Patent Application
20060088274
1. Field of the Invention
The present invention relates to a video signal processing apparatus and a video signal processing method that determine pixels that have been dropout compensated from a video signal and perform an interpolation process for the determined pixels.
2. Description of the Related Art
In recent years, digital recording mediums such as digital versatile discs (DVDs) on which digital signals are recorded have been outspread. Analog video signals that were recorded by video cassette recorders (VCRs) are converted into digital signals and then recorded on digital recording mediums.
When an analog video signal recorded on a magnetic tape such as a VCR tape is reproduced, a part of a video signal may be dropped out due to dirt and a scratch on the surface of the tape. Thus, the video signal may not be correctly read, namely dropped out. When an analog VCR detects a dropout, since lines of a video signal are correlated, the analog VCR compensates the dropout by outputting the immediately preceding line of the video signal.
The following patent document 1 describes a dropout compensating apparatus that replaces a dropout portion with a past video signal that is strongly correlated with the dropout portion. [Patent Document 1] Japanese Patent Laid-Open Publication No. HEI 7-123365
However, depending on a video signal, it may be insufficient to compensate a dropout portion with the immediately preceding-line of the video signal. The dropout may cause the picture quality of the video signal to deteriorate. In particular, when a correlation between a line to be compensated and the immediately preceding line is weak, the picture quality of the video signal largely deteriorates. In this case, when a video signal that has been dropout compensated is recorded on a recording medium, the video signal whose picture quality is deteriorated is recorded on the recording medium.
Thus, it is necessary to detect pixels that have been dropout compensated (these pixels are referred to as compensated pixels) and interpolate the compensated pixels with good pixels of the video signal. Thus, it is necessary to determine compensated pixels of a video signal.
In the past, when a dropout occurred in an analog VCR, it output a dropout indication signal that indicates that the dropout occurred. When a compensated pixel is detected with the dropout indication signal, since a circuit that detects the dropout indication signal is disposed, the structure of the apparatus becomes complicated. In addition, this circuit can be used for only predetermined hardware.
In view of the foregoing, it would be desirable to provide a video signal processing apparatus and a video signal processing method that determine compensated pixels with a video signal that has been dropout compensated, replace the compensated pixels with interpolated pixels, and generate a video signal closer to an original video signal.
According to an embodiment, there is provided a video signal processing apparatus that determines whether pixels of a video signal have been dropout compensated and compensates pixels that have been dropout compensated according to a correlation of adjacent pixels, corresponding pixels in adjacent fields, or corresponding pixels in adjacent frames.
According to an embodiment, there is provided a video signal processing apparatus that determines pixels that have been dropout compensated from a video signal that has been dropout compensated. The video signal processing apparatus has a detection section and a determination section. The detection section detects the absolute value or the square value of the difference of pixel data of two pixels that exist at the same horizontal position of two adjacent lines of the video signal. The determination section determines a pixel that satisfies a first condition of which the absolute value or the square value is smaller than a predetermined threshold value to be a pixel that has been dropout compensated.
According to an embodiment, there is provided a video signal processing apparatus that determines pixels that have been dropout compensated from a video signal that has been dropout compensated. The video signal processing apparatus has a first detection section, a second detection section, and a determination section. The first detection section detects the absolute value or the square value of the first difference of pixel data of two pixels that exist at the same position in a predetermined field or frame and the temporally adjacent field or frame that is immediately preceded by the predetermined field or frame. The second detection section detects the absolute value or the square value of the second difference of pixel data of two pixels that exist at the same position in the predetermined field or frame and the temporally adjacent field or frame that is immediately followed by the predetermined field or frame. The determination section determines a pixel that satisfies a second condition of which the absolute value or the square value of the first difference is larger than a predetermined threshold value and the absolute value or the square value of the second difference is larger than the predetermined threshold value to be a pixel that has been dropout compensated.
According to an embodiment, there is provided a video signal processing apparatus that determines pixels that have been dropout compensated from a video signal that has been dropout compensated. The video signal processing apparatus has a first detection section, a second detection section, a third detection section, and a determination section. The first detection section detects the absolute value or the square value of the first difference of pixel data of two pixels that exist at the same horizontal position of adjacent two lines of the video signal. The second detection section detects the absolute value or the square value of the second difference of pixel data of two pixels that exist at the same position in a predetermined field or frame and the temporally adjacent field or frame that is immediately preceded by the predetermined field or frame. The third detection section detects the absolute value or the square value of the third difference of pixel data of two pixels that exist at the same position in the predetermined field or frame and the temporally adjacent field or frame that is immediately followed by the predetermined field or frame. The determination section determines a pixel that satisfies both a first condition of which the absolute value or the square value of the first difference is smaller than a first threshold value and a second condition of which the absolute value or the square value of the second difference is larger than a second threshold value and the absolute value or the square value of the third difference is larger than the second threshold value to be a pixel that has been dropout compensated.
According to an embodiment, there is provided a video signal processing method of determining pixels that have been dropout compensated from a video signal that has been dropout compensated. The absolute value or the square value of the difference of pixel data of two pixels that exist at the same horizontal position of two adjacent lines of the video signal is detected. A pixel that satisfies a first condition of which the absolute value or the square value is smaller than a predetermined threshold value is determined to be a pixel that has been dropout compensated.
According to an embodiment, there is provided a video signal processing method of determining pixels that have been dropout compensated from a video signal that has been dropout compensated. The absolute value or the square value of the first difference of pixel data of two pixels that exist at the same position in a predetermined field or frame and the temporally adjacent field or frame that is immediately preceded by the predetermined field or frame is detected. The absolute value or the square value of the second difference of pixel data of two pixels that exist at the same position in the predetermined field or frame and the temporally adjacent field or frame that is immediately followed by the predetermined field or frame is detected. A pixel that satisfies a second condition of which the absolute value or the square value of the first difference is larger than a predetermined threshold value and the absolute value or the square value of the second difference is larger than the predetermined threshold value is determined to be a pixel that has been dropout compensated.
According to an embodiment, there is provided a video signal processing method of determining pixels that have been dropout compensated from a video signal that has been dropout compensated. The absolute value or the square value of the first difference of pixel data of two pixels that exist at the same horizontal position of adjacent two lines of the video signal is detected. The absolute value or the square value of the second difference of pixel data of two pixels that exist at the same position in a predetermined field or frame and the temporally adjacent field or frame that is immediately preceded by the predetermined field or frame is detected. The absolute value or the square value of the third difference of pixel data of two pixels that exist at the same position in the predetermined field or frame and the temporally adjacent field or frame that is immediately followed by the predetermined field or frame is detected. A pixel that satisfies both a first condition of which the absolute value or the square value of the first difference is smaller than a first threshold value and a second condition of which the absolute value or the square value of the second difference is larger than a second threshold value and the absolute value or the square value of the third difference is larger than the second threshold value are determined to be a pixel that has been dropout compensated.
According to an embodiment of the present invention, compensated pixels are determined from a video signal. The compensated pixels are replaced with interpolated values. As a result, a video signal closer to an original video signal can be generated.
According to an embodiment of the present invention, since a video signal that has been dropout compensated is used to determine compensated pixels, the positions of compensated pixels can be detected and the compensated pixels can be interpolated without need to use particular hardware.
According to an embodiment of the present invention, since a dropout detection signal is not used for a pixel that has been dropout compensated, a dubbed video signal whose pixels have been dropout compensated can become closer to an original video signal. In addition, these processes can be performed by software.
These and other objects, features and advantages of the present invention will become more apparent in light of the following detailed description of a best mode embodiment thereof, as illustrated in the accompanying drawings.
The invention will become more fully understood from the following detailed description, taken in conjunction with the accompanying drawings, wherein similar reference numerals denote similar elements, in which:
Next, with reference to the accompanying drawings, an embodiment of the present invention will be described. According to the embodiment, input signals are component signals of which luminance signal Y and color difference signals Cr and Cb are Y/C separated.
The dropout detection section 11 performs a process (that will be described later) to determine compensated pixels of the video signal that is input to the input terminal 2. In addition, the dropout detection section 11 generates a dropout detection signal DO that represents whether a pixel has been dropout compensated. The detection signal DO is supplied to an interpolated pixel generation section 4 and a switch SW1. Hereafter, pixels that have not been dropout compensated are referred to as normal pixels.
The switch SW1 is connected to a terminal a1 or a terminal b1 according to the supplied detection signal DO. When a compensated pixel is detected, the switch SW1 is connected to the terminal a1 so that the compensated pixel is replaced with pixel data of an interpolated pixel generated by the interpolated pixel generation section 4, which will be described later. When a normal pixel is detected, the switch SW1 is connected to the terminal b1 so that pixel data supplied from the A/D converter 3 are output. A delay circuit (DL) 5 is disposed. The DL 5 delays the output of the A/D converter 3 for a process time of the interpolated pixel generation section 4. Pixel data selected by the switch SW1 are obtained from an output terminal 6.
The foregoing process applies to the color difference signals Cr and Cb as well as the luminance signal Y. In other words, the color difference signal Cr and the color difference signal Cb that are input to an input terminal 12 and an input terminal 22 are converted into digital signals by an A/D converter 13 and an A/D converter 23, respectively. In addition, the detection signal DO supplied from the dropout detection section 11 is supplied to a switch SW2 and a switch SW3.
Switching processes of the switch SW2 and the switch SW3 are controlled according to the detection signal DO. When a compensated pixel is detected, the switch SW2 is connected to a terminal a2 so that the compensated pixel is replaced with interpolated pixel data generated by an interpolated pixel generation section 14 and the interpolated pixel data are output. When a normal pixel is detected, the switch SW2 is connected to a terminal b2. The selected pixel data are output to an output terminal 16. Likewise, when a compensated pixel is detected, the switch SW3 is connected to a terminal a3 so that the compensated pixel is replaced with interpolated pixel data generated by an interpolated pixel generation section 24 and the interpolated pixel data are output. When a normal pixel is detected, the switch SW3 is connected to a terminal b3. The selected pixel data are supplied to an output terminal 26.
In addition, a delay circuit 15 and a delay circuit 25 are disposed. The delay circuit 15 and the delay circuit 25 delay outputs of the A/D converter 13 and the A/D converter 23 for process times of the interpolated pixel generation section 14 and the interpolated pixel generation section 24, respectively.
Output signals of the output terminals 6, 16, and 26 are converted into signals according to a predetermined format. The converted signals are recoded on a recording medium (not shown), a hard disc drive (HDD), or the like and output to a monitor (not shown).
The dropout detection section 11 performs a process that determines a compensated pixel. Next, this process will be described. In the following description, pixel data are denoted by I (y, x, t) where y represents a vertical position (line), x represents a horizontal position (column), and t represents a frame number or a field number. In the interlace system, since one frame is composed of two fields, when y is an even value, it represents an even field; when y is an odd value, it represents an odd field.
The dropout detection section 11 performs a process that obtains the difference between two pixels at the same horizontal position of adjacent two lines in the same field to determine a compensated pixel (hereinafter, this difference is referred to as line difference). The line difference is obtained by formula (1).
DA(y, x, t)=I(y, x, t)−I(y−2, x, t) (1)
The absolute value of the difference DA and a predetermined threshold value TA are compared. A pixel of which the difference DA is smaller than the threshold value TA is selected. Whether a pixel is selected is represented by data of one bit. When a pixel is selected, it is represented by “1” (that means logical “1”). When a pixel is not selected, it is represented by “0” (that means logical “0”). A condition of which the difference DA is smaller than the threshold value TA is referred to as a first condition.
For easy understanding,
Mask A=|DA|<TA (2)
A dropout compensation is performed by replacing a dropout with the immediately preceding line of a video signal. Thus, the difference of pixel data of two pixels at the same horizontal position of the two adjacent lines becomes small. As a result, a pixel whose value is “1” in the mask A, which represents that the difference of pixel data is small, is a compensated pixel.
In addition, the dropout detection section 11 determines the chronological correlation of pixel data using the frame difference. The dropout detection section 11 obtains a first difference between pixel data I (y, x, t) and pixel data I (y, x, t−1), where t−1 represents the temporally immediately preceding frame of t, as expressed by formula (3).
DP(y, x)=I(y, x, t)−I(y, x, t−1) (3)
In addition, a second difference between pixel data I (y, x, t) and pixel data I (y, x, t+1), where t+1 represents the temporally immediately following frame of t, is obtained as expressed by formula (4).
Dn(y, x)=I(y, x, t)−I(y, x, t+1) (4)
The absolute value of the first frame difference Dp and the absolute value of the second frame difference Dn are compared with a predetermined threshold value TB. A pixel that satisfies a second condition of which the absolute value of the difference Dp is larger than the threshold value TB and the absolute value of the difference Dn is larger than the threshold value TB is selected. A mask B is generated with a pixel that satisfies the second condition. The mask B can be expressed by formula (5).
Mask B=(|Dp|>TB)AND(|Dn|>TB) (5)
The second condition may be expressed by formula (6) instead of formula (5).
Mask B=[(Dp<−TB)AND(Dn<−TB)]OR[Dp>TB]AND(Dn>TB)] (6)
A pixel that satisfies the second condition is “1” (that means logical “1”) and a pixel that does not satisfy the second condition is “0” (that means logical “0”). For easy understanding,
Since a video signal is continuous, a pixel in a predetermined frame and corresponding pixels in the immediately preceding and following frames have strong correlations. However, when the dropout compensation is performed, since a pixel of a line is replaced with a corresponding pixel of the immediately preceding line, a pixel in a predetermined frame and corresponding pixels in the immediately preceding and following frames have weak correlations and appear as noise. The strength and weakness of correlations can be determined by comparing the frame difference with a threshold value. Thus, it can be determined that a pixel that is “1” in the mask B, which represents a weak correlation, is a compensated pixel.
According to the embodiment, the mask B is generated by obtaining the difference between frames. Instead, the mask B may be generated by obtaining the difference between fields.
When a pixel that satisfies both the first condition and the second condition is “1”, a mask AB can be generated. The mask AB can be expressed by formula (7).
Mask AB=mask A & mask B (7)
A dropout tends to occur as a burst, not at random. Because of this tendency, it can be determined that when the number of pixels that are “1” and that are successive in the horizontal direction is not larger than a predetermined threshold value, they not be pixels that have not been dropout compensated (for example, noise). Pixels that are “1” of the mask AB and that are not successive for a predetermined threshold value are bit 0. These pixels are a mask C. A condition of which pixels that are “1” are successive in the horizontal direction for the threshold value or more is a third condition.
Thus, a moving portion is removed from an interlaced picture. Statistically, a dropout tends to occur in the same field, whereas a moving portion of an interlaced picture tends to occur in the same frame.
It is assumed that a dropout is detected in a line yn in which pixel data (y, x) exist. When a dropout is long, it occupies not only the line yn but line Yn+2, line Yn+4, and so forth, namely a plurality of lines of the same field. However, statistically, a dropout tends not to occur in different fields such as line yn+1 and line yn+1.
In contrast, a moving portion of a video signal of an interlaced picture occupies different fields. Thus, if a line yn and the upper and lower lines, line yn+1 or line yn+3 are determined as dropouts, it is determined that they be a moving portion of an interlaced picture and pixels that are “1” of the line yn are changed to “0”.
In other words, the number of pixels that are determined as “1” is calculated for each line as expressed by formula (8).
S(y)=Σ MASKc(y, x) where x=1, 2, (8)
Next, the differences between the value S (y) obtained for each line and the value S (y−1) for the immediately preceding line and the difference between the value S (y) obtained for each line and the value S (y+1) obtained for the immediately following line are obtained as expressed by formula (9) and formula (10), respectively.
DSn(y)=S(y)−S(y−1) (9)
DSp(y)=S(y)−S(y+1) (10)
When at least one of the obtained differences DSn and DSp is smaller than the threshold value Td, the pixels are normal pixels. When at least one of the differences DSn and DSp is equal to or larger than the threshold value Td, the pixels are compensated pixels. In other words, pixels that satisfy at least one of the conditions of formula (11) and formula (12) are “0” as expressed by formula (13).
DSn(yn)<Td (11)
DSp(yn)<Td (12)
MASKc(yn, x)=0 (13)
When the difference of the number of compensated pixels of one line and the number of compensated pixels of an adjacent line is smaller than the threshold value, it can be determined that the compensated pixels be a moving portion of an interlaced picture and occupy a plurality of lines. Thus, when a pixel of which at least one of the obtained differences is smaller than the threshold value is determined to be a normal pixel and “1” and a pixel of which at least one of the obtained differences is equal to or larger than the threshold value is determined to be a compensated pixel and “0”, a moving portion of an interlaced picture can be removed. The condition of which at least one of the differences obtained as expressed by formula (11) and formula (12) is smaller than the predetermined threshold value is referred to as a fourth condition.
Next, generation of interpolated values against compensated pixels will be described. When interpolated values are generated against compensated pixels, pixels that are strongly correlated with the compensated pixels are used. As a result, a video signal close to the original video signal can be generated. Thus, the interpolated pixel generation section 4, the interpolated pixel generation section 14, and the interpolated pixel generation section 24 detect pixels strongly correlated with compensated pixels.
The absolute values of the differences of a compensated pixel and each adjacent pixel that exists in the horizontal direction (D1), the vertical direction (D2), the left diagonal direction (D3), and the right diagonal direction (D4) are calculated. The direction of the two pixels of which the absolute value of the difference is the smallest has a strong correlation.
The absolute value of the difference of two pixels is obtained by for example formulas (12) to (15).
D1(y, x, t)=|I(y, x, t−1)−I(y, x, t+1)| (12)
D2(y, x, t)=|I(y−1, x, t)−I(y+1, x, t)| (13)
D3(y, x, t)=|I(y−1, x+1, t)−I(y+1, x−1, t)| (14)
D4(y, x, t)=|I(y−1, x−1, t)−I(y+1, x+1, t)| (15)
The average value of a compensated pixel and the adjacent pixel in the direction having a strong correlation with the compensated pixel is obtained as an interpolated value. The interpolated value is generated by the interpolated pixel generation section 4, the interpolated pixel generation section 14, and the interpolated pixel generation section 24 shown in
(1) An interpolated value is generated with the average value of a compensated pixel and a corresponding pixel in adjacent frames.
I′(y, x, t)=(I(y, x, t−1)+I(y, x, t+1))/2
(2) An interpolated value is generated with the average value of a compensated pixel and an upper or lower pixel in the same frame (or the same field).
I′(y, x, t)=(I(y−1, x, t)+I(y+1, x, t))/2
(3) An interpolated value is generated with the average value of a compensated pixel and an upper or lower left pixel in the same frame (or the same field).
I′(y, x, t)=(I(y−1, x+1, t)+I(y+1, x−1, t))/2
(4) An interpolated value is generated with the average value of a compensated pixel and an upper or lower right pixel in the same frame (or the same field)
I′(y, x, t)=(I(y−1, x−1, t)+I(y+1, x+1, t))/2
Compensated pixels are replaced with generated interpolated values and selectively output by the switch SW1, the switch SW2, and the switch SW3.
In the video signal processing apparatus shown in
It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and alternations may occur depending on design requirements and other factors insofar as they are within the scope of the appended claims or the equivalents thereof.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2004-309205 | Oct 2004 | JP | national |
The present invention contains subject matter related to Japanese Patent Application No. 2004-309205 filed in the Japanese Patent Office on Oct. 25, 2004, the entire contents of which being incorporated herein by reference.
