JP2010021689A - Video signal processing apparatus and nonlinear low-pass filter - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a video signal processing apparatus capable of expanding bit accuracy even when noise is superimposed on video signals, reducing the adverse effects of attenuating feeble texture and keeping the texture. <P>SOLUTION: The nonlinear LPF 1 executes nonlinear LPF processing to 8-bit digital video signals and outputs the low frequency component signals of 12-bit accuracy. A bit accuracy conversion part 3 converts the low frequency component signals to 8-bit accuracy. A subtractor 4 subtracts the output signals of the bit accuracy conversion part 3 from the digital video signals and outputs the high/low frequency component signals of the 8-bit accuracy. An adder 5 adds the low frequency component signals output from the nonlinear LPF 1 and the high/low frequency component signals output from the subtractor 4. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a video signal processing apparatus and a nonlinear low-pass filter that perform processing for extending the bit accuracy of a video signal, and in particular, can retain a weak texture of a video and superimpose noise on the video signal. In particular, the present invention relates to a video signal processing device and a nonlinear low-pass filter that can obtain the effect of the bit precision extension processing.

  Conventionally, the display resolution of the liquid crystal panel has been about 8 bits. However, in recent years, liquid crystal panels having a display resolution of 10 bits or more have appeared, and the contrast of the display image has been improved. As video signal processing technology advances, it is natural to perform noise reduction processing and gradation correction. The video signal input to the liquid crystal panel is generally a digital video signal quantized with 8 bits such as a DVD or a high-definition broadcast. When an 8-bit video signal is displayed on a liquid crystal panel having a display resolution of 10 bits, contour gradation stripes resulting from 8-bit quantization may be seen, and smooth gradations may not be displayed. This phenomenon appears more prominently when tone correction processing such as gamma correction processing or contrast enhancement is performed.

Therefore, various video signal processing apparatuses that perform processing for extending the bit accuracy of the video signal in order to display smooth gradation have been proposed. For example, Japanese Patent Application Laid-Open No. 2004-054210 (Patent Document 1) describes that a gradation part is detected and a lower bit is added to the gradation part to extend the bit accuracy. Japanese Patent Application Laid-Open No. 2007-181189 (Patent Document 2) describes that a gradation portion where gradation stripes are conspicuous is detected and a desirable gradation is obtained by linear interpolation with high bit accuracy. Japanese Patent Laid-Open No. 2008-047986 (Patent Document 3) describes a video signal processing apparatus that performs processing for extending the bit accuracy of a video signal using a nonlinear low-pass filter or a coring circuit.
JP 2004-054210 A JP 2007-181189 A JP 2008-047986 A

  When detecting the gradation part as described in Patent Documents 1 and 2 and extending the bit precision, erroneous detection occurs when the gradation part is detected. If noise larger than the least significant bit of the video signal is superimposed, the gradation portion cannot be detected and a sufficient effect cannot be obtained. There is a demand for a video signal processing apparatus capable of effectively extending the bit accuracy even when noise is superimposed on the video signal. In addition, a video signal processing apparatus that can maintain a weak texture such as a skin texture is desired.

The present invention has been made in view of such problems, and even if a certain amount of noise is superimposed on the video signal, the bit accuracy can be effectively extended, and there is little adverse effect of attenuating a weak texture. Another object of the present invention is to provide a video signal processing apparatus capable of retaining the texture.
It is another object of the present invention to provide a non-linear low-pass filter suitable for use in a video signal processing apparatus that performs processing for extending the bit accuracy of a video signal.

In order to solve the above-described problems of the prior art, the present invention applies a non-linear low-pass filter process to a digital video signal having a first number of quantization bits, and outputs a second larger than the first number of quantization bits. A non-linear low-pass filter (1) that outputs a low-frequency component signal with quantization bit number accuracy, and a bit-accuracy conversion that converts the low-frequency component signal output from the non-linear low-pass filter to the first quantization bit number accuracy Part (3), the digital video signal and the output signal of the bit precision conversion part are input in a state where their phases are aligned, and the output signal of the bit precision conversion part is subtracted from the digital video signal A subtractor (4) for outputting the first quantization bit number precision high and low frequency component signal; and the low frequency component signal output from the nonlinear low pass filter and the subtraction To provide a video signal processing apparatus characterized by comprising a first adder and (5) output by adding the more output said high and low-frequency component signal.
Here, the non-linear low-pass filter (1) includes a pixel signal of the target pixel and the plurality of pixels within a predetermined region including the target pixel located at the center and a plurality of peripheral pixels positioned around the target pixel. A plurality of differentiators (12a to 12x) for obtaining a difference from the pixel signal of each of the peripheral pixels, and positive or negative difference values output from the plurality of differentiators are input, and a first input value is positive. An input value is output in the range between the value and the negative second value, and when the input value exceeds the first value and is equal to or less than the positive third value, the input value is attenuated. When the input value exceeds the second value and is equal to or greater than the negative fourth value, the input value is amplified and the input value exceeds the third and fourth values. Are a plurality of nonlinear arithmetic units (13a to 13x) that output 0, and the plurality of nonlinear arithmetic units A cumulative addition unit (14) that adds all of the output values, an average value unit (16) that averages the output values from the cumulative addition unit by the number of surrounding pixels, and a pixel signal of the pixel of interest And a second adder (17) for adding the output signal of the averaging unit.
Further, a limiter (15) to which a positive or negative output value from the cumulative addition unit is input is provided between the cumulative addition unit (14) and the averaging unit (16), When the input value is in the range between the positive fifth value and the negative sixth value, the input value is output, and when the input value exceeds the fifth value and is equal to or less than the positive seventh value A value obtained by attenuating the input value is output. When the input value exceeds the sixth value and is equal to or greater than the negative eighth value, a value obtained by amplifying the input value is output. When the value exceeds the eighth value, it is preferable to output 0.

  Furthermore, in order to solve the above-described problems of the conventional technology, the present invention provides a pixel of interest within a predetermined area including a pixel of interest located at the center and a plurality of peripheral pixels located around the pixel of interest. A plurality of differentiators (12a to 12x) for obtaining a difference between a pixel signal and a pixel signal of each of the plurality of surrounding pixels, and each of positive or negative difference values output from the plurality of differentiators is input, Outputs an input value in the range between the positive first value and the negative second value, and when the input value exceeds the first value and is equal to or less than the positive third value, the input value When the input value exceeds the second value and is a negative fourth value or more, a value obtained by amplifying the input value is output, and the input value is the third and third values. A plurality of nonlinear arithmetic units (13a to 13x) that output 0 when a value of 4 is exceeded; A cumulative addition unit (14) for adding all output values from the nonlinear arithmetic unit, and a positive or negative output value from the cumulative addition unit are input, and a positive fifth value and a negative sixth value are input. An input value is output in a range between values, and when the input value exceeds the fifth value and is equal to or less than a positive seventh value, a value obtained by attenuating the input value is output. A limiter that outputs an amplified input value when it exceeds the sixth value and is equal to or greater than the negative eighth value, and outputs 0 when the input value exceeds the seventh and eighth values. 15), an averaging unit (16) that averages the output value from the limiter by the number of surrounding pixels, and an addition that adds the pixel signal of the pixel of interest and the output signal of the averaging unit A non-linear low-pass filter, characterized in that it comprises a device (17).

  According to the video signal processing device and the non-linear low-pass filter of the present invention, even if a certain amount of noise is superimposed on the video signal, the bit accuracy can be effectively expanded, and there is little adverse effect of attenuating a weak texture, The texture can be maintained.

  Hereinafter, a video signal processing apparatus and a nonlinear low-pass filter according to the present invention will be described with reference to the accompanying drawings. FIG. 1 is a block diagram showing an embodiment of a video signal processing apparatus of the present invention. In the present embodiment, an example in which a video signal having a quantization bit number of 8 bits is extended to 12-bit precision is taken as an example. The present invention is not limited to extending 8 bits to 12-bit accuracy, but extends a video signal having a first bit number to a second bit number accuracy greater than the first bit number. I just need it.

  In FIG. 1, a pixel signal of an 8-bit video signal is sequentially input to a nonlinear low-pass filter (nonlinear LPF) 1 and a delay unit 2. The video signal may be a luminance signal or a color signal. A low-frequency component signal with 12-bit accuracy is output from the nonlinear LPF1. The specific configuration and operation of the nonlinear LPF 1 will be described in detail later. The output signal of the nonlinear LPF 1 is input to the bit precision conversion unit 3 and the adder 5. The bit precision conversion unit 3 rounds off the lower 4 bits of 12 bits and then rounds down the lower 4 bits to output an 8-bit signal. The 8-bit output signal of the bit precision conversion unit 3 is input to the subtracter 4.

  The delay unit 2 delays the input video signal (pixel signal) by a processing time in the nonlinear LPF 1 and the bit precision conversion unit 3. The 8-bit pixel signal from the delay unit 2 is input to the subtracter 4. As a result, the phase of the output signal of the bit precision conversion unit 3 input to the subtracter 4 and the pixel signal output from the delay unit 2 coincide with each other. The subtracter 4 outputs an 8-bit precision high-frequency component signal by subtracting the output signal of the bit precision conversion unit 3 from the pixel signal output from the delay device 2. The adder 5 adds the output signal of the nonlinear LPF 1 and the output signal of the subtractor 4 and outputs the result.

  Comparing the input video signal and the output video signal of the video signal processing apparatus shown in FIG. 1, the bit accuracy of the low frequency component signal is increased from 8 bits to 12 bits, but the high frequency component signal is substantially 8 bits. Bit remains unchanged. Therefore, as the processing for the video signal of the entire band, only the 8-bit LSB is changed between the input video signal and the output video signal, and it is difficult to cause a great adverse effect. If an edge component is included in the low-frequency component signal, the edge component changes (attenuates), and the sharpness of the entire image is lowered or the texture is deteriorated. Therefore, in this embodiment, the low-frequency component signal is prevented from including an edge component by the nonlinear LPF 1.

  Next, a specific configuration and operation of the nonlinear LPF 1 will be described with reference to FIG. In FIG. 2, as an example, the in-region pixel signal generator 11 generates pixel signals in a region of a total of 25 pixels of 5 pixels in the horizontal direction and 5 pixels in the vertical direction, as shown in FIG. 3. A pixel P33 located at the center of the pixels P11 to P15, P21 to P25, P31 to P35, P41 to P45, and P51 to P55 in FIG. 3 is a target pixel to be processed by the video signal processing device. Although the area is 25 pixels here, more pixels may be used. Further, it is not always necessary to use a pixel region continuous in the horizontal direction and the vertical direction. For example, pixels may be thinned out in a horizontal direction and a vertical direction in an area of 81 pixels as a whole to form an area of 25 pixels.

  FIG. 4 shows a specific configuration example of the in-region pixel signal generation unit 11 for generating the pixel signal of 25 pixels shown in FIG. In FIG. 4, a pixel signal is input to an input terminal 1100, and this pixel signal is sequentially delayed pixel by pixel by pixel delay units 1101 to 1104. The pixel signal input to the input terminal 1100 is delayed by one line by the line memory 1105. The pixel signals output from the line memory 1105 are sequentially delayed pixel by pixel by the pixel delay units 1106 to 1109 and further delayed by one line by the line memory 1110. Pixel signals output from the line memory 1110 are sequentially delayed pixel by pixel by the pixel delay units 1111 to 1114 and further delayed by one line by the line memory 1115.

  Pixel signals output from the line memory 1115 are sequentially delayed pixel by pixel by the pixel delay units 1116 to 1119 and further delayed by one line by the line memory 1120. Pixel signals output from the line memory 1120 are sequentially delayed pixel by pixel by the pixel delay units 1121 to 1124. The pixel signals of the pixels P51 to P55 are output from the output terminals 1125 to 1129. Pixel signals of the pixels P41 to P45 are output from the output terminals 1130 to 1134. The pixel signals of the pixels P31 to P35 are output from the output terminals 1135 to 1139. From the output terminals 1140 to 1144, pixel signals of the pixels P21 to P25 are output. The pixel signals of the pixels P11 to P15 are output from the output terminals 1145 to 1149. The pixel signal output from the output terminal 1137 is the pixel signal of the target pixel P33.

  Returning to FIG. 2, the pixel signal of the target pixel P <b> 33 output from the in-region pixel signal generation unit 11 is supplied to 24 differentiators 12 a to 12 x and the adder 17. The differentiators 12a to 12x are supplied with pixel signals other than the pixel signal of the pixel of interest P33 output from the in-region pixel signal generator 11. The differentiators 12a to 12x take the difference between the pixel signal of the pixel of interest P33 and the pixel signals of the other 24 pixels, and supply positive or negative difference values to the nonlinear calculators 13a to 13x.

  As shown in FIG. 5, the nonlinear arithmetic units 13a to 13x output the value of the input signal as it is when the input signal is in the range from 0 to the threshold values a and -a, and the value twice the threshold value a. In the range up to 2a, a value obtained by subtracting the value of the input signal from the value 2a is output, and in the range from the threshold value -a to -2a which is twice that value, a value obtained by adding the value of the input signal to the value -2a is output. When the value is 2a or more and the value -2a or less, 0 is output. The cumulative addition unit 14 adds all the output signals of the nonlinear calculation units 13a to 13x. In FIG. 2, for the sake of convenience, reference numerals P11, P12, P33, P54, and P55 are attached as reference numerals indicating pixel signals from the respective pixels.

  Here, the characteristics of the nonlinear processing in the nonlinear computing units 13a to 13x are triangular in the positive and negative directions as shown in FIG. 5, but may be trapezoidal. The characteristics of the non-linear processing are not limited to those shown in FIG. 5, and the input value is not within the range between the positive first value (a) and the negative second value (−a). When the input value exceeds the first value (a) and is less than or equal to the positive third value (2a), the input value is attenuated and the input value exceeds the second value. If the input value is greater than or equal to the negative fourth value (−2a), a value obtained by amplifying the input value is output, and if the input value exceeds the third and fourth values, 0 may be output. .

  The positive or negative output of the cumulative adder 14 is limited by the limiter 15 and input to the averaging unit 16. The limiting characteristics in the limiter 15 will be described later. The averaging unit 16 calculates an average value by dividing the output of the limiter 15 by 24, which is the cumulative addition number in the cumulative addition unit 14. The adder 17 adds the pixel signal of the target pixel P33 and the average value output from the average value unit 16, and outputs the result.

  With the configuration shown in FIG. 2, the nonlinear LPF 1 can be subjected to a low-pass filter after removing an edge component, and the influence on the edge portion can be suppressed to a small level. In particular, in the non-linear operation units 13a to 13x, the value larger than the absolute value of the values 2a and -2a is set to 0 so that the low-pass filter is not applied. it can. Also, since the absolute value of the threshold values a and -a is the upper limit of the input signal to the cumulative adder 14, the difference values themselves from the differentiators 12a to 12x are cumulatively added, and the average value is obtained by the averaging unit 16. Compared with the case of making it, the numerical value handled becomes small. Therefore, the circuit scale when the nonlinear LPF 1 is configured by hardware is reduced.

  Here, the reason why the limiter 15 is provided and the limiting characteristics of the limiter 15 will be described. As described above, the nonlinear LPF 1 can suppress the influence on a clear edge component. However, if the limiter 15 is not provided, the following problem may occur. That is, the texture of skin wrinkles and freckles expressed by weak signals of about 4 or 5 (values of about 4/256, 5/256) from LSB in 8-bit signals, subtle like touch of painting There is a risk that the quality of the video will be lost due to the slight reduction of the expression. The limiter 15 is provided to avoid such a problem.

  As shown in FIG. 6, as an example, the limiter 15 outputs the value of the input signal as it is when the input signal is in the range from 0 to the threshold values b and -b, and the value 2b that is twice the threshold value b. The value obtained by subtracting the value of the input signal from the value 2b is output in the range up to 2, and the value obtained by adding the value of the input signal to the value -2b is output in the range from the threshold value -b to -2b which is twice that value. When the value is 2b or more and the value -2b or less, 0 is output. Here, the limiting characteristics of the limiter 15 are triangular in the positive and negative directions as shown in FIG. 6, but may be trapezoidal. The limiting characteristic of the limiter 15 is not limited to that shown in FIG. 6, and the input value is within the range between the positive fifth value (b) and the negative sixth value (−b). When the input value exceeds the fifth value and is less than or equal to the positive seventh value (2b), the input value is attenuated and the input value exceeds the sixth value and is negative. When the input value is greater than or equal to the eighth value (−2b), a value obtained by amplifying the input value is output, and when the input value exceeds the seventh and eighth values, 0 may be output.

  An example of an image pattern in the region of FIG. 3 and an output value of the cumulative addition unit 14 in each pattern will be described with reference to FIGS. FIG. 7A shows pixel values (luminance values) in the pixels P11 to P15, P21 to P25, P31 to P35, P41 to P45, and P51 to P55 of FIG. Are all pixel values 104. In this case, as shown in FIG. 7B, the pixel P33 is an isolated point and is in a weak edge state that is not a clear edge. The output value of the cumulative adder 14 in the case of FIG.

  FIG. 8A shows a case where the pixels P13, P23, P33, P43, and P53 have a pixel value of 100 and all others have a pixel value of 104. In this case, as shown in FIG. 8 (B), a linear weak edge is not a clear edge. The output value of the cumulative adder 14 in the case of FIG. 8 is 80, which is 4 × 20 + 0 × 4. FIG. 9A shows a case where the pixels P11, P12, P21, P22, P31, P32, P41, P42, P51, and P52 have the pixel value 101 and all others have the pixel value 100. In this case, as shown in FIG. 9B, a clear gradation portion having a difference in pixel value 1 between the left two columns and the right three columns. The output value of the cumulative adder 14 in the case of FIG. 9 is 10 as 1 × 10 + 0 × 14.

  As can be seen from FIGS. 7 to 9, when the output value of the cumulative addition unit 14 is relatively large, it can be determined that the target pixel P33 is not a gradation portion but a weak edge. The limiting characteristic of the limiter 15 shown in FIG. 6 may be an optimum characteristic based on the output value of the cumulative adder 14 in the weak edge and gradation part shown in FIGS. Specifically, when the output value of the cumulative addition unit 14 is a weak edge, the limiter 15 limits the input to the averaging unit 16 to an extremely small value or zero. Since the output value of the cumulative adder 14 varies depending on the number of pixels in the region calculated by the nonlinear LPF1, the value of the threshold value b and the value of the input signal with the output signal set to 0 (the value 2b in the example of FIG. 6). ) May be set to an optimum value according to the number of pixels in the region calculated by the nonlinear LPF 1.

  Here, a processing example of the signal waveform of the gradation part and the weak edge by the video signal processing apparatus of FIG. 1 will be described with reference to FIGS. FIG. 10 is an operation explanatory diagram in the case of the signal waveform of the gradation portion where noise is not superimposed. FIG. 10A shows one-dimensionally the signal waveform of the gradation portion in which the signal level increases step by step by 8 bits of LSB. When nonlinear LPF is applied to the signal waveform of FIG. 10A by nonlinear LPF1, FIG. 10B is obtained. When the signal waveform of FIG. 10B is changed to 8 bits by the bit precision conversion unit 3, FIG. 10C is obtained. When the difference between the signal waveforms of FIG. 10A and FIG. 10C is taken by the subtractor 4, it becomes 0 as shown in FIG. 10D. When the signal waveforms of FIGS. 10B and 10D are added by the adder 5, the signal waveform of FIG. 10E is obtained. Thus, in the gradation portion, the nonlinear LPF processing by the nonlinear LPF 1 is performed.

  Next, processing in the case of a weak edge signal waveform will be described with reference to FIGS. FIG. 11 is an operation explanatory diagram when the limiter processing by the limiter 15 is not performed, and FIG. 12 is an operation explanatory diagram of the present embodiment in which the limiter processing by the limiter 15 is performed. FIG. 11A and FIG. 12A one-dimensionally show a signal waveform of a linear weak edge. The difference in pixel value between the weak edge and the other part is a value of 4 in 8 bits. First, in the case of FIG. 11 in which the limiter process by the limiter 15 is not performed, if the nonlinear LPF is applied to the signal waveform in FIG. 11A without performing the limiter process by the limiter 15, FIG. When the signal waveform of FIG. 11B is converted to 8 bits by the bit precision conversion unit 3, FIG. 11C is obtained. When the difference between the signal waveforms of FIG. 11A and FIG. 11C is taken by the subtractor 4, a signal waveform as shown in FIG. 11D is obtained. When the signal waveforms of FIGS. 11B and 11D are added by the adder 5, the signal waveform of FIG. 11E is obtained.

  As can be seen from FIG. 11, in the configuration in which the limiter 15 is not provided, the output waveform of the adder 5 becomes distorted as shown in FIG.

  On the other hand, in the case of FIG. 12 in which the limiter process by the limiter 15 is performed, when the nonlinear LPF is applied to the signal waveform of FIG. When the signal waveform of FIG. 12B is converted to 8 bits by the bit precision conversion unit 3, FIG. 12C is obtained. When the difference between the signal waveforms of FIG. 12A and FIG. 12C is taken by the subtracter 4, it becomes 0 as shown in FIG. When the signal waveforms shown in FIGS. 12B and 12D are added by the adder 5, the signal waveform shown in FIG. 12E is obtained. In the case of a weak edge signal waveform, the cumulative addition value output from the cumulative addition unit 14 by the limiter 15 is set to 0, so that the output of the nonlinear LPF 1 maintains the original signal. Therefore, the output waveform of the adder 5 is not affected by the nonlinear LPF 1 as shown in FIG.

  FIG. 13 is an operation explanatory diagram in the case of a signal waveform of a gradation portion on which random noise is superimposed. FIG. 13A shows one-dimensionally the signal waveform of the gradation portion on which random noise is superimposed. When nonlinear LPF is applied to the signal waveform of FIG. 13A by nonlinear LPF1, FIG. 13B is obtained. When the signal waveform of FIG. 13B is changed to 8 bits by the bit precision conversion unit 3, FIG. 13C is obtained. When the difference between the signal waveforms of FIG. 13A and FIG. 13C is taken by the subtractor 4, a signal waveform as shown in FIG. 13D is obtained. When the signal waveforms shown in FIGS. 13B and 13D are added by the adder 5, the signal waveform shown in FIG. 13E is obtained.

  As can be seen from FIG. 13 (E), the signal shown in FIG. 13 (E) has a global slope (that is, a low frequency component) of 12-bit accuracy, and noise of 8-bit accuracy is included in this low frequency component. This is a signal in which components (that is, high frequency components) are superimposed. According to the present embodiment, since a weak high frequency component is maintained, a noise component such as random noise is maintained. If noise reduction processing is performed using, for example, a frame cyclic noise reduction device that performs noise reduction using noise randomness at the subsequent stage of the video signal processing device of the present embodiment, the signal shown in FIG. can get. However, even if such noise reduction processing is performed, the weak edge is preserved.

  In the present embodiment, as described above, the output of the nonlinear LPF 1 is a low-frequency component signal, and the output of the subtractor 4 is a high-frequency component signal. Both signals are separated and processed. Since a signal with strong randomness is canceled by the cumulative addition in the nonlinear LPF 1, even if a certain amount of noise is superimposed on the video signal, the bit accuracy can be effectively extended. In addition, since the high frequency component signal is added to the output signal of the nonlinear LPF 1 as it is, it is possible to hold a weak edge, and it is possible to hold the texture with little adverse effect of attenuating the weak texture. It is.

  In the inventions described in Patent Documents 1 and 2 described above, since the gradation part is detected and the bit is expanded, if the noise is superimposed on the gradation part and it is determined that the gradation part is not the gradation part, the bit accuracy is expanded. Not. Since an actual television signal often has noise superimposed thereon, the inventions described in Patent Documents 1 and 2 are poor in the effect of the bit precision extension processing. On the other hand, in this embodiment, the bit accuracy can be expanded without being significantly affected by noise as described above.

  The present invention is not limited to the embodiment described above, and various modifications can be made without departing from the scope of the present invention. The present invention may be configured by hardware or software. Further, both may be mixed.

It is a block diagram which shows one Embodiment of this invention. It is a block diagram which shows the specific structural example of the nonlinear low pass filter 1 of FIG. It is a figure which shows the example of the area | region of the pixel signal processed by one Embodiment of this invention. FIG. 2 is a block diagram illustrating a specific configuration example of an in-region pixel signal generation unit 11 in FIG. 1. It is a figure which shows the example of a characteristic of the nonlinear calculation in the nonlinear calculating parts 13a-13x of FIG. It is a figure which shows the example of the limiter characteristic in the limiter 15 of FIG. It is a figure which shows the example of the weak edge of an isolated point. It is a figure which shows the example of a linear weak edge. It is a figure which shows the example of a gradation part. It is operation | movement explanatory drawing in the case of the signal waveform of the gradation part in which noise is not superimposed. It is operation | movement explanatory drawing in the case of the signal waveform of the weak edge when not performing a limiter process. It is operation | movement explanatory drawing in the case of the signal waveform of the weak edge in the case of performing a limiter process. It is operation | movement explanatory drawing in the case of the signal waveform of the gradation part on which noise was superimposed.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Nonlinear low pass filter 2 Delay device 3 Bit precision conversion part 4 Subtractor 5,17 Adder 11 Intra-region pixel signal generation part 12a-12x Differentiator 13a-13x Nonlinear operation part 14 Cumulative addition part 15 Limiter 16 Average value part

Claims (4)

A non-linear low-pass filter that performs non-linear low-pass filter processing on a digital video signal having a first quantization bit number and outputs a low-frequency component signal having a second quantization bit number accuracy greater than the first quantization bit number When,
A bit precision conversion unit that converts the low-frequency component signal output from the nonlinear low-pass filter into the first quantization bit number precision;
The digital video signal and the output signal of the bit precision conversion unit are input in a state where their phases are aligned, and the first quantization is performed by subtracting the output signal of the bit precision conversion unit from the digital video signal. A subtractor that outputs high and low frequency component signals with bit number accuracy;
A video signal processing apparatus comprising: a first adder that adds and outputs the low frequency component signal output from the nonlinear low-pass filter and the high and low frequency component signal output from the subtractor. . The nonlinear low-pass filter is
A plurality of pixels for obtaining a difference between a pixel signal of the pixel of interest and a pixel signal of each of the plurality of peripheral pixels within a predetermined region including the pixel of interest located in the center and a plurality of peripheral pixels located around the pixel of interest; Differencer of
Each of the positive or negative difference values output from the plurality of differentiators is input, and an input value is output in a range between the positive first value and the negative second value. When the value exceeds the first value and is less than or equal to the positive third value, the input value is attenuated, and the input value exceeds the second value and exceeds the negative fourth value. A plurality of nonlinear arithmetic units that output an amplified input value when there is, and output 0 when the input value exceeds the third and fourth values;
A cumulative addition unit that adds all output values from the plurality of nonlinear arithmetic units;
An averaging unit that averages the output value from the cumulative addition unit by the number of the surrounding pixels;
The video signal processing apparatus according to claim 1, further comprising: a second adder that adds a pixel signal of the target pixel and an output signal of the averaging unit. Between the cumulative addition unit and the averaging unit, a limiter to which a positive or negative output value from the cumulative addition unit is input,
The limiter outputs an input value in a range between a positive fifth value and a negative sixth value, and the input value exceeds the fifth value and is equal to or less than a positive seventh value. When the input value is equal to the output value, the input value is attenuated. When the input value exceeds the sixth value and is equal to or greater than the negative eighth value, the input value is amplified and the input value is output. 3. The video signal processing apparatus according to claim 2, wherein 0 is output when the value exceeds the seventh and eighth values. A plurality of pixels for obtaining a difference between a pixel signal of the pixel of interest and a pixel signal of each of the plurality of peripheral pixels within a predetermined region including the pixel of interest located in the center and a plurality of peripheral pixels located around the pixel of interest; Differencer of
Each of the positive or negative difference values output from the plurality of differentiators is input, and an input value is output in a range between the positive first value and the negative second value. When the value exceeds the first value and is less than or equal to the positive third value, the input value is attenuated, and the input value exceeds the second value and exceeds the negative fourth value. A plurality of nonlinear arithmetic units that output an amplified input value when there is, and output 0 when the input value exceeds the third and fourth values;
A cumulative addition unit that adds all output values from the plurality of nonlinear arithmetic units;
A positive or negative output value from the cumulative addition unit is input, and an input value is output in a range between a positive fifth value and a negative sixth value, and the input value is the fifth value. When the input value exceeds the sixth value and is equal to or less than the positive seventh value, the input value is attenuated. When the input value exceeds the sixth value and is equal to or greater than the negative eighth value, the input value is output. A limiter that outputs a value obtained by amplifying the value and outputs 0 when the input value exceeds the seventh and eighth values;
An averaging unit that averages the output value from the limiter by the number of the surrounding pixels;
A non-linear low-pass filter comprising: an adder that adds a pixel signal of the target pixel and an output signal of the averaging unit.

Images