JP4266227B2 - Video signal processing device - Google Patents

Description

  The present invention relates to a moving image signal processing apparatus, and in particular, a moving image encoding apparatus that compresses and encodes a moving image signal with high efficiency and a moving image decoding that decodes the compression-coded signal to reproduce the original moving image signal. The present invention relates to a moving picture signal processing apparatus suitable for removing coding distortion caused by compression coding and removing noise of a moving picture signal prior to coding.

  In systems that transmit / store moving image signals, such as TV telephones, TV conference systems, portable information terminals, digital video disk systems, and digital TV broadcasting systems, moving images are compressed and encoded to a small amount of information. The code sequence is transmitted / accumulated to the transmission path / accumulation medium, and the transmitted / accumulated code sequence is decoded to reproduce the original moving image signal.

  As a compression encoding technique of a moving image signal applied to such a system, various methods such as a method such as motion compensation, discrete cosine transform (DCT), subband encoding, pyramid encoding, and a combination thereof are available. Has been developed. In addition, ISO / MPEG1, MPEG2, ITU-T / H. 261, H.H. 262 is defined. All of these coding methods use motion compensated adaptive predictive cosine transform coding. Reference 1: edited by Hiroshi Yasuda, “International Standard for Multimedia Coding”, Maruzen, (June 1991), etc. Details.

  If a moving image signal can be compressed and encoded to a small amount of information using a compression coding technique, for example, a rate of several kbps to several tens of kbps, then the moving image is transmitted through a wireless telephone line such as an analog telephone line, cellular phone, or PHS. Thus, moving image communication is possible between a TV phone device, a portable information terminal, a personal computer, and the like.

  By the way, when transmitting a moving image signal that has undergone motion compensation predictive coding on a transmission path in which many transmission path errors are mixed, such as a wireless transmission path, a system called refresh may be used to increase error tolerance. Refreshing is an operation of encoding a part of a moving image signal using intraframe encoding in which the influence of errors is not transmitted to subsequent frames. However, when refreshing, the refreshed part, that is, the part that has been subjected to intraframe coding, and the other part that has undergone motion compensation predictive coding have different encoding methods, so the properties of the decoded image are different. A sense of incongruity occurs between the other parts, and the subjective quality of the decoded image is lowered.

  On the other hand, as a method for improving the prediction efficiency of motion compensation prediction, there is a method called patch motion compensation in which a motion vector is interpolated for each pixel. However, while patch motion compensation has the advantage of being able to express smooth deformation of an object efficiently, motion compensation is not performed in areas where the motion changes greatly, such as the boundary between the stationary background and the moving object. It cannot be performed efficiently, and a large coding distortion occurs.

  Also, if noise is mixed in the moving image signal to be encoded due to a camera input problem or flickering occurs, the encoding efficiency is greatly reduced, resulting in a decrease in encoding quality. This problem is particularly noticeable when the encoding bit rate is low. As a method for solving this problem, there is a method of applying a band limiting filter to an input image signal prior to encoding. However, in this method, not only noise but also a part of the original signal component is lost, so that the encoded image is blurred.

  Furthermore, due to camera input problems, the pixel values such as black level pixels are very low and the surrounding pixels and pixel values are large at the image edge of the video signal to be encoded, that is, at the edge of the screen. Different pixels may exist in a strip shape. If such low-level pixels exist, the accuracy of motion vector detection near the edge of the image decreases, leading to a decrease in encoding efficiency, or the low-level pixels spread around due to motion compensation. The quality of the converted image is degraded.

  As described above, in a moving image encoding / decoding device that uses refresh encoding such as intraframe encoding in order to increase resistance to transmission path errors, a portion that has been subjected to refresh encoding and other portions are used. There is a problem that the subjective quality of the decoded image is deteriorated because the properties of the decoded image are different.

  In addition, a moving image encoding / decoding device that performs patch motion compensation has a problem in that large encoding distortion occurs in a portion where the motion greatly changes. Further, in the conventional moving image encoding / decoding device, there is a problem that the encoding efficiency decreases when noise or flicker exists in the input moving image signal, and the encoding quality decreases when low level pixels exist at the image end. was there.

  The present invention has been made to solve the above-described problems of the conventional moving image encoding / decoding device. The decoding distortion of the decoded moving image signal and the encoding of the moving image signal before encoding are not possible. It is an object of the present invention to provide a moving image signal processing apparatus that can effectively remove noise, flicker, and the like.

  In order to solve the above-described problems, a moving image signal processing apparatus according to the present invention includes a time filter unit that performs filtering in the time axis direction according to a predetermined filter coefficient on an input moving image signal and outputs an output moving image signal. A filter coefficient determining means for determining a filter coefficient in the time filter means according to at least a difference between a signal obtained by delaying an input moving image signal and a signal obtained by delaying an output moving image signal or a signal obtained by delaying an input moving image signal; It is characterized by having.

  Further, when the moving image signal processing apparatus uses the decoded moving image signal decoded by the moving image decoding apparatus from the encoded data obtained by the moving image encoding apparatus having a plurality of encoding modes as the input moving image signal, In the filter coefficient determination means, at least the filter coefficient in the time filter means is (a) the magnitude of the difference between the input video signal and the delayed output video signal or the delayed input video signal, and (b ) Determined according to the coding mode of the moving picture coding apparatus.

  In this moving image signal processing apparatus, for example, a relatively strong time filtering is applied to a portion where the difference between the input moving image signal and the output moving image signal or a signal obtained by delaying the input moving image signal is small, that is, a stationary portion. By determining the filter coefficient so that the output signal is output as an output moving image signal, coding distortion that is particularly noticeable in a stationary portion is effectively reduced. In addition, in the portion where the difference is large, the decoded moving image signal is left as it is or the filter coefficient is determined so as to be output as an output moving image signal after performing weak temporal filtering, so that the image is changed by a scene change or the like. Even when there is a large change, it is possible to prevent a large change in the image from remaining as an afterimage.

  Further, when the filter coefficient is further determined in consideration of the encoding mode, for example, the refresh coefficient is set by setting the filter coefficient so that the time filtering is relatively strong in the portion where the refresh encoding is performed. Of the decoded image due to the difference in the characteristics of the decoded image due to the difference in the encoding method between the encoding mode of the part (intra-frame encoding mode) and the encoding mode of the other part (motion compensation prediction encoding mode). Naturalness is eliminated.

  In addition, the filter coefficients may be determined taking into account the activity, i.e. the evaluation value indicating the complexity of the image, in which case, for example, the temporal filtering may be weakened or input at high edges of the activity or fine picture parts. By outputting the moving image signal as it is, the blurring of the image caused by temporal filtering is reduced, and the flat part where the activity is low and the encoding distortion is conspicuous is applied to the flat part strongly to suppress the encoding distortion effectively. it can.

  Furthermore, in the present invention, after performing low-level pixel signal removal processing for performing processing for removing pixel signals of low-level pixels whose pixel value existing at the image edge is equal to or less than a threshold value from the input moving image signal, The time filtering process and the filter coefficient determination may be performed.

  In this way, by removing pixel signals of low level pixels such as black level pixels existing at the edge of the image caused by a camera input problem or the like from the input moving image signal before compression encoding, encoding efficiency is improved. As a result, the coding quality is improved.

  Further, according to the present invention, the low-level pixel signal removal processing for removing the pixel signal of the low-level pixel whose pixel value existing at the image edge is equal to or less than the threshold value from such an input video signal is the input video signal. Based on the number of low-level pixels in which the low-level pixel whose pixel value is equal to or less than the threshold is detected, the number of low-level pixels is searched for at the edge of the image, and the pixel value of this low-level pixel is searched This is realized by replacing with a pixel value other than the determined low-level pixel.

  According to the image signal processing apparatus of the present invention, it is possible to remove coding distortion caused by compression coding of a moving image signal and improve the quality of a decoded image. In addition, noise, flicker, black edges, and the like existing in a moving image signal to be compression-encoded can be removed to increase the efficiency of compression-encoding and improve encoding quality.

  Embodiments of the present invention will be described below.

(First Embodiment) FIG. 1 is a block diagram showing a configuration of a coding distortion removing apparatus which is a moving image signal processing apparatus according to a first embodiment of the present invention. This coding distortion removing apparatus receives a decoded moving picture signal 121 from a moving picture decoding apparatus as an input moving picture signal, and for removing the coding distortion by the cyclic adaptive time filter 100 with respect to the decoded moving picture signal 121. Then, a filtering process in the time axis direction (hereinafter referred to as a time filtering process) is performed, and an output moving image signal 123 is output.

  The time filter 100 includes a first multiplier 101, an adder 102, a frame delay unit 103, and a second multiplier 104. Further, the filter coefficient of the time filter 100 (hereinafter referred to as a weight coefficient) is determined by the weight coefficient determination unit 105. The weighting factor determination unit 105 includes an interframe difference calculation unit 106 and a weighting factor calculation unit 107.

  Next, the operation of the coding distortion removing apparatus according to the present embodiment will be described. The input decoded video signal 121 is input to the adder 102 after being multiplied by the first weighting coefficient 125 output from the weighting coefficient determining unit 105 in the first multiplier 101. An output signal from the adder 102 is output to the outside of the apparatus as an output moving image signal 123 and also input to the frame delay unit 103. The frame delay unit 103 outputs a signal 124 obtained by delaying the output moving image signal 123 by one frame time. The output signal 124 from the frame delay unit 103 is multiplied by the second weighting factor 126 output from the weighting factor determining unit 105 by the second multiplier 104. The output signal from the second multiplier 104 is added to the output signal from the first multiplier 101 by the adder 102.

Based on the decoded moving image signal 121 and the output moving image signal 124 one frame before, the weighting factor determination unit 105 determines the weighting factors 125 and 126 for each pixel as follows. First, the inter-frame difference calculation unit 106 calculates the magnitude D of the difference between the decoded moving image signal 121 and the output moving image signal 124 of the previous frame using ± M pixels surrounding the processing target pixel according to the following equation.

  Then, the weighting coefficient calculation unit 107 calculates and determines the weighting coefficient from the difference size D thus calculated. FIG. 2 shows an example of the relationship between the difference magnitude D and the weighting factor. In the figure, W on the vertical axis is the second weighting coefficient 126, and the first weighting coefficient 125 is (1-W). Further, ath1, ath2, and WL are predetermined constants.

  In this way, the weighting factor determination unit 105 adaptively determines the weighting factors 125 and 126 according to the magnitude D of the difference between the decoded moving image signal 121 and the output moving image signal 124 one frame before. That is, when the magnitude D of the difference is small, W is determined to be large. As a result, in the still part where coding distortion is conspicuous, the decoded moving image signal 121 to which the coding distortion is added is not output as it is, but the time filter 100 outputs the previous frame for the decoded moving image signal 121. A signal subjected to temporal filtering processing by weighted addition with the moving image signal 124 is output as the output moving image signal 124, thereby reducing coding distortion.

  On the other hand, when the difference magnitude D is equal to or greater than the threshold value ath2, W = 0, and the decoded video signal 121 is output as it is as the output video signal 123. For this reason, even if the image changes greatly from the previous frame due to a scene change or the like, a large change in the image does not remain as an afterimage.

  In the present embodiment, the process in units of pixels has been described, but the process may be performed in units of blocks.

(Second Embodiment) FIG. 3 is a block diagram showing the configuration of a coding distortion removing apparatus which is a moving image signal processing apparatus according to a second embodiment of the present invention, and shows a difference size D and a weighting factor. This is an example in which the relationship of W is switched according to the encoding mode of the video encoding device, the size of the motion vector obtained by the video encoding device, and the like. Therefore, in this embodiment, a signal (referred to as a side information signal) 128 indicating a coding mode and a motion vector is a weight coefficient calculation unit in the filter coefficient determination unit 105 separately from the decoded video signal 121 from the video decoding device. 107 is input.

  FIG. 4 shows the difference in the difference between the portion where refresh coding is performed and the other portion when refresh coding is performed by the moving image coding apparatus in order to suppress the propagation of image quality degradation caused by transmission path errors. In this example, the relationship between the length D and the weighting factor W is switched. The 401 characteristic is used for the part subjected to refresh encoding, and the 402 characteristic is used for the other parts.

  The characteristic 401 in which the weight coefficient W is likely to be large in the portion subjected to the refresh encoding is that the encoding mode of the portion subjected to the refresh encoding is the intra-frame encoding mode, and the encoding mode of the other portion. This is because the nature of the decoded image is different because the encoding method is different from (for example, motion compensation prediction encoding mode), and unnaturalness is likely to occur in the decoded image. In particular, in a still part where coding distortion is conspicuous or a part where the motion is small, when the part that was in the normal motion compensation prediction coding mode until the previous frame is changed to the intra-frame coding mode that is the refresh coding mode, the decoded image The change in the property of the image becomes visually noticeable, and the subjective quality of the decoded image is likely to deteriorate.

  On the other hand, if the characteristic 401 in FIG. 4 is used, the weighting coefficient W, that is, the weighting coefficient 126 for the output moving image signal 124 one frame before becomes large in the still part where the difference D is small or the part where the motion is small. . Accordingly, the portion of the decoded moving image signal 121 whose property has been changed by refreshing is subjected to weighted addition with the output moving image signal 124 one frame before, and the time filtering process is strongly applied, so that the image quality deterioration can be suppressed to a small level. Become.

  FIG. 5 shows another example in which the relationship between the difference magnitude D and the weighting factor W is switched in the present embodiment. In the motion compensation predictive coding mode, the characteristic 501 in FIG. This is to suppress coding distortion caused by motion compensation of the stationary part by the motion vector of the moving object at the boundary part between the stationary part and the moving object, and is likely to cause large distortion at the boundary part of the object. This is effective in reducing the coding distortion of patch motion compensation.

  On the other hand, since such distortion does not occur in the intra-frame coding mode, the characteristic 502 in which the weight coefficient W is reduced is used.

(Third Embodiment) FIG. 6 is a block diagram showing the configuration of a coding distortion removing apparatus which is a moving image signal processing apparatus according to a third embodiment of the present invention. In this example, the relationship of W is changed according to the activity indicating the complexity of the image. For this purpose, in the present embodiment, the activity calculation unit 108 calculates the activity of the decoded moving image signal 121 in the filter coefficient determination unit 105, and the threshold value calculation unit 109 generates the threshold value ath1 in FIG. And ath2 are changed.

First, in the activity calculation unit 108 and the inter-frame difference calculation unit 106, the magnitude 127 of the difference between the activity 131 of the decoded moving image signal 121 and the output moving image signal 124 one frame before is expressed by ± M pixels around the processing target pixel. Calculate according to the following formula.

  Next, the threshold value calculation unit 109 determines the first threshold value 132 (ath1) and the second threshold value 133 (ath2) from the activity 221 based on the characteristics shown in FIGS. To do. In FIG. 7, th1, th2, and α are predetermined constants. The characteristics shown in FIGS. 7A and 7B are expressed as follows.

ath1 = th1-α · act · th1 / th2ath2 = th2-α · act
Then, the weighting factor calculation unit 107 determines the weighting factors 125 and 126 according to the characteristics shown in FIG. 2 from the magnitude 127 of the difference and the first and second threshold values 132 and 133.

  In this way, by changing the characteristics of the time filter 100 in accordance with the activity, the threshold values ath1 and ath2 are reduced and the weighting factor W is reduced in a portion where the activity is high, that is, an edge or a fine pattern portion. As a result, the strength of temporal filtering in the temporal filter 100 is weakened, and blurring of the image caused by the temporal filter 100 is suppressed to a small level.

  On the other hand, since a flat portion where coding distortion is conspicuous has low activity, the weighting factor W tends to be large and temporal filtering is strongly applied, so that the coding distortion suppression effect becomes large.

  (Fourth Embodiment) FIG. 8 is a block diagram showing the configuration of a coding distortion removing apparatus which is a moving image signal processing apparatus according to a fourth embodiment of the present invention. When the same reference numerals are assigned to the parts corresponding to those in FIG. 1 and the difference from the first embodiment will be mainly described, the spatial filter 110 is newly added in this embodiment, and the output signal of the adder 102 is The difference is that after the encoding distortion removal processing in the spatial direction is performed by the spatial filter 110, it is extracted as the output moving image signal 123.

  The spatial filter 110 removes coding distortion by performing a filtering process in the spatial direction. Specifically, for example, the strength of the filter is adaptively changed according to the difference between adjacent pixels. The part where the inter-pixel difference is small strengthens the filter, and the part where the inter-pixel difference is large weakens the filter. Thereby, it is possible to remove coding distortion such as mosquito distortion while preserving edges.

  As described above, in the present embodiment, the temporal displacement of the decoded moving image signal 121 and the time of one frame is achieved by combining the temporal distortion correction by the temporal filter 100 with the spatial distortion removal by the spatial filter. The afterimage that rarely occurs due to the weighted addition processing of the output moving image signal 124 having a certain level can be removed, and the quality of the output moving image signal 123 is further improved as compared with the case of using only the time filter.

  (Fifth Embodiment) FIG. 9 is a block diagram showing the configuration of a coding distortion removing apparatus which is a moving image signal processing apparatus according to a fifth embodiment of the present invention. This coding distortion removing apparatus inputs a decoded moving picture signal 221 of the current frame input from the moving picture decoding apparatus and a decoded moving picture signal 224 of the previous frame stored in the frame memory of the moving picture decoding apparatus. Then, the non-recursive adaptive time filter 200 performs coding distortion removal processing and outputs an output moving image signal 223.

  The time filter 200 includes a first multiplier 201, an adder 202, and a second multiplier 204. Further, the filter coefficient (referred to as a weighting coefficient) of the time filter 200 is determined by the weighting coefficient determination unit 205.

  The input decoded video signal 221 of the current frame and the decoded video signal 228 of the previous frame are input to the first and second weights 201 and 204 from the weight coefficient determination unit 205, respectively. After the coefficients are multiplied, they are input to the adder 202 and added together, and the output signal of the adder 202 is output as an output moving image signal 223 to the outside of the apparatus.

The weighting factor determination unit 205 determines a weighting factor in the same manner as the weighting factor determination unit 105 in any of the first to fourth embodiments. However, the interframe difference D is calculated as a difference between the decoded moving image signal 221 and the decoded moving image signal 224 one frame before according to the following equation.

  As described above, in the coding distortion removal apparatus of the present embodiment, the coding distortion removal processing is performed by the non-cyclic time filter 200 on the cyclic adaptive time filter 100 used in the first to fourth embodiments. . For this reason, although the encoding distortion removal performance is slightly reduced as compared with the first to fourth embodiments, the output moving image signal of the previous frame and the decoded video of the current frame as in the first to fourth embodiments. The afterimage produced by weighted addition with the image signal is reduced. In the present embodiment, since the decoded video signal 224 one frame before is taken out from the video decoding device, it is not necessary to provide a frame delay in the coding distortion removal device, and the amount of hardware is reduced.

  However, depending on the configuration of the video decoding device, the decoded video signal 224 one frame before may not be output. In this case, as shown in FIG. 9, a frame delay unit 203 for delaying the decoded video signal 221 of the current frame by one frame time is provided in the encoding distortion removing apparatus, and the decoded video signal 224 of the previous frame is obtained. It only has to be generated.

  (Sixth Embodiment) FIG. 10 is a block diagram showing the configuration of a video signal processing apparatus according to the sixth embodiment of the present invention. Unlike the first to fifth embodiments, FIG. The apparatus suppresses noise, flicker, and the like prior to compression encoding of a moving image signal.

  The input moving image signal 321 is first input to the image black edge removal processing unit 306, and is a black level pixel existing at the image edge caused by a camera input problem or the like, that is, a low level whose pixel value is equal to or less than a predetermined threshold value. Processing for removing the pixel signal of the pixel from the moving image signal 321 is performed.

  FIG. 11 is a diagram for explaining the image black edge removal processing, and shows an example of processing for removing black level pixels existing at the left end of the image. In FIG. 11, black circles represent black level pixels, and white circles represent pixels other than the black level.

  [Step 1] First, black level pixels are detected in order from the left end of the image. Whether the pixel is a black level pixel may be determined based on whether the pixel value is equal to or less than a predetermined threshold value.

  [Step 2] Next, a search is performed for each horizontal line to determine how many black level pixels are continuously counted from the left end of the image. The number of black level pixels from the left edge of the image searched in this way is Nsb.

  [Step 3] Then, the pixel signal of the black level pixel is replaced with the pixel signal of the pixel Nf (arbitrary integer) right of the pixel signal. The above processing will be described with an example of the uppermost horizontal line in FIG. 11. Since there are two black level pixels 1101 and 1102 from the left end of the image, the black level pixel count Nsb is 2. If Nf = 2, Nsb + Nf−1 pixels 1101 to 1103 from the left end of the image are replaced with the pixel value of the Nsb-th pixel 1104 from the left end of the image.

  When black level pixels are also present at the top, right, and bottom edges of the image, the processing for replacing the pixel signal of the black level pixel with the pixel signal of another pixel is performed in the same manner as the processing for the pixel existing at the left end of the image. Just do it.

  When a large black level pixel value exists at the edge of the image, if the above pixel replacement process is performed on the black level pixel value, an uncomfortable feeling may occur in the image. In order to prevent this, the pixel replacement process may not be performed when the black level pixel count Nsb is equal to or greater than a certain value.

  Returning to FIG. 10, the moving image signal after the processing for removing the pixel signal of the black level pixel existing at the image end by the black edge removal processing unit 306 is performed by the cyclic adaptive time filter 300 such as noise and flicker. Is removed. The cyclic adaptive time filter 300 and the weight coefficient determination unit 305 that determines the filter coefficient (weight coefficient) have the same configuration as that described in the first to fourth embodiments. For example, the multiplier 302, the frame delay unit 303, and the multiplier 304 perform the same operations as the multiplier 101, the adder 102, the frame delay unit 103, and the multiplier 104 in FIG. The determination of the weighting coefficient in the weighting coefficient unit 305 is performed in the same manner as in the first to fourth embodiments. However, the constants such as WL, th1, th2, and α described above are the input moving image signal 321 or the image. It is assumed that it is determined according to the nature of the moving image signal that has passed through the black edge removal 306.

In the above embodiment, the inter-frame difference D and the activity act are calculated using the absolute value sum. However, the calculation is performed using the sum of squares, the weighted absolute value sum, and the weighted square sum as in the following equation. May be.

Here, wa and wD are coefficients for weighting according to k and l.

1 is a block diagram showing a configuration of a moving image signal processing apparatus according to a first embodiment of the present invention. The figure which shows the relationship between the difference between frames in the embodiment, and a weighting coefficient. The block diagram which shows the structure of the moving image signal processing apparatus which concerns on the 2nd Embodiment of this invention. The figure which shows the 1st example of the relationship between the difference between frames in the same embodiment, and a weighting coefficient. The figure which shows the 2nd example of the relationship between the difference between frames in the same embodiment, and a weighting coefficient. The block diagram which shows the structure of the moving image signal processing apparatus which concerns on the 3rd Embodiment of this invention. The figure which shows the relationship between the activity and threshold value in the same embodiment The block diagram which shows the structure of the moving image signal processing apparatus which concerns on the 4th Embodiment of this invention. The block diagram which shows the structure of the moving image signal processing apparatus which concerns on the 5th Embodiment of this invention. The block diagram which shows the structure of the moving image signal processing apparatus which concerns on the 6th Embodiment of this invention. The figure for demonstrating operation | movement of the moving image signal processing apparatus which concerns on the same embodiment

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 ... Cyclic adaptive time filter 101 ... Multiplier 102 ... Adder 103 ... Frame delay device 104 ... Multiplier 105 ... Filter coefficient determination part 106 ... Inter-frame difference calculation part 107 ... Weight coefficient calculation part 108 ... Activity calculation part 109 ... Threshold calculation unit 110 ... spatial filter 121 ... decoded video signal 123 ... output video signal 124 ... output video signal 125, 126 one frame before ... weight coefficient 127 ... inter-frame difference 128 ... side information signal 131 ... activity 132, 133 ... threshold 200 ... acyclic adaptive time filter 201 ... multiplier 202 ... adder 203 ... frame delay device 204 ... multiplier 205 ... filter coefficient determination unit 300 ... cyclic adaptive time filter 301 ... multiplier 302 ... adder 303 ... frame delay 304 ... multiplier 305 ... filter Number determining unit 306 ... image black edge removing unit 321 ... input moving image signal 323 ... output moving image signal 324 ... output moving image signals 325 and 326 one frame before ... weighting factor 401 ... interframe difference and weighting factor in refresh mode Relationship 402: Inter-frame difference and weight coefficient relationship 501 in modes other than refresh 501 ... Inter-frame difference and weight coefficient relationship 502 in motion compensation mode 502 ... Inter-frame difference and weight coefficient relationship in intra-frame coding mode

Claims (3)

A moving image signal for processing an input moving image signal using a decoded moving image signal decoded by a moving image decoding device from encoded data obtained by a moving image encoding device having a plurality of encoding modes as an input moving image signal In the processing device,
Time filter means for performing filtering processing in a time axis direction according to a predetermined filter coefficient to the input moving image signal and outputting an output moving image signal;
Filter coefficient determining means for determining a filter coefficient in the time filter means in accordance with an encoding mode of the video encoding device;
And the filter coefficient determination means sets the filter coefficient so that the filter coefficient becomes large in the motion compensation prediction coding mode and becomes small in the intra-frame coding mode. Processing equipment. The filter coefficient determination unit further determines the filter coefficient in the time filter unit according to the magnitude of a difference between an input moving image signal and a signal obtained by delaying the output moving image signal or a signal obtained by delaying the input moving image signal. The moving image signal processing device according to claim 1, wherein The moving image signal processing apparatus according to claim 1, wherein the filter coefficient determining unit determines the filter coefficient in consideration of an evaluation value indicating complexity of the input moving image signal.

Images