD_p×(r−1)+2 to DD_p×r.

Having thus determined the deleted bit positions, the near-instantaneous decompression section 71 adds a bit(s) of predetermined values to each of the compressed differential data CDD_p×(r−1)+2 to CDD_p×r, thereby generating the decompressed differential data DDD_p×(r−1)+2 to DDD_p×r, which can be regarded as substantially the same as the aforementioned differential data DD_p×(r−1)+2 to DD_p×r.

More specifically, when receiving the level value LV₁, as shown in FIGS. 18A and 18B, the near-instantaneous decompression section 71 adds a bit sequence BS₁₁ or bit sequence BS₁₂ immediately after each of the compressed differential data CDD_p×(r−1)+2 to CDD_p×r. The bit sequences BS₁₁ and BS₁₂ each have a predetermined bit pattern composed of t bits. Note that the bit sequences BS₁₁ and BS₁₂ shown in FIGS. 18A and 18B are directed to the same exemplary case which FIG. 12 is drawn to, i.e., n=8, t=5, and u=4. As shown in FIG. 18A, the bit sequence BS₁₁, whose first bit is “1” and other (t−1) bits are “0”, is added to those compressed differential data CDD_p×(r−1)+2 to CDD_p×r which have positive values. On the other hand, the bit sequence BS₁₂, which has a reversed bit pattern (i.e., the first bit of the bit sequence BS₁₂ is “0” and the other (t−1) bits are “1”) relative to the bit sequence BS₁₁, is added to those compressed differential data CDD_p×(r−1)+2 to CDD_p×r which have negative values. Thus, the decompressed differential data DDD_p×(r−1)+2 to DDD_p×r are generated. The bit patterns of the bit sequences BS₁₁ and BS₁₂ are prescribed as above in order to ensure that the differences between the values of the decompressed differential data DDD_p×(r−1)+2 and the values of their corresponding original differential data DD_p×(r−1)+2 are small.

When receiving the level value LV₂, as shown in FIG. 19A, the near-instantaneous decompression section 71 places the MSBs of the respective compressed differential data CDD_p×(r−1)+2 to CDD_p×r at the first bit, while placing the other (u−1) bits at the third bit (i.e., the reference bit position RBL₂) through the (u+1)^th bit. Furthermore, for those compressed differential data CDD_p×(r−1)+2 to CDD_p×r which have positive values, the near-instantaneous decompression section 71 adds “0” at the second bit, while adding a bit sequence BS₂₁ composed of (t−1) bits at the (u+2)^th through n^th bits. Note that the bit sequence BS₂₁ is a sequence in which the only instance of “1” is at the first bit, as in the bit sequence BS₁₁. On the other hand, for those compressed differential data CDD_p×(r−1)+2 to CDD_p×r which have negative values, as shown in FIG. 19B, the near-instantaneous decompression section 71 adds “1” at the second bit, while adding a bit sequence BS₂₂ composed of (t−1) bits at the (u+2)^th through n^th bits. Note that the bit sequence BS₂₂ is a sequence in which the only instance of “0” is at the first bit.

Thereafter, when the level values LV₃ to LV_t are received, similarly to when the level value LV₂ is received, the near-instantaneous decompression section 71 places the MSBs of the respective compressed differential data CDD_p×(r−1)+2, CDD_p×(r−1)+3, . . . CDD_p×r at the first bit, while placing the other (u−1) bits at the reference bit positions RBL₃, RBL₄, . . . RBL_t through the (u+1)^th bit. Furthermore, for those compressed differential data CDD_p×(r−1)+2, CDD_p×(r−1)+3, . . . CDD_p×r which have positive values, the near-instantaneous decompression section 71 sets “0” at the second bit through the bit immediately before the reference bit positions RBL₃, RBL₄, . . . RBL_t, “1” at the (u+2)^th bit, and “0” at the (u+3)^th through n^th bits. For those compressed differential data CDD_p×(r−1)+2, CDD_p×(r−1)+3, . . . CDD_p×r which have negative values, bits which are reverses of those set for the compressed differential data having positive values are set. As shown in FIG. 16, the decompressed differential data DDD_p×(r−1)+2 to DDD_p×r which have been thus generated are outputted to the adder section 722 in the DPCM decoding section 72.

As shown in FIG. 16, the DPCM decoding section 72 receives the pixel value XV_p×(r−1)+1 from the packet deassembling section 6 and the decompressed differential data DDD_p×(r−1)+2 to DDD_p×r from the near-instantaneous decompression section 71. The DPCM decoding section 72 performs an inverse process of the process which is performed by the DPCM encoding section 22 so as to generate the decoded pixel values DXV_p×(r−1)+2 to DXV_p×r from the received pixel value XV_p×(r−1)+1 and the decompressed differential data DDD_p×(r−1)+2 to DDD_p×r, and sequentially outputs the decoded pixel values to the image data reproduction section 8.

More specifically, the delay section 721 in the DPCM decoding section 72 receives the pixel value XV_p×(r−1)+1 from the packet deassembling section 6. The delay section 721 applies a delay amount DL₂ to the received pixel value XV_p×(r−1)+1, and outputs the resultant pixel value to the adder section 722 as a delayed pixel value LXV_p×(r−1)+1. Herein, the delay amount DL₂ is typically an amount of time corresponding to predetermined clocks. More specifically, the delay amount DL₂ is prescribed to a value which ensures that the decompressed differential data DDD_p×(r−1)+2 from the near-instantaneous decompression section 71 and the delayed pixel value LXV_p×(r−1)+1 are received by the adder section 722 in the DPCM decoding section 72 substantially simultaneously.

The adder section 722 also sequentially receives sets of decompressed differential data DDD_p×(r−1)+2 to DDD_p×r. The adder section 722 adds the decompressed differential data DDD_p×(r−1)+2 (which is received before any other decompressed differential data) and the concurrently-received delayed pixel value LXV_p×(r−1)+1 to generate a decoded pixel value DXV_p×(r−1)+2. The decoded pixel value DXV_p×(r−1)+2 which has been thus generated is outputted to the image data reproduction section 8 as mentioned above, and is also fed back to the adder section 722. Next, the adder section 722 adds the decompressed differential data DDD_p×(r−1)+3 from the near-instantaneous decompression section 71 and the concurrently-received decoded pixel value DXV_p×(r−1)+2 to generate a decoded pixel value DXV_p×(r−1)+3. The decoded pixel value DXV_p×(r−1)+3 which has been thus generated is outputted to the image data reproduction section 8, and also fed back to the adder section 722. Thereafter, in a repetition of similar processing, the adder section 722 adds the decompressed differential data DDD_p×(r−1)+4, DDD_p×(r−1)+5, . . . DDD_p×r−1 from the near-instantaneous decompression section 71 and the previously-generated decoded pixel values DXV_p×(r−1)+3, DXV_p×(r−1)+4, . . . DXV_p×r−2 to generate decoded pixel values DXV_p×(r−1)+4, DXV_p×(r−1)+5, . . . DXV_p×r−1, respectively, which are outputted to the image data reproduction section 8 and itself. Furthermore, the adder section 722 adds the decompressed differential data DDD_p×r from the near-instantaneous decompression section 71 and the previously-generated decoded pixel value DXV_p×r−1 to generate a decoded pixel value DXV_p×r, which is outputted only to the image data reproduction section 8. Thus, the DPCM decoding section 72 generates the decoded pixel values DXV_p×(r−1)+2 to DXV_p×r, and outputs these decoded pixel values to the image data reproduction section 8.

As a result of the above-described processing, the image data reproduction section 8 sequentially receives q sets of decoded pixel values DXV_p×(r−1)+2 to DXV_p×r. Furthermore, prior to the arrival of each set of decoded pixel values DXV_p×(r−1)+2 to DXV_p×r, the pixel value XV_p×(r−1)+1 is received from the packet deassembling section 6. In summary, the image data reproduction section 8 first receives the pixel value XV₁ and the decoded pixel values DXV₂ to DXV_p. The pixel value XV₁ and the decoded pixel values DXV₂ to DXV_p which are thus generated are substantially identical to the pixel value XV₁ and the pixel values XV₂ to XV_p in the first line (along the width direction HD) of the image MG (see FIG. 2A or 2B). Subsequently, as the image data reproduction section 8 receives the q^th set of decoded pixel values, all of the pixel value XV_p×(r−1)+1 and the decoded pixel values DXV_p×(r−1)+2 to DXV_p×r which are necessary for the reproduction of the image MG are on hand. Thus, as shown in FIG. 20, the image data reproduction section 8 generates reproduced image data RTD, which represents an image which hardly presents any difference to the human eye from the image MD represented by the image data TD shown in FIG. 3, by arranging the pixel value XV_p×(r−1)+1 and the decoded pixel values DXV_p×(r−1)+2 to DXV_p×r in the order in which they are received.

As described above, according to the present embodiment of the invention, encoding and compression are performed for a fixed-length data block DB_r which is composed of p pixel values XV_p×(r−1)+1 to XV_p×r arranged in line along the width direction HD. In other words, unlike in the conventional image compression schemes (MPEG or DVC) where image correlation on a macro block-by-macro block basis is utilized, the correlation between pixels arranged in line along the width direction HD is utilized to compress an image MG. As a result, the delay time which is incurred before the reproduced image data RTD is generated by the receiver Rx can be reduced.

In the above embodiment of the invention, the blocking section 1 is illustrated as generating data blocks DB_r each composed of p received pixel values XV₁ to XV_i×j, where p is a divisor of i. However, i does not need to be an exact integer multiple of p, but any number a of pixel values XV may be left as a remainder. In such cases, as shown in FIG. 21, the blocking section 1 adds a padding bit sequence PBS composed of n×(p−a) bits after the a pixel values XV to generate padded data blocks PDB having the same size as that of the respective data blocks DB_r. Also in the case where i×j is not an exact integer multiple of p, it is preferable to generate similar padded data blocks PDB.

Based on the above consideration, the data compression section 2 can subject the padded data blocks to the same processing as that for the data blocks to generate compressed blocks. Furthermore, since the blocking section 1 only needs to add a padding bit sequence PBS for the pixel values XV arranged in line along the width direction HD, and there is no need to add any bit sequence along the longitudinal direction VD, the total number of extra bits to be added to the compressed data CD_r is much smaller than that required for a typical conventional image compression scheme such as the MPEG scheme.

Next, a variant of the above-described data compression section 2a will be described with reference to FIGS. 22 and 23. In the following description, the variant of the data compression section 2a will be referred to as a “data compression section 2b”. As shown in FIG. 22, the data compression section 2b differs from the data compression section 2a in that a DPCM encoding section 25 is employed instead of the DPCM encoding section 22. This difference is underlined by the fact that the DPCM encoding section 25 receives compressed differential data CDD_p×(r−1)+1 to CDD_p×r and a level value LV_y from a near-instantaneous compression section 23. Since the data compression section 2b is otherwise identical to the data compression section 2a, any component elements in the data compression section 2b which find their counterparts in the data compression section 2a are denoted by the same reference numerals as those employed in connection with the data compression section 2a, and the descriptions thereof are omitted.

As shown in FIG. 23, the DPCM encoding section 25 differs from the DPCM encoding section 22 in that the DPCM encoding section 25 additionally includes the near-instantaneous decompression section 251 and the adder section 252, and that the delay section 252 is employed instead of the delay section 221. Since the DPCM encoding section 25 is otherwise identical to the DPCM encoding section 22, any component elements in the DPCM encoding section 25 which find their counterparts in the DPCM encoding section 22 are denoted by the same reference numerals as those employed in connection with the DPCM encoding section 22, and the descriptions thereof are omitted.

In the DPCM encoding section 25, the near-instantaneous decompression section 251 receives the level value LV_y from the level determination section 232 and the compressed differential data CDD_p×(r−1)+2 to CDD_p×r from the data reduction section 233. By performing a near-instantaneous decompression similar to that performed by the near-instantaneous decompression section 71 shown in FIG. 16, the near-instantaneous decompression section 251 generates decompressed differential data DDD_p×(r−1)+2 to DDD_p×r from the received compressed differential data CDD_p×(r−1)+2 to CDD_p×r in accordance with the level value LV_y, and outputs the decompressed differential data DDD_p×(r−1)+2 to DDD_p×r to the adder section 252.

The delay section 253 sequentially receives the pixel values XV_p×(r−1)+1 to XV_p×r−1, as does the aforementioned delay section 221. The delay section 253 applies a delay amount DL₃ to each of the received pixel values XV_p×(r−1)+1 to XV_p×r−1 to generate delayed pixel values LXV_p×(r−1)+1 to LXV_p×r−1, which are outputted to the adder section 252. The delay amount DL₃ is prescribed to a value which ensures that the delayed pixel value LXV_{p×(r−1)×2} from the delay section 253 and the decompressed differential data DDD_p×(r−1)+2 from the near-instantaneous decompression section 251 are received by the adder section 252 substantially simultaneously. Generally speaking, the delay amount DL₃ is prescribed to a value which ensures that the delayed pixel values LXV_p×(r−1)+2, LXV_p×(r−1)+3, . . . LXV_p×r−1 and the compressed differential data CDD_p×(r−1)+2, CDD_p×(r−1)+3, . . . CDD_p×r−1 (which are generated on the basis of the same pixel values XV_p×(r−1)+2, XV_p×(r−1)+3, . . . XV_p×r−1, respectively) are received by the adder section 252 substantially simultaneously.

The adder section 252 adds the delayed pixel values LXV_p×(r−1)+2, LXV_p×(r−1)+3, . . . LXV_p×r−1 and the concurrently-received compressed differential data CDD_p×(r−1)+2, CDD_p×(r−1)+3, . . . CDD_p×r−1 to generate added pixel values AXV_p×(r−1)+2, AXV_p×(r−1)+3, . . . AXV_p×r−1, which is outputted to the subtraction section 222. Note that, when the delayed pixel value LXV_p×(r−1)+1 from the delay section 253 is received by the adder section 252, the adder section 252 is receiving no input from the near-instantaneous decompression section 251, so that the adder section 252 passes the received delayed pixel value LXV_p×(r−1)+1 through to the subtraction section 222 as the added pixel value AXV_p×(r−1)+1.

As described earlier, the subtraction section 222 receives the pixel values XV_p×(r−1)+2 to XV_p×r from the splitter section 21, and the added pixel values AXV_p×(r−1)+1 to AXV_p×r−1 from the adder section 252. First, the subtraction section 222 subtracts the added pixel value AXV_p×(r−1)+1 from the currently-received pixel value XV_p×(r−1)+2, and generates a differential data DD_p×(r−1)+2 representing a difference value therebetween, such that the generated differential data DD_p×(r−1)+2 is in the same format as that shown in FIG. 8. Furthermore, the subtraction section 222 subtracts the added pixel value AXV_p×(r−1)+2 from the concurrently-received pixel value XV_p×(r−1)+3 to generate a differential data DD_p×(r−1)+3. Thereafter, the subtraction section 222 repeats similar processes until it generates the aforementioned differential data DD_p×r. As shown in FIG. 23, the subtraction section 222 sequentially outputs the generated differential data DD_p×(r−1)+2 to DD_p×r to the near-instantaneous compression section 23.

For various reasons such as transmission errors, it might be possible for the receiver Rxa to miss, i.e., fail to correctly receive, all of the data packet DP_r which have been sent from the transmitter Tx. Next, with reference to FIG. 24, a variant of the receiver Rxa which can solve the above problem will be described. In the following description, the variant of the receiver Rxa will be referred to as a “receiver Rxb”. As shown in FIG. 24, the receiver Rxb differs from the receiver Rxa in that the receiver Rxb includes a missing block recovery section 9 immediately before the image data reproduction section 8. Since the receiver Rxb is otherwise identical to the receiver Rxa, any component elements in the receiver Rxb which find their counterparts in the receiver Rxa are denoted by the same reference numerals as those employed in connection with the receiver Rxa, and the descriptions thereof are omitted.

The missing block recovery section 9 receives q sets of decoded pixel values DXV_p×(r−1)+2 to DXV_p×r from the decompression/decoding section 7. Prior to the arrival of each set of decoded pixel values DXV_p×(r−1)+2 to DXV_p×r, the missing block recovery section 9 receives the pixel value XV_p×(r−1)+1 from the packet deassembling section 6. Thus, granted that there is no transmission error or the like, the missing block recovery section 9 will first receive the pixel value XV₁ and the decoded pixel values DXV₂ to DXV_p. Subsequently, as the missing block recovery section 9 receives the q^th set of decoded pixel values, all of the pixel value XV_p×(r−1)+1 and the decoded pixel values DXV_p×(r−1)+2 to DXV_p×r which are necessary for the reproduction of the image MG are on hand. Thus, when the pixel value XV_p×(r−1)+1 and the set of decoded pixel values DXV_p×(r−1)+2 to DXV_p×r, which together compose one frame, are all correctly received, the missing block recovery section 9 sequentially outputs these values to the image data reproduction section 8.

However, due to the aforementioned transmission error or the like, the receiver Rxb may miss or fail to receive the pixel value XV_p×(r−1)+1 and the decoded pixel values DXV_p×(r−1)+2 to DXV_p×r to be generated from one or more data packets DP_r. In the following description, any data packets DP_r which the receiver Rxb fails to receive will be referred to as “missing data packets DDP_r”. In such cases, the missing block recovery section 9 is able to virtually reproduce the pixel values XV_p×(r−1)+1 to DXV_p×r which have been contained in the missing data packet DDP_r from the correctly-generated set of the pixel value XV_p×(r−1)+1 and the decoded pixel values DXV_p×(r−1)+2 to DXV_p×r.

For example, as shown in FIG. 25, it is assumed that a data packet DP_r1 which is generated as an r₁^th data packet in the transmitter Tx happens to become a missing data packet DDP_r1, and that the receiver Rxb is correctly receiving all the other data packets DP_r. In this exemplary case, the missing block recovery section 9 first selects the pixel value XV_{p×((r1−i/p)−1)+1} and the decoded pixel values DXV_{p×((r1−i/p)−1)+2} to DXV_{p×(r1−i/p)} which have been generated from a data packet DP_r1−i/p which lies (i/p) data packets before the missing data packet DP_r1. Note that the pixel value XV_{p×((r1−i/p)−1)+1} and the decoded pixel values DXV_{p×((r1−i/p)−1)+2} to DXV_{p×(r1−i/p)} are located one line above (along the longitudinal direction VD) the pixel value XV_{p×(r1−1)+1} and the decoded pixel values DXV_{p×(r1−1)+2} to DXV_p×r1 which would have been generated from the missing data packet DP_r1. Furthermore, the missing block recovery section 9 selects the pixel value XV_{p×((r1+i/p)−1)+1} and the decoded pixel values DXV_{p×((r1+i/p)−1)+2} to DXV_p×(r1+i/p) which have been generated from a data packet DP_r1+i/p which lies (i/p) after the missing data packet DP_r1. Note that the pixel value XV_{p×((r1+i/p)−1)+1} and the decoded pixel values DXV_{p×((r1+i/p)−1)+2} to DXV_p×(r1+i/p) are located one line below (along the longitudinal direction VD) the pixel value XV_{p×(r1−1)+1} and the decoded pixel values DXV_{p×(r1−1)+2} to DXV_p×r1 which would have been generated from the missing data packet DP_r1.

Next, the missing block recovery section 9 assigns an average value AV_{p×(r1−1)+1} of the pixel value XV_{p×((r1−i/p)−1)+1} and the XV_{p×((r1+i/p)−1)+1} as the pixel value XV_{p×(r1−1)+1}. Moreover, the missing block recovery section 9 assigns an average value AV_{p×(r1−1)+2} of the decoded pixel value DXV_{p×((r1−i/p)−1)+2} and DXV_{p×((r1+i/p)−1)+2} as the decoded pixel value XV_{p×(r1−1)+2}. Thereafter, similar average values AV_{p×(r1−1)+3} to AV_p×r1 are assigned as the decoded pixel values DXV_{p×(r1−1)+2} to DXV_p×r1. Instead of the pixel value XV_p×(r−1)+1 and the decoded pixel values DXV_p×(r−1)+2 to DXV_p×r which were never obtained, the missing block recovery section 9 outputs the average values AV_{p×(r1−1)+1} to AV_p×r1 which have virtually been reproduced in the above manner to the image data reproduction section 9. Based on the above-described average values AV_{p×(r1−1)+1} to AV_p×r1 as well as the pixel value XV_p×(r−1)+1 and the decoded pixel values DXV_p×(r−1)+2 to DXV_p×r which have been generated from all the data packets DP_r except for the missing data packet DDP_r1, the image data reproduction section 9 generates reproduced image data RTD₂ which is similar to the aforementioned reproduced image data RTD₁.

The reproduced image data RTD₂ and the aforementioned reproduced image data RTD₁ are substantially identical (i.e., indistinguishable to the human eye) for the following reasons. The transmitter Tx performs DPCM encoding and near-instantaneous compression on the basis of the data blocks DB_r, so that the missing data packet DDP_r does not exert any influence on the near-instantaneous decompression and DPCM decoding performed for the other data packets DP_r. Furthermore, each data block DB_r only includes p pixel values XV_p×(r−1)+1 to DXV_p×r. Therefore, the reproduced image data RTD₂ containing the aforementioned average value AV_{p×(r1−1)+1} to AV_p×r1 would hardly present any difference to the human eye from the reproduced image data RTD₁.

As can be seen from the above, in accordance with the receiver Rxb, any missing data packets DDP_r can be approximately recovered based on the other data packets DP_r, thereby solving the aforementioned problem and generating reproduced image data RTD₂ which is visually acceptable.

Alternatively, the missing block recovery section 9 may utilize, in order to deal with the missing data packet DDP_r, the pixel value XV_p×(r−1)+1 and a set of decoded pixel values DXV_p×(r−1)+2 to DXV_p×r from one line above or below (along the longitudinal direction VD) the missing data packet DDP_r. Further alternatively, the missing block recovery section 9 may utilize the pixel value XV_p×(r−1)+1 and a set of decoded pixel values DXV_p×(r−1)+2 to DXV_p×r from one pixel value right or left (along the width direction HD) of the missing data packet DDP_r. In particular, in the case where moving pictures are represented by the image data TD, the pixel value XV_p×(r−1)+1 and a set of decoded pixel values DXV_p×(r−1)+2 to DXV_p×r which are generated from a data packet DP_r pertaining to a preceding and/or succeeding frame may be employed in the receiver Rxb.

Next, with reference to FIG. 26, a second implementation of the data compression section 2 shown in FIG. 1 will be described. In the following description, the second implementation of the data compression section 2 will be referred to as a “data compression section 2c”. In order to perform the aforementioned second compression process, the data compression section 2c includes an orthogonal transform section 26 and a data reduction section 27, as shown in FIG. 26.

Next, the second compression process which is performed by the data compression section 2c will be specifically described. Data blocks DB_r from the aforementioned blocking section 1 (see FIG. 1) are received by an orthogonal transform section 26 in the data compression section 2c. The orthogonal transform section 26 performs an orthogonal transform to multiply the received set of pixel values XV_p×(r−1)+1 to XV_p×r by a predetermined orthogonal transform matrix, thereby generating a set of coefficients CF_p×(r−1)+1 to CF_p×r. As shown in FIG. 27, these coefficients CF_p×(r−1)+1 to CF_p×r which are derived through the orthogonal transform are expressed in n bits, as are the pixel values XV_p×(r−1)+1 to XV_p×r. Furthermore, the coefficients CF_p×(r−1)+1 to CF_p×r represent respectively different frequency components in the frequency domain. The above set of coefficients CF_p×(r−1)+1 to CF_p×r are outputted from the orthogonal transform section 26 to the data reduction section 27.

Hereinafter, an Hadamard transform will be described as an example of the aforementioned orthogonal transform, and the process performed by the orthogonal transform section 26 will be more specifically described. As described earlier, each data block DB_r includes p pixel values XV_p×(r−1)+1 to XV_p×r. In the following description, p is conveniently assumed to be 16. Under this assumption, the orthogonal transform section 26 retains a (16×16) Hadamard transform matrix H as expressed by eq. 1 below:

$\begin{array}{cc}H=\frac{1}{\sqrt{16}}\hspace{1em}\left[\begin{array}{cccccccccccccccc}1& 1& 1& 1& 1& 1& 1& 1& 1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1& -1& -1& -1& -1& -1& -1& -1& -1\\ 1& 1& 1& 1& -1& -1& -1& -1& -1& -1& -1& -1& 1& 1& 1& 1\\ 1& 1& 1& 1& -1& -1& -1& -1& 1& 1& 1& 1& -1& -1& -1& -1\\ 1& 1& -1& -1& -1& -1& 1& 1& 1& 1& -1& -1& -1& -1& 1& 1\\ 1& 1& -1& -1& -1& -1& 1& 1& -1& -1& 1& 1& 1& 1& -1& -1\\ 1& 1& -1& -1& 1& 1& -1& -1& -1& -1& 1& 1& -1& -1& 1& 1\\ 1& 1& -1& -1& 1& 1& -1& -1& 1& 1& -1& -1& 1& 1& -1& -1\\ 1& -1& -1& 1& 1& -1& -1& 1& 1& -1& -1& 1& 1& -1& -1& 1\\ 1& -1& -1& 1& 1& -1& -1& 1& -1& 1& 1& -1& -1& 1& 1& -1\\ 1& -1& -1& 1& -1& 1& 1& -1& -1& 1& 1& -1& 1& -1& -1& 1\\ 1& -1& -1& 1& -1& 1& 1& -1& 1& -1& -1& 1& -1& 1& 1& -1\\ 1& -1& 1& -1& -1& 1& -1& 1& 1& -1& 1& -1& -1& 1& -1& 1\\ 1& -1& 1& -1& -1& 1& -1& 1& -1& 1& -1& 1& 1& -1& 1& -1\\ 1& -1& 1& -1& 1& -1& 1& -1& -1& 1& -1& 1& -1& 1& -1& 1\\ 1& -1& 1& -1& 1& -1& 1& -1& 1& -1& 1& -1& 1& -1& 1& -1\end{array}\right]& \mathrm{eq}.1\end{array}$

For conciseness, the pixel values XV_16×r−15 to XV_16×r contained in the currently-received data block DB_r are represented as a matrix expressed by eq. 2 below. The coefficients CF_16×r−15 to CF_16×r obtained through the Hadamard transform are represented as a matrix expressed by eq. 3 below.x=[pixel value XV_16×r−15, pixel value XV_16×r−14, . . . , pixel value XV_16×r]^t  eq. 2y=[coefficient CF_16×r−15, coefficient CF_16×r−14, . . . , coefficient CF_16×r]^t  eq. 3

In eq. 2 and eq. 3, “t” means transpose.

Under the definitions expressed by eq. 1 to eq. 3 above, the orthogonal transform section 26 multiplies the Hadamard transform matrix H by the matrix “x” beginning from the right side thereof, as expressed by eq. 4 below.y=H×x/4  eq. 4

The resultant coefficient CF_16×r−15 is an integer in the range from 0 to 255, and the resultant coefficients CF_16×r−14 to CF_16×r are integers in the range from −127 to 127. Therefore, these coefficients can all be expressed in n bits as mentioned above. According to the Hadamard transform matrix H expressed by eq. 1 above, the coefficient CF_16×r−15 represents a component associated with the lowest frequency region. Likewise, the coefficients CF_16×r−15 to CF_16×r having greater suffix values represent components which are associated with respectively higher frequency regions.

As described earlier, the data reduction section 27 sequentially receives sets of coefficients CF_p×(r−1)+1 to CF_p×r. From each received set of coefficients CF_p×(r−1)+1 to CF_p×r, the data reduction section 27 deletes b coefficients CF_p×r−b to CF_p×r which represent components associated with predetermined high-frequency regions, thereby generating compressed coefficients CF_p×(r−1)+1 to CCF_{p×r−(b−1)}. Herein, “b” is a natural number in the range from 1 to n.

According to the present embodiment of the invention, as shown in FIG. 28, those of the coefficients CF_p×(r−1)+1 to CF_p×r which are to be deleted are predetermined. FIG. 28 shows example coefficients CF which are to be deleted in an exemplary case where n=8 and p=16. In principle, among the currently-received coefficients CF_16×r−15 to CF_16×r, the data reduction section 27 leaves intact those which represent predetermined low-frequency components, while deleting those which represent any higher frequency components. In the exemplary case shown in FIG. 28, the coefficients CF_16×r−15 to CF_16×r−12, which represent relatively low-frequency components, are left intact (see the hatched portions). On the other hand, the coefficients CF_16×r−7 to CF_16×r representing relatively high frequency components are deleted. The reason why all of the coefficients CF_16×r−7 to CF_16×r representing relatively high-frequency components can be deleted is as follows. In a set of coefficients (i.e., matrix y (as represented by eq. 3)) which are produced through the aforementioned orthogonal transform, the greater coefficients are generally concentrated in the low frequency regions, whereas the smaller coefficients are generally concentrated in the high frequency regions, due to certain correlations residing in the image MG.

According to the present embodiment of the invention, the next bit to the MSB and the least significant bit (hereinafter referred to as “LSB”) are deleted from both coefficients CF_16×r−11 and CF_16×r−10. Furthermore, the next bit to the MSB and the lower two bits are deleted from both coefficients CF_16×r−9 and CF_16×r−8. More specifically, as shown in FIG. 29, the LSBs in the coefficients CF_16×r−11 to CF_16×r−8 are all deleted. If any of the coefficients CF_16×r−11 to CF CF_16×r−8 has a positive value and the next bit to the MSB that is to be deleted is “1”, then the other bits are all set to “1”, so the resultant compressed coefficient CCF_16×r−11 to CCF_16×r−8 will be a bit sequence composed only of “1”. If any of the coefficients CF_16×r−11 to CF_16×r−8 has a positive value and the next bit to the MSB that is to be deleted is “0”, then all the other bits are left intact. If any of the coefficients CF_16×r−11 to CF_16×r−8 has a negative value and the next bit to the MSB that is to be deleted is “1”, then all the other bits are set to “0”. If any of the coefficients CF_16×r−11 to CF_16×r−8 has a negative value and the next bit to the MSB that is to be deleted is “0”, then all the other bits are left intact.

The present embodiment is not limited to the exemplary cases shown in FIGS. 28 and FIG. 29. It is possible to determine which one of the coefficients CF_p×(r−1)+1 to CF_p×r are to be deleted and which bit(s) thereof is to be deleted in accordance with the band width of the transmission path N, the image quality which is required on the receiver Rx side, and the like.

In the above-described manner, the data reduction section 27 generates a set of compressed coefficients CCF_p×(r−1)+1 to CCF_{p×r−(b−1)}, which are outputted to the data sending section 3 as the aforementioned compressed block CB_r.

As shown in FIG. 1, the data sending section 3 includes the buffer section 31 and the sending control section 32. The buffer section 31 stores the set of compressed coefficients CCF_p×(r−1)+1 to CCF_{p×r−(b−1)} as a fixed-length compressed block CB_r. As in the first implementation, since the set of compressed coefficients CCF_p×(r−1)+1 to CCF_{p×r−(b−1)} is of a fixed length, any delay time incurred will be minimized. The sending control section 32 receives the set of compressed coefficients CCF_p×(r−1)+1 to CCF_{p×r−(b−1)} from the buffer section 31, and sends these onto the transmission path N. As shown in FIG. 1, the set of compressed coefficients CCF_p×(r−1)+1 to CCF_{p×r−(b−1)}, as an example compressed block CB_r, are transmitted via the transmission path N so as to be received by the receiver RX.

The receiver Rx subjects the received data packet DP_r to predetermined processing to reproduce the image data TD. Hereinafter, a second implementation of the receiver Rx of FIG. 1 will be described with reference to FIG. 30. In the following description, the second implementation of the receiver Rx will be referred to as a “receiver Rxc”. The receiver Rxc differs from the above-described receiver Rxa (see FIG. 15) in that the receiver Rxc includes a bit decoding section 10 and an inverse orthogonal transform section 11 instead of the packet deassembling section 6 and the decompression/decoding section 7. Since the receiver Rxc is otherwise identical to the receiver Rxa, any component elements in FIG. 30 which find their counterparts in FIG. 15 are denoted by the same reference numerals as those employed in connection with the receiver Rxa, and the descriptions thereof are omitted.

Next, the reproduction processing for the image data TD which is performed by the receiver Rxc having the aforementioned structure will be described in detail. In the data receiving section 5, the buffer section 51 stores a set of compressed coefficients CCF_p×(r−1)+1 to CCF_{p×r−(b−1)} from the transmission path N. The buffer section 51, which is only required to store the set of fixed-length compressed coefficients CCF_p×(r−1)+1 to CCF_{p×r−(b−1)}, as is the case with the buffer section 31, contributes to the minimization of the delay time. After the buffering, the compressed coefficients CCF_p×(r−1)+1 to CCF_{p×r−(b−1)} are outputted to the bit decoding section 10 via the reception control section 52.

The bit decoding section 10 performs an inverse process of the process which is performed by the data reduction section 27 so as to generate the decompressed coefficients DCF_p×(r×1)+1 to DCF_p×r from the compressed coefficients CCF_p×(r−1)+1 to CCF_{p×r−(b−1)}, which are then outputted to the inverse orthogonal transform section 11. Herein, the decompressed coefficients DCF_p×(r−1)+1 to DCF_p×r have such small differences from the coefficients CF_p×(r−1)+1 to CF_p×r that the decoded pixel values DXV_p×(r−1)+1 to DXV_p×r (described later) and the pixel values XV_p×(r−1)+1 to XV_p×r would hardly present any difference to the human eye. To describe the bit restoration process more specifically, an average value of a bit sequence of 1 and/or 0, i.e., values which can be expressed in two bits, is added after the LSB of each of the compressed coefficients CCF_16×r−9 and CCF_16×r−8. Furthermore, in the case where the MSB in the compressed coefficient CCF_16×r−9 and/or CCF_16×r−8 has a positive value, “0” is inserted after the MSB of that compressed coefficient; otherwise, “1” is inserted after the MSB of that compressed coefficient. Thus, the decompressed coefficients DCF_16×r−9 and DCF_16×r−8 are generated. Moreover, a bit “1” or “0” is added after the LSB of each of the compressed coefficients CCF_16×r−11 and CCF_16×r−10. Furthermore, in the case where the MSB in the compressed coefficient CCF_16×r−11 and/or CCF_16×r−10 has a positive value, “0” is inserted after the MSB of that compressed coefficient; otherwise, “1” is inserted after the MSB of that compressed coefficient. Thus, the decompressed coefficients DCF_16×r−11 and DCF_16×r−10 are generated. In order to deal with the coefficients CF_16×r−7 and DCF_16×r from which all bits have been deleted, the bit decoding section 10 generates decompressed coefficients DCF_16×r−7 and DCF_16×r whose eight bits are all “0”.

The inverse orthogonal transform section 11 performs an inverse orthogonal transform, i.e., an inverse process of the process which is performed by the orthogonal transform section 26, to multiply an inverse matrix of the aforementioned orthogonal transform matrix by the received decompressed coefficients DCF_p×(r−1)+1 to DCF_p×r, thereby generating a set of decoded pixel values DXV_p×(r−1)+1 to DXV_p×r. The resultant decoded pixel values DXV_p×(r−1)+to DXV_p×r, which hardly present any difference to the human eye as compared to the pixel values XV_p×(r−l)+1 to XV_p×r, are outputted to the image data reproduction section 8.

As described earlier, the present embodiment of the invention is directed to the case where an Hadamard transform is performed. Next, the processing to be performed by the inverse orthogonal transform section 11 in this specific case will be described more specifically. The following description also assumes that p=16. Under this assumption, the inverse orthogonal transform section 11 retains an inverse transform matrix H⁻¹ of eq. 1 above. For conciseness, the currently-received decompressed coefficients DCF_p×(r−1)+1 to DCF_p×r are represented as a matrix expressed by eq. 5 below. The decoded pixel values DXV_p×(r−1)+1 to DXV_p×r which are obtained through the inverse transform of the Hadamard transform are represented as a matrix expressed by eq. 6 below.y=[decompressed coefficient CF_16×r−15, . . . , decompressed coefficient CF_16×r]^t  eq. 5z=[decoded pixel value XV_16×r−15, . . . , decoded pixel value XV_16×r]^t  eq. 6In eq. 5 and eq. 6, “t” means transpose.

Under the definitions expressed by eq. 5 and eq. 6 above, the inverse orthogonal transform section 11 multiplies the matrix “y” by the inverse transform matrix H⁻¹ of the Hadamard transform matrix H beginning from the right side thereof, as expressed by eq. 7 below.z=H⁻¹×y×4  eq. 7

As a result of the above processing, the image data reproduction section 8 sequentially receives q sets of decoded pixel values DXV_p×(r−1)+1 to DXV_p×r. The image data reproduction section 8 generates a reproduced image data RTD similar to that shown in FIG. 20.

As described above, in accordance with the second implementation of the data compression section 2c, too, encoding and compression are performed for the fixed-length data block DB_r composed of p pixel values XV_p×(r−1)+1 to XV_p×r arranged in line along the width direction HD, it is possible to minimize any delay time elapsing prior to the generation of the reproduced image data RTD in the receiver Rxc.

Although the above illustration is directed to the case where an Hadamard transform is performed in the orthogonal transform section 26, the present invention is not limited thereto. Alternatively, the orthogonal transform section 26 may perform a DCT or discrete sine transform (DST).

“First Application”

In recent years, there has been plenty of work directed to the research and development of driving assistant systems for assisting a driver in his/her driving of a vehicle by capturing an image of the surroundings of the vehicle via image capturing devices and providing such an image to the driver. Next, a driving assistant system TS₁ incorporating the above-described transmitter Tx and receiver Rx will be described. FIG. 31 is a block diagram illustrating the overall configuration of the driving assistant system TS₁. The driving assistant system TS₁ shown in FIG. 31 includes two image capturing devices 13, two image processing sections 14, two transmitters Tx, a transmission path N, a receiver Rx, an image synthesis section 15, and a display section 16. The driving assistant system TS₁ is mounted in a vehicle V_ur.

The image capturing devices 13, each of which is disposed so as to be able to capture an image of an area in the rear of the vehicle V_ur, capture images of respectively different regions in the rear of the vehicle V_ur, and generate captured image data CTD representing the captured images MG (see FIG. 32). After each image capturing device 13, an implementation of the aforementioned image processing section 14 is coupled so as to receive the respective captured image data CTD from that image capturing device 13.

As shown in FIG. 32, from the received image data CTD, each image processing section 14 selects a number of pixels composing a partial image PMG representing a predetermined portion of the image indicative of the surroundings of the vehicle, thereby generating a partial image data PTD (this process being referred to as “clipping”). Herein, it is conveniently assumed that each partial image data PTD is of the format shown in FIG. 3 in the present embodiment of the invention. After each image processing section 14, an implementation of the aforementioned transmitter Tx is coupled so as to receive the respective image data PTD from that image processing section 14.

Each transmitter Tx performs the processing described in any of the earlier embodiments of the invention for the partial image data PTD it receives, thereby generating compressed block CB_r. The respective compressed blocks CB_r are transmitted to the receiver Rx via the transmission path N.

The receiver Rx performs the processing described in any of the earlier embodiments of the invention for each received compressed block CB_r, thereby generating reproduced partial image data RPTD. Herein, as will be appreciated from the foregoing description, each reproduced partial data RPTD represents a reproduced partial image RMG which is substantially the same as the image MG represented by each partial image data PTD (as shown in FIG. 32). The respective reproduced partial data RPTD are outputted to the subsequent image synthesis section 15.

The image synthesis section 15 performs a synthesis process for both received reproduced partial image data RPTD to generate a merged image data MTD representing a single synthesized image MMG which is composed of the two partial images PMG. The merged image data MTD is outputted to the display section 16.

The display section 16 subjects the received merged image data MTD to display processing, thereby providing the aforementioned synthesized image MMG to the driver of the vehicle V_ur.

Thus, in accordance with the present driving assistant system TS₁, each image processing section 14 generates a partial image data PTD which is required on the receiver Rx side, so that the amount of data which is transmitted over the transmission path N can be minimized.

Moreover, in accordance with the transmitters Tx and the receiver Rx, as described earlier, the delay time which is incurred after the generation of captured image data CTD by the image capturing devices 13 and before the display processing of the merged image data MTD by the display section 16 can be minimized. By incorporating the transmitters Tx and the receiver Rx having such characteristics in the driving assistant system TS₁, a driver is enabled to grasp the surroundings of the vehicle V_ur in real time. As a result, the driver can drive the vehicle V_ur with increased safety.

Next, with reference to FIG. 33, the above-described technological effects will be more specifically described. In FIG. 33, the vehicle V_ur is moving “in reverse” (i.e., backing up) in a direction shown by arrow A₁. An obstacle BST is present at a distance from the vehicle V_ur in the direction (indicated by arrow A₁) in which the vehicle V_ur is moving. The driver will try to drive the vehicle V_ur so as not to collide with the obstacle BST by checking the synthesized image MTD which is displayed on the display section 16. In FIG. 33, a point P₀ is a position where the rear end of the vehicle V_ur passes at the current time T₀. Given that the aforementioned delay time is DT, the synthesized image MTD which is displayed on the display section 16 at the current time T₀ is an image generated based on the captured image data CTD which was generated by the image capturing devices 13 at a time (T₀−t).

In FIG. 33, a point P₁ is a position where the rear end of the vehicle V_ur passed at the aforementioned time (T₀−t), i.e., where the captured image data CTD was generated. Herein, the distance Δd between the points P₀ and P₁ is determined by the velocity SP of the vehicle V_ur and the delay time DT, as expressed by eq. 8 below.Δd=SP×DT  eq. 8

Assuming that the vehicle V_ur moves at a constant velocity, as seen from eq. 8 above, the distance Δd increases as the delay time DT increases. If the delay time DT=0, then the distance Δd=0, so that the points P₀ and P₁ will coincide. If there is substantial delay time DT, on the other hand, the vehicle V_ur may collide into the obstacle BST before the display section 16 displays the sight of the collision in vain. From this perspective, it can be seen how useful it is to incorporate transmitters Tx and a receiver Rx having a sufficiently small delay time DT in the driving assistant system TS₁.

In the above application, the driving assistant system TS₁ comprises two sets of image capturing devices 13, image processing sections 14, and image transmitters Tx. However, the present invention is not limited to such a configuration. The driving assistant system TS₁ may comprise one or more set of such elements. Although the above image capturing devices 13 are illustrated as being capable of capturing images of objects at the rear of the vehicle V_ur, the image capturing devices 13 may alternatively be fixed on the vehicle V_ur so as to be capable of capturing images of objects at the front and/or sides of the vehicle V_ur, as necessary.

While the missing block recovery section 9 in the earlier-described embodiment is illustrated as approximately recovering the pixel values XV_p×(r−1)+1 to DXV_p×r which have been contained in a missing data packet DDP_r from a correctly-generated set of the pixel value XV_p×(r−1)+1 and the decoded pixel values DXV_{p×(r−1 )+2} to DXV_p×r, it may be more preferable in the driving assistant system TS₁ not to reproduce any missing data packets DDP_r, but simply “black out” that portion, for example, thereby warning the driver of the occurrence of the missing data packets DDP_r, because the driver is in need of accurate information concerning the surroundings of the vehicle V_ur.

“Second Application”

In the field of FA (Factory Automation), there has been some research and development work on remote control systems for acting on an object via remote control. Next, a remote control system TS₂ incorporating the above-described transmitter Tx and the receiver Rx will be described. FIG. 34 is a block diagram illustrating the overall structure of the remote control system TS₂. The remote control system TS₂ of FIG. 34 includes an image capturing device 17, a transmitter Tx, a transmission path N, a receiver Rx, a display section 18, a control section 19, a control data generation section 110, a control data sending section 111, a control data receiving section 112, and a manipulator section 113. The remote control system TS₂ is employed to exert an action on an object TG.

The image capturing device 17, which is disposed in the neighborhood of the object TG, captures an image of the object TG, and generates captured image data CTD representing the captured image. The generated captured image data CTD is outputted to the transmitter Tx. The transmitter Tx subjects the received captured image data CTD to the processing which has been described in any of the foregoing embodiments of the invention, thereby generating compressed block CB_r as described above. The compressed block CB_r is transmitted to the receiver Rx via the transmission path N.

The receiver Rx subjects the received compressed block CB_r to the processing which has been described in the foregoing embodiment of the invention, thereby generating reproduced image data RTD. Herein, as will be appreciated from the foregoing description, the reproduced image data RTD represents an image of the object TG which is substantially the same as the image represented by the captured image data CTD. The reproduced image data RTD is outputted to the display section 18. The display section 18 subjects the received reproduced image data RTD to display processing, thereby providing an image representing the object TG to an operator.

While checking on the display section 18, the operator operates the control section 19 to instruct as to what sort of action to exert on the object TG. In response to the instruction from the control section 19, the control data generation section 110 generates control data CTLD representing an action to be exerted on the object, which is outputted to the control data sending section 111. The control data sending section 111 sends the received control data CTLD onto the transmission path N, which then is transmitted to the control data receiving section 112. The manipulator section 113 exerts an action on the object TG in accordance with the control data CTLD which is received by the control data receiving section 112.

As described above, in accordance with the transmitter Tx and the receiver Rx, the delay time which is incurred after the generation of captured image data CTD by the image capturing device 17 and before the display processing by the display section 18 can be minimized. By incorporating the transmitters Tx and the receiver Rx having such characteristics in the remote control system TS₂, an operator is enabled to grasp the situation concerning the object TG in real time. As a result, the operator can properly act on the object TG from a remote place.

Although the above-described application assumes that the operator operates the control section 19 while looking at the display section 18, it is also possible to automatically control the operation of the robot (or the manipulator section 113) by applying image recognition techniques.

While the invention has been described in detail, the foregoing description is in all aspects illustrative and not restrictive. It is understood that numerous other modifications and variations can be devised without departing from the scope of the invention.

Claims

1. An image transmitter for compressing image data and transmitting the image data to a receiver via a transmission path, wherein the image data at least contains i pixel values of pixels arranged in line along a single direction, each pixel value being expressed in n bits, the image transmitter comprising:a blocking section for taking every p pixel values among the i pixel values in the image data to form a data block, and sequentially outputting a plurality of data blocks each including p pixel values;a data compression section for reducing an amount of data from each data block outputted from the blocking section to output a compressed block for each data block, the data compression section compressing each data block based on a level value derived from a differential pulse code modulation result of two adjoining pixel values; anda data sending section for sending each compressed block outputted from the data compression section onto the transmission path,wherein i, n, and p are predetermined natural numbers, and the data compression section deletes a bit, among all bits of a data block, from each data block based on the level value,wherein the data compression section comprises: a DPCM encoding section for performing a differential pulse code modulation based on two adjoining pixel values in each data block generated by the blocking section to generate and output differential data;a near-instantaneous compression section for performing a near-instantaneous compression for the differential data outputted from the DPCM encoding section to reduce a number of bits in the differential data, and generating and outputting compressed differential data; anda packet assembling section for generating and outputting a data packet to the data sending section as the compressed block, the data packet containing a first pixel value in the data block generated by the blocking section and the respective compressed differential data outputted from the near-instantaneous compression section,wherein the DPCM encoding section comprises: a delay section for applying a predetermined delay amount to 1st to (p−1)th pixel values in each data block generated by the blocking section, and outputting delayed pixel values; anda subtraction section for calculating respective differences between 2 to p pixel values in each data block generated by the blocking section and 1st to (p−1)th delayed pixel values outputted from the delay section to generate (p−1) differential data as the differential data, and outputting the (p−1) differential data to the near-instantaneous compression section, andwherein the near-instantaneous compression section comprises: a buffer section for storing the (p−1) differential data outputted from the DPCM encoding section;a level determination section for generating the level value based on each of the (p−1) differential data outputted from the buffer section and outputting the level value, the level value determining a bit to be deleted from each of the (p−1) differential data; anda data reduction section for deleting the bit, as determined by the level value outputted from the level determination section, from each differential data outputted from the buffer section to generate (p−1) compressed differential data as the compressed differential data, and outputting the (p−1) compressed differential data to the packet assembling section.

2. The image transmitter according to claim 1, wherein: the compressed block has a fixed length; andthe data sending section comprises: a buffer section for storing the compressed block; anda sending control section for sending the compressed block stored in the buffer section onto the transmission path.

3. The image transmitter according to claim 1, wherein the level determination section comprises: a differential data selection section for selecting one of the respective (p−1) differential data outputted from the buffer section which has a greatest absolute value, and outputting the selected differential data as a maximum differential data; anda level selection section for selecting one of a predetermined number of the level values based on the maximum differential data outputted from the differential data selection section, and outputting the selected level value to the data reduction section.

4. The image transmitter according to claim 3, wherein: the data reduction section deletes t bits, as determined by the level value outputted from the level determination section, from each differential data outputted from the buffer section to generate the (p−1) compressed differential data; andt is a natural number.

5. The image transmitter according to claim 1, wherein the data packet generated by the packet assembling section contains the first pixel value in each data block generated by the blocking section, the level value outputted from the level determination section, and the (p−1) compressed differential data outputted from the data reduction section.

6. The image transmitter according to claim 1, wherein, when p pixel values among the i pixel values in the image data cannot be taken to form a data block, the blocking section generates a padded data block by using a predetermined padding bit, the padded data block having the same size as that of the data block.

7. A transmission system including the receiver and the image transmitter according to claim 1, wherein the receiver comprises: a data receiving section for receiving the data packet transmitted via the transmission path and outputting the received data packet;a packet disassembling section for disassembling the data packet outputted from the data receiving section into the first pixel value in the data block and the respective compressed differential data generated by the near-instantaneous compression section, and outputting the first pixel value in the data block and the respective compressed differential data;a near-instantaneous decompression section for restoring the bits having been deleted from the compressed differential data generated by the near-instantaneous compression section to generate and output decompressed differential data;a DPCM decoding section for performing an inverse process of the differential pulse code modulation performed by the DPCM encoding section, generating decoded pixel values from the respective decompressed differential data outputted from the near-instantaneous decompression section, and outputting the decoded pixel values; andan image data reproduction section for generating reproduced image data from the first pixel value outputted from the packet disassembling section and the respective decoded pixel values outputted from the DPCM decoding section, the reproduced image data representing an image.

8. The transmission system according to claim 7, wherein: the data packet has a fixed length; andthe data receiving section comprises: a buffer section capable of storing at least the data packet of the fixed length; anda reception control section for outputting the data packet stored in the buffer section to the packet disassembling section.

9. The transmission system according to claim 7, wherein the receiver further comprises: a missing block recovery section for, when a data packet is missed by the data receiving section, reproducing pixel values contained in the missing data packet based on a first pixel value from at least one other data packet from the packet disassembling section and decoded pixel values from the at least one other data packet from the DPCM decoding section, and outputting the pixel values to the image data reproduction section.

10. A driving assistant system for assisting in the driving of a vehicle, the driving assistant system comprising: a transmission path;a plurality of image capturing devices each of which is fixed on the vehicle and which captures an image of surroundings of the vehicle and outputs captured image data;a plurality of image processing sections each of which is connected to an image capturing device of the image capturing devices and performs a clipping process for the captured image data outputted from the respective image capturing device to generate partial image data, wherein each partial image data at least contains i pixel values of pixels arranged in line along a single direction, each pixel value being expressed in n bits; anda plurality of transmitters each of which is associated with one of the plurality of image processing sections,wherein each transmitter comprises: a blocking section for taking every p pixel values among the i pixel values in the partial image data from the associated image processing section to form a data block, and sequentially outputting a plurality of data blocks each including p pixel values,a data compression section for reducing an amount of data from each data block outputted from the blocking section to output a compressed block for each data block, the data compression section reducing the amount of data based on a level value derived from a different pulse code modulation result of two adjoining pixel values, anda data sending section for sending each compressed block outputted from the data compression section onto the transmission path;a receiver for receiving and decompressing each compressed block from the transmission path, restoring each data block, and reproducing the respective partial image data;an image synthesis section for performing a synthesis process for the partial image data from the receiver and outputting merged image data, the merged image data representing an image in which images represented by the respective partial image data are synthesized; anda display section for displaying the image represented by the merged image data outputted from the image synthesis section,wherein i, n, and p are predetermined natural numbers, and the data compression section deletes a bit, among all bits of a data block, from each data block based on the level value,wherein the data compression section of each transmitter comprises: a DPCM encoding section for performing a differential pulse code modulation based on two adjoining pixel values in each data block generated by the blocking section to generate and output differential data;a near-instantaneous compression section for performing a near-instantaneous compression for the differential data outputted from the DPCM encoding section to reduce a number of bits in the differential data, and generating and outputting compressed differential data; anda packet assembling section for generating and outputting a data packet to the data sending section as the compressed block, the data packet containing a first pixel value in the data block generated by the blocking section and the respective compressed differential data outputted from the near-instantaneous compression section,wherein the DPCM encoding section of each transmitter comprises: a delay section for applying a predetermined delay amount to 1st to (p−1)th pixel values in each data block generated by the blocking section, and outputting delayed pixel values; anda subtraction section for calculating respective differences between 2nd to pth pixel values in each data block generated by the blocking section and 1st to (p−1)th delayed pixel values outputted from the delay section to generate (p−1) differential data as the differential data, and outputting the (p−1) differential data to the near-instantaneous compression section, andwherein the near-instantaneous compression section of each transmitter comprises: a buffer section for storing the (p−1) differential data outputted from the DPCM encoding section;a level determination section for generating the level value based on each of the (p−1) differential data outputted from the buffer section and outputting the level value, the level value determining a bit to be deleted from each of the (p−1) differential data; anda data reduction section for deleting the bit, as detenriined by the level value outputted from the level determination section, from each differential data outputted from the buffer section to generate (p−1) compressed differential data as the compressed differential data, and outputting the (p−1) compressed differential data to the packet assembling section.

11. A remote control system for exerting an action on an object via remote control, the remote control system comprising: a transmission path;an image capturing device which is provided in a neighborhood of the object and captures an image of surroundings of the object and outputs captured image data, wherein the captured image data at least contains i pixel values of pixels arranged in line along a single direction, each pixel value being expressed in n bits; anda transmitter,wherein the transmitter comprises: a blocking section for taking every p pixel values among the i pixel values in the captured image data from the image capturing device to form a data block, and sequentially outputting a plurality of data blocks each including p pixel values;a data compression section for reducing an amount of data from each data block outputted from the blocking section to output a compressed block for each data block, the data compression section reducing the amount of data based on a level value derived from a differential pulse code modulation result of two adjoining pixel values, anda data sending section for sending each compressed block outputted from the data compression section onto the transmission path;a receiver for receiving and decompressing each compressed block from the transmission path, restoring each data block, and reproducing the captured image data;a display section for displaying the image represented by the captured image data outputted from the receiver for viewing by an operator;a control data generation section for generating and outputting control data for exerting an action on the object in accordance with control made by the operator; anda control data sending section for sending control data from the control data generation section onto the transmission path;a control data receiving section for receiving and outputting the control data from the transmission path; anda manipulator section for exerting the action on the object in accordance with the control data received from the control data receiving section,wherein i, n, and p are predetermined natural numbers, and the data compression section deletes a bit, among all bits of a data block, from each data block based on the level value,wherein the data compression section comprises: a DPCM encoding section for performing a differential pulse code modtilation based on two adjoining pixel values in each data block generated by the blocking section to generate and output differential data;a near-instantaneous compression section for performing a near-instantaneous compression for the differential data outputted from the DPCM encoding section to reduce a number of bits in the differential data, and generating and outputting compressed differential data; anda packet assembling section for generating and outputting a data packet to the data sending section as the compressed block, the data packet containing a first pixel value in the data block generated by the blocking section and the respective compressed differential data outputted from the near-instantaneous compression section,wherein the DPCM. encoding section comprises: a delay section for applying a predetermined delay amount 1st to (p−1)th pixel values in each data block generated by the blocking section, and outputting delayed pixel values; anda subtraction section for calculating respective differences between 2nd to pth pixel values in each data block generated by the blocking section and 1st to (p−1)th delayed pixel values outputted from the delay section to generate (p−1) differential data as the differential data, and outputting the (p−1) differential data to the near-instantaneous compression section, andwherein the near-instantaneous compression section comprises: a buffer section for storing the (p−1) differential data outputted from the DPCM encoding section;a level determination section for generating the level value based on each of the (p−1) differential data outputted from the buffer section and outputting the level value, the level value determining a bit to be deleted from each of the (p−1) differential data; anda data reduction section for deleting the bit, as detennined by the level value outputted from the level determination section, from each differential data otitputted from the buffer section to generate (p−1) compressed differential data as the compressed differential data, and outputling the (p−1) compressed differential data to the packet assembling section.

12. An image compression and transmission method for compressing image data and transmitting the image data to a receiver via a transmission path, wherein the image data at least contains i pixel values of pixels arranged in line along a single direction, each pixel value being expressed in n bits, the image compression and transmission method comprising:a blocking operation of taking every p pixel values among the i pixel values in the image data to form a data block, and sequentially outputting a plurality of data blocks each including p pixel values;a data compression operation of reducing an amount of data from each data block outputted in the blocking operation to output a compressed block for each data block based on a level value derived from a differential pulse code modulation result of two adjoining pixel values; andsending each compressed block outputted in the data compression operation onto the transmission path,wherein i, n, and p are predetermined natural numbers, and the data compression operation deletes a bit, among all bits of a data block, from each data block based on the level value,wherein the data compression operation comprises: a DPCM encoding operation of performing a differential pulse code modulation based on two adjoining pixel values in each data block generated by the blocking operation to generate and output differential data;a near-instantaneous compression operation of performing a near-instantaneous compression for the differential data outputted by the DPCM encoding operation to reduce a number of bits in the differential data, and generating and outputting compressed differential data; anda packet assembling operation of generating and outputting a data packet for the sending operation as the compressed block, the data packet containing a first pixel value in the data block generated by the blocking operation and the respective compressed differential data outputted by the near-instantaneous compression operation,wherein the DPCM encoding operation comprises: a delay operation of applying a predetermined delay amount to 1st to (p−1)th pixel values in each data block generated by the blocking operation, and outputting delayed pixel values; anda subtraction operation of calculating respective differences between 2nd to pth pixel values in each data block generated by the blocking operation and 1st to (p−1)th delayed pixel values outputted by the delay operation to generate (p−1) differential data as the differential data, and outputting the (p−1) differential data for the near-instantaneous compression operation, andwherein the near-instantaneous compression operation comprises: a buffer operation of storing the (p−1) differential data outputted by the DPCM encoding operation;a level determination operation of generating the level value based on each of the (p−1) differential data outputted by the buffer operation and outputting the level value, the level value determining a bit to be deleted from each of the (p−1) differential data; anda data reduction operation of deleting the bit, as determined by the level value outpuffed by the level determination operation from each differential data outputted by the buffer operation to generate (p−1) compressed differential data as the compressed differential data, and outputting the (p−1) compressed differential data for the packet assembling operation.