Patent Grant
7895251
The present invention relates to a technique for generating a random number.
As one of random number generating methods using logical operations, an L-bit length random number method called an M-sequence is known. This method uses an equation given by xN=xN−K+xN−L, and generates a 1-bit random number using only numerals 0 and 1. Since this random number has a logically maximum length for specific K upon determining K, it is called an M-sequence using M for Maximum Length. If the internal state (xN−K, xN−K+1, . . . , xN−1) of K bits is given, the values of xn, xN+1, . . . are uniquely determined and, hence, there are 2K different internal states. The number of internal states is 2K−1 except for a self-evident solution xN=0 of the above formula, and the M-sequence has it as a length. The circuit is configured by a shift register, and an EXOR (exclusive OR) which feeds back an intermediate value of the shift register, and the feedback configuration is determined in correspondence with a primitive polynomial to be implemented.
As described above, since the conventional random number generating method uses an LFSR (Linear Feedback Shift Register), neighboring numerical values have correlation and a unique spectrum. Especially, middle and low frequency ranges have peaks, and if this method is used in image processing or the like, it poses a problem of image quality disturbance.
When a K-bit LFSR is used, its length is 2K−1, and it is difficult to generate random numbers of a plurality of lengths from a single random number generator. Since only one primitive polynomial can be implemented, a random number pattern is also limited to one.
It is impossible to generate an identical random number sequence in the reverse direction, and the use application of the method is limited.
The present invention has been made to solve the aforementioned problems, and has as its object to provide a technique for generating an arbitrary M-sequence random number of K bits or less.
It is another object of the present invention to generate a visibly satisfactory random number including a small number of low-frequency components.
In order to achieve an object of the present invention, for example, a random number generating method of the present invention comprises the following arrangement.
That is, a random number generating method for generating a random number expressed by K bits, comprising:
a mask step of masking a bit sequence of K bits held in a memory by a predetermined bit pattern;
an operation step of making an exclusive-OR operation of the bit sequence masked in the mask step, and setting a bit value indicating the exclusive-OR operation result to be included in a result obtained by shifting the bit sequence of K bits held in the memory in a predetermined direction;
a first inversion control step of controlling inversion/non-inversion of bit values in the bit sequence by the operation step in accordance with a designated bit value in the predetermined bit pattern; and
an output step of outputting the bit sequence as the processing result in the first inversion control step as a random number expressed by K bits.
In order to achieve an object of the present invention, for example, a random number generating method of the present invention comprises the following arrangement.
That is, a random number generating method for generating a random number expressed by K bits, comprising:
a mask step of masking a bit sequence of (K−1) bits held in a memory by a predetermined bit pattern;
a first operation step of making an exclusive-OR operation of the bit sequence masked in the mask step, and setting a bit value indicating the exclusive-OR operation result to be included in a result obtained by shifting the bit sequence of (K−1) bits held in the memory in a predetermined direction;
a selection step of selecting a bit value in the predetermined bit pattern in accordance with a position of a pixel to be processed on an image;
an inversion control step of controlling a process for inverting the bit value selected in the selection step in accordance with the position of the pixel to be processed on the image; and
an output step of outputting a bit sequence obtained by adding a bit of the processing result in the inversion control step to a head of the bit sequence by the first operation step as a random number expressed by K bits.
In order to achieve an object of the present invention, for example, a random number generating apparatus of the present invention comprises the following arrangement.
That is, a random number generating apparatus for generating a random number expressed by K bits, comprising:
a register for holding a bit sequence of K bits;
mask means for masking the bit sequence held in the register by a predetermined bit pattern;
shift means for shifting the bit sequence held in the register in a predetermined direction;
operation means for making an exclusive-OR operation of the bit sequence masked by the mask means, and setting a bit value indicating the exclusive-OR operation result to be included in the bit sequence in the register, which is shifted by the shift means;
first inversion control means for controlling a process for inverting bit values in the bit sequence by the operation means in accordance with a designated bit value in the predetermined bit pattern; and
output means for outputting the bit sequence as the processing result of the first inversion control means as a random number expressed by K bits.
In order to achieve an object of the present invention, for example, a random number generating apparatus of the present invention comprises the following arrangement.
That is, a random number generating apparatus for generating a random number expressed by K bits, comprising:
a register for holding a bit sequence of (K−1) bits;
mask means for masking the bit sequence held in the register by a predetermined bit pattern;
shift means for shifting the bit sequence held in the register in a predetermined direction;
first operation means for making an exclusive-OR operation of the bit sequence masked by the mask means, and setting a bit value indicating the exclusive-OR operation result to be included in the bit sequence in the register, which is shifted by the shift means;
second operation means for calculating a bit value determined in accordance with a bit value depending on whether or not a bit sequence for a predetermined number of bits in the bit sequence by the first operation means is a specific bit sequence, and a bit value determined in accordance with a designated bit value in the predetermined bit pattern; and
output means for outputting a bit sequence obtained by adding the bit calculated by the second operation means to a head of the bit sequence calculated by the first operation means as a random number expressed by K bits.
In order to achieve an object of the present invention, for example, a random number generating apparatus of the present invention comprises the following arrangement.
That is, a random number generating apparatus for generating a random number expressed by K bits, comprising:
a first register for holding a bit sequence of K bits;
first selection means for selecting a predetermined bit from the bit sequence held in the first register;
first shift means for shifting the bit sequence held in the first register in a predetermined direction;
first operation means for making an exclusive-OR operation of a bit sequence selected by the first selection means, and setting a bit value indicating the exclusive-OR operation result to be included in the bit sequence in the first register, which is shifted by the first shift means;
inversion control means for controlling a process for inverting bit values in the bit sequence by the first operation means in accordance with a bit value determined depending on whether a bit sequence for a predetermined number of bits in the bit sequence by the first operation means is a specific bit sequence, and a position of a pixel to be processed on an image; and
output means for outputting the bit sequence as the processing result of the inversion control means as a random number expressed by K bits.
In order to achieve an object of the present invention, for example, a random number generating apparatus of the present invention comprises the following arrangement.
That is, a random number generating apparatus for generating a random number expressed by K bits, comprising:
a first register for holding a bit sequence of (K−1) bits;
first selection means for selecting a predetermined bit from the bit sequence held in the first register;
first shift means for shifting the bit sequence held in the first register in a predetermined direction;
first operation means for making an exclusive-OR operation of a bit sequence selected by the first selection means, and setting a bit value indicating the exclusive-OR operation result to be included in the bit sequence in the first register, which is shifted by the first shift means;
selection means for selecting a bit value in a predetermined bit pattern in accordance with a position of a pixel to be processed on an image;
inversion control means for controlling a process for inverting a value of the bit selected by the selection means in accordance with the position of the pixel to be processed on the image; and
output means for outputting a bit sequence obtained by adding the bit as the processing result of the inversion control means to a head of the bit sequence by the first operation means as a random number expressed by K bits.
In order to achieve an object of the present invention, for example, a random number generating method of the present invention comprises the following arrangement.
That is, a random number generating method for generating a random number expressed by K bits, comprising:
a first selection step of selecting a predetermined bit from a bit sequence of K bits held in a first register;
a first shift step of shifting the bit sequence held in the first register in a predetermined direction;
a first operation step of making an exclusive-OR operation of a bit sequence selected in the first selection step, and setting a bit value indicating the exclusive-OR operation result to be included in the bit sequence in first register, which is shifted in the first shift step;
an inversion control step of controlling a process for inverting bit values in the bit sequence by the first operation step in accordance with a bit value determined depending on whether a bit sequence for a predetermined number of bits in the bit sequence by the first operation step is a specific bit sequence, and a position of a pixel to be processed on an image; and
an output step of outputting the bit sequence as the processing result of the inversion control step as a random number expressed by K bits.
In order to achieve an object of the present invention, for example, a random number generating method of the present invention comprises the following arrangement.
That is, a random number generating method for generating a random number expressed by K bits, comprising:
a first selection step of selecting a predetermined bit from a bit sequence of (K−1) bits held in a first register;
a first shift step of shifting the bit sequence held in the first register in a predetermined direction;
a first operation step of making an exclusive-OR operation of a bit sequence selected in the first selection step, and setting a bit value indicating the exclusive-OR operation result to be included in the bit sequence in the first register, which is shifted in the first shift step;
a selection step of selecting a bit value in a predetermined bit pattern in accordance with a position of a pixel to be processed on an image;
an inversion control step of controlling a process for inverting a value of the bit selected in the selection step in accordance with the position of the pixel to be processed on the image; and
an output step of outputting a bit sequence obtained by adding the bit as the processing result of the inversion control step to a head of the bit sequence by the first operation step as a random number expressed by K bits.
Other features and advantages of the present invention will be apparent from the following description taken in conjunction with the accompanying drawings, in which like reference characters designate the same or similar parts throughout the FIGS. thereof.
The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.
Preferred embodiments in which a random number generating apparatus of the present invention is applied to a random number generator will be described in detail hereinafter with reference to the accompanying drawings. Of course, the following embodiments provide disclosures for facilitating implementations by those who are skilled in the art of the present invention, and are only some embodiments included in the technical scope of the present invention defined by the appended claims. Therefore, as can be apparent to those who are skilled in the art, even embodiments which are not directly described in the present specification are included in the technical scope of the present invention as long as they have a common technical idea.
Note that a plurality of embodiments will be described for the sake of convenience. As is easily understood by those who are skilled in the art, not only these embodiments are individually achieved as inventions, but also an invention is achieved by combining a plurality of embodiments as needed.
Referring to
The shift register 1 shifts the held bit sequence 1 bit to the left every time it generates a random number. The mask circuit 2 masks only predetermined bits of the bit sequence output from the shift register 1, and outputs them to the EXOR circuit 3. More specifically, the mask circuit 2 masks bits other than those which correspond to a primitive polynomial. For example, when the shift register 1 holds 9 bits “110010100” (“0” is held in a memory part with index 0, “0” in a memory part with index 1, “1” in a memory part with index 2, “0” in a memory part with index 3, “1” in a memory part with index 4, “0” in a memory part with index 5, “0” in a memory part with index 6, “1” in a memory part with index 7, and “1” in a memory part with index 8), and a random number based on a primitive polynomial X7+X+1 is to be generated, the mask circuit 2 masks bits other than bits 6 and 0 (i.e., it computes a logical product of the shift register value and the mask value “001000001”), and outputs a bit sequence “000000000”. When a random number based on primitive polynomial X9+X4+1 is to be generated, the mask circuit 2 masks bits other than bits 8 and 3 (i.e., it computes a logical product of the shift register value and the mask value “100001000”), and outputs a bit sequence “100000000”. The generated bit sequence is input to the EXOR circuit 3.
The EXOR circuit 3 calculates a feedback value with reference to the bit sequence received from the mask circuit 2. The EXOR circuit 3 executes the same process as a so-called parity bit generating circuit. That is, the EXOR circuit 3 counts the number of bits “1” in the bit sequence input from the mask circuit 2, and outputs as the feedback value 0 if the count is an even number or 1 if it is an odd number. This feedback value is input (fed back) to the lower bit (a position indicated by “0” in
Therefore, by changing the mask pattern, arbitrary primitive polynomials of K-th order or less can be realized.
The pattern table 4 is a memory that holds H bits in the vertical direction, and W bits in the horizontal direction. In
The line counter 6 is used to read out a pattern of the pattern table 4 corresponding to a line of an image to be processed. More specifically, the count value counted by the line counter 6 indicates a line of the pattern table 4 which corresponds to a line including a pixel to be processed. The count value of the line counter 6 is input to the selector 5. Let c be the count value of the line counter 6. Then, the selector 5 reads out bits held in respective grids of the c-th column, i.e., a bit sequence of 4 bits, and outputs it to the selector 7.
The pixel counter 8 is used to read out a pattern of the pattern table 4 corresponding to the horizontal position of a pixel to be processed. That is, the count value counted by the pixel counter 8 indicates the position of a pattern n line data of the pattern table 4 selected by the line counter 6 which corresponds to the pixel to be processed. The count value of the pixel counter 8 is input to the selector 7. Let f be the count value of the pixel counter 8. Then, the selector 7 reads out the f-th bit from the bit sequence of 4 bits output from the selector 5, and outputs it to the inverter 9.
More specifically, one of 16 bits held by the pattern table 4 is designated by the line counter 6 and pixel counter 8, and the designated bit is output to the inverter 9. In other words, the designated bit is determined by the position of the line to be processed, and that of the pixel to be processed in that line.
The inverter 9 refers to 1 bit output from the selector 7. If the referred bit value is 1, the inverter 9 inverts bit values which form the bit sequence of 9 bits output from the shift register 1, and externally outputs them; if the referred bit value is 0, the inverter 9 externally outputs the bit sequence of 9 bits output from the shift register 1 without any process. In this way, the inverter 9 outputs a random number expressed by 9 bits.
With the aforementioned process, one of 16 bits held by the pattern table 4 is determined in accordance with the position of the pixel to be processed, and inversion/non-inversion of the bit sequence held by the shift register 1 can be controlled in accordance with the determined bit value. In this way, the bit sequence held by the shift register 1 is two-dimensionally modulated. For example, when the pattern table 4 stores 16 bits “1010-0101-1010-0101”, the low-frequency component of the random number generated by the shift register 1 is shifted toward the high-frequency side.
According to the random number generating method of this embodiment, the length of the random number to be generated is 2×(2K−1), i.e., twice the conventional length 2K−1, and a uniform random number with a large length can be generated.
With the above inversion operation, the random number value becomes 0 (when the random number value with all bits “1” is inverted) in some case. If 0 poses a problem, inversion when all the bits of the random number are “1” is inhibited, or such random number is substituted by data with all bits “1”, thus generating the same uniform random numbers as the conventional ones. In order to generate the same random number sequence in the reverse direction, the shift direction of the shift register is reversed to replace the LSB position and the MSB position (the highest order position in case of K bits or less) of the mask.
The shift register 11, mask circuit 12, and EXOR circuit 13 are respectively the same as the shift register 1, mask circuit 2, and EXOR circuit 3, and operate, as described in the first embodiment. By setting a mask of the mask circuit 12 to be different from that of the mask circuit 2, a random number having a different sequence (primitive polynomial) from that generated by the shift register 1, mask circuit 2, and EXOR circuit 3 is generated. By loading the value of the shift register 11 to the shift register 1 at the end (or head) of the line, correlation between lines of a random number generated by the shift register 1, mask circuit 2, and EXOR circuit 3 can be eliminated.
The pattern table 14 stores patterns for the length (2K−1) of M-sequence random numbers. In this embodiment, since K=9, the pattern table 14 stores patterns of bit sequences for 511 bits.
The pixel counter 18 is used to select the value of the pattern table 14 corresponding to the horizontal position of a pixel to be processed on an image.
The selector 17 selects one of 511-bit bit sequences held by the pattern table 14 in accordance with the value of the pixel counter 18, and outputs it to the EXOR circuit 19.
The pattern detector 15 refers to the bit sequence of 9 bits output from the shift register 1 or 11, and checks if this bit sequence is that of a specific pattern which is determined in advance. If the bit sequence of interest forms the specific pattern, the count value counted by the pixel counter 18 is reset to 0 and the count value of the pattern counter 20 is inverted.
The pattern counter 20 is a 1-bit counter, and always outputs the same count value (0 or 1) unless it receives an instruction from the pattern detector 15. As described above, when the pattern detector 15 detects a specific pattern, the pattern counter 20 inverts this count value. In this manner, an inversion pattern which uniquely corresponds to a random number sequence is input to the EXOR circuit 19 via the pixel counter 18 and selector 17.
The line counter 16 is a 1-bit counter, and inverts its count value for each line. Therefore, the inverted pattern is also inverted for each line.
Since the “bit sequence with a specific pattern which is determined in advance” is used to attain synchronization with the random number sequence, the pattern of the bit sequence is basically not limited. For example, in order to facilitate detection, a pattern with all bits=1 may be adopted. Alternatively, the pixel counter 18 may be configured as a (2K−1)-ary counter to invert the value of the pattern counter by a carry of this counter. In this case, the pattern detector 15 is omitted.
The EXOR circuit 19 EXORs 1 bit as the output from the selector 17, 1 bit as the output from the pattern counter 20, and 1 bit as the output from the line counter 16, and outputs 1 bit as the EXOR result to the inverter 9. When this 1 bit as the EXOR result is 1, the inverter 9 inverts bits which form 9 bits from the shift register 1, and externally outputs inverted 9 bits. On the other hand, when 1 bit as the EXOR result of the EXOR circuit 19 is 0, the inverter 9 externally outputs 9 bits from the shift register 1 intact.
With the above inversion pattern setting, a random number having a spectrum (frequency characteristics) according to the use application can be generated.
When the mask pattern of the mask circuit 12 is matched with that of the mask circuit 2, correlation may be provided in the vertical direction, or random number sequences with different lengths may be generated. Alternatively, in order to improve the randomness, a random number generated by the shift register 11, mask circuit 12, and EXOR circuit 13 may be generated N times for respective lines, and may then be loaded to the shift register 1. Especially, after a random number is idly generated in correspondence with the number of bits of the shift register (or with the highest-order of a primitive polynomial), it may be loaded to the shift register 1, thus eliminating correlation between lines.
The pattern table 14 may have a two-dimensional structure to shape the spatial frequency. In this case, the line counter 16 is used to select a line of the pattern table 14 in synchronism with a random number generated by the shift register 11, mask circuit 12, and EXOR circuit 13. Note that the EXOR circuit 19 may have two inputs since inversion by the line counter 16 is not required.
The selector 23 receives 9 bits from the shift register 11, selects a predetermined number of bits equal to or smaller than 9 bits (4 bits in
Likewise, the selector 24 receives 9 bits from the shift register 1, selects a predetermined number of bits equal to or smaller than 9 bits (4 bits in
The selector 25 extracts a bit sequence for a predetermined number of bits from a bit corresponding to the height order number of a primitive polynomial from the bit sequence of 9 bits from the shift register 1, and inputs the extracted bit sequence to the pattern detector 26.
The pattern detector 26 checks if the bit sequence for a predetermined number of bits input from the selector 25 matches that of a specific pattern. If the bit sequence matches the specific pattern, the pattern detector 26 inputs bit “1” to the EXOR circuit 28; if they do not match, the detector 26 inputs bit “0” to the EXOR circuit 28.
On the other hand, the pixel counter 27 holds 1-bit data, and inverts that bit every time a random number sequence is generated. That is, the pixel counter 27 alternately outputs 0 and 1 like 0, 1, 0, 1, . . . every time a random number sequence is generated. The count value of the pixel counter 27 is input to the EXOR circuit 28.
The EXOR circuit 28 EXORs 1 bit from the pattern detector 26 and 1 bit from the pixel counter 27, and outputs 1 bit as the EXOR result to the EXOR circuit 19. That is, when the bit sequence for the predetermined number of bits input from the selector 25 matches the bit sequence of the specific pattern, the EXOR circuit 28 executes a process for inverting the bit from the pixel counter 27, and outputs the inverted bit to the EXOR circuit 19.
For example, when the values of 4 bits from the highest order are 0101, since the M-sequence random number uses a shift register, the MSBs of the subsequent random numbers are in the order of 0, 1, 0, and 1. The count value of the pixel counter 27 is inverted every time a random number is generated. For example, when the MSBs of the generated random numbers are output in the order of 0, 1, 0, 1, . . . , and these values are inverted, the MSBs of the output random numbers consecutively assume the same values like 0000. This means that the high-frequency component of the random number is modulated to the low-frequency side, and the random number becomes visually readily detectable. Hence, a bit sequence for a predetermined number of bits from the MSB side of the random number is extracted to detect a pattern such as 0101, and the value of the pixel counter 27 is inverted if such pattern is detected, thus preventing the random number from being modulated to the low-frequency side.
Since the selector 24 is switched in response to the pattern sync signal (the output from the pattern detector 15 or a carry of a (2K−1)-ary counter), a primitive polynomial can be switched to different ones in turn. In this manner, the length can be further extended while maintaining the characteristics of uniform random numbers. Note that the length in the line direction can be extended by switching the selector 23. Since the switching timing at that time can be set while the shift operation of the shift register 11 stops, easy switching is allowed.
The EXOR circuit 19 EXORs all input bits, and inputs 1 bit as its EXOR result to the inverter 9. The subsequent processes are the same as those in the first embodiment.
In this embodiment, the EXOR circuits 21 and 22 have four inputs. However, the present invention is not limited to this, and these EXOR circuits may have, e.g., two inputs. In this case, some primitive polynomials (those of 8th order in this case) cannot be implemented. However, since the number of primitive polynomials which cannot be implemented is small, no serious problem is posed.
The process for generating a random number for each line by the LFSR configured by the shift register 31, mask circuit 32, and EXOR circuit 33 is the same as that in the above embodiments. The same applies to the LFSR configured by the shift register 34, mask circuit 35, and EXOR circuit 36. In this embodiment, since the MSB of a random number is generated by modulating the MSB table 29, the number of bits of a bit sequence handled by each LFSR is smaller by 1 bit.
The line counter 30 is used to select a line of the MSB table 29. The line counter 30 is sequentially counted up every time a line is switched as in the line counter 16. When the count value of the line counter 30 reaches the number of lines of the MSB table 29, it is reset to zero to repeat counting for the number of lines of the MSB table 29.
The MSB table 29 two-dimensionally stores the MSBs of random numbers, and one MSB is selected by the line counter 30, pixel counter 18, and selector 17. This selection is made by specifying a column by the line counter 30 and selecting one bit of this column by the pixel counter 18 as in the first embodiment.
The EXOR circuit 19 EXORs 1 bit from the selector 17, and 1 bit output from the pattern detector 15 and pattern counter 20, and outputs 1 bit as an EXOR result. The output one bit is appended to the head of an 8-bit bit sequence from the shift register 31 or 34 as the MSB, thus externally outputting a bit sequence of a total of 9 bits.
With the above arrangement, the size of the MSB table 29 in the pixel direction matches the length of a random number formed by the shift register 34, mask circuit 35, and EXOR circuit 36, and the size of the MSB table 29 in the line direction matches the length of a random number formed by the shift register 31, mask circuit 32, and EXOR circuit 33. Hence, a random number having desired spatial frequency characteristics can be generated.
The random number generating method described in each of the first to fourth embodiments is implemented by hardware with the arrangement shown in each of
A method of supplying the program of this software is not particularly limited. As is well known, the program may be loaded from a hard disk drive to the RAM, or the program of the software may be supplied from an external apparatus to the RAM of this computer via a network such as a LAN, Internet, or the like.
The objects of the present invention are also achieved when a CPU or MPU of a device such as a camera, printer, or the like reads out and executes a program code from a recording medium (or storage medium), which records a program code of a software program that can implement the functions of the above-mentioned embodiments. In this case, the program code itself read out from the recording medium implements the functions of the above-mentioned embodiments, and the recording medium which stores the program code constitutes the present invention.
The functions of the above-mentioned embodiments may be implemented not only by executing the readout program code by the device such as a camera, printer, or the like, but also by some or all of actual processing operations executed by an operating system (OS) running on the device such as a camera, printer, or the like on the basis of an instruction of the program code.
Furthermore, the functions of the above-mentioned embodiments may be implemented by some or all of actual processing operations executed by a CPU or the like arranged in a function extension card or a function extension unit, which is inserted in or connected to the device such as a camera, printer, or the like, after the program code read out from the recording medium is written in a memory of the extension card or unit.
When the present invention is applied to the recording medium, that recording medium stores program codes corresponding to the aforementioned functional arrangement.
As many apparently widely different embodiments of the present invention can be made without departing from the spirit and scope thereof, it is to be understood that the invention is not limited to the specific embodiments thereof except as defined in the claims.
This application claims priority from Japanese Patent Application No. 2004-174607 filed on Jun. 11, 2004, the entire contents of which are hereby incorporated by reference herein.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2004-174607 | Jun 2004 | JP | national |
| Number | Name | Date | Kind |
|---|---|---|---|
| 5805738 | Kaburagi et al. | Sep 1998 | A |
| 6339781 | Sasaki | Jan 2002 | B1 |
| 7426528 | Denk | Sep 2008 | B2 |
| 20030236803 | Williams | Dec 2003 | A1 |
| 20040078401 | Hilton | Apr 2004 | A1 |
| Number | Date | Country |
|---|---|---|
| 6-276067 | Sep 1994 | JP |
| 10-42132 | Feb 1998 | JP |
| Number | Date | Country | |
|---|---|---|---|
| 20050278402 A1 | Dec 2005 | US |
