JP2007336075A - Block distortion reducing device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve problems that conventionally, an image having small block distortion is erroneously determined to have block noise at its image outline part etc., and then unnecessary filtering is performed and that countermeasures for block distortion caused in a local place in a picture of one frame are not sufficient. <P>SOLUTION: Block distortion quantity detectors 21<SB>1</SB>to 21<SB>m</SB>obtain block distortion quantity detection results by measuring independently, the number of block boundaries where block distortion having distortion intensity larger than a certain value is caused among all block boundaries in allocated areas among respective (m) divided areas of one picture of decoded image data. A filter switch 23 selects one of filters 24<SB>1</SB>to 24<SB>m</SB>based upon detection levels, detected in pixel units, from block distortion detectors 22 in the pixel units and the block distortion quantity detection results, detected by the areas, from the block distortion quantity detectors 21<SB>1</SB>to 21<SB>m</SB>, and supplies image data decoded by a decoder 11. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a block distortion reducing apparatus, and more particularly to a block generated at a block boundary of decoded image data when encoded image data obtained by compressing and encoding an image by a block unit encoding method is decoded. The present invention relates to a block distortion reduction device that reduces distortion (block noise).

  In image compression coding such as the MPEG (Moving Picture Experts Group) standard, image data is divided into block units of a predetermined size composed of a plurality of predetermined pixels adjacent in the horizontal and vertical directions, and within the divided blocks. A method (block unit coding method) in which compression coding is performed by performing processing such as discrete cosine transform (DCT) in units of blocks using the high correlation between adjacent pixels of Yes.

  In the case of image data encoded by this block unit encoding method, when the compression rate is increased in order to reduce the amount of information, the boundary between adjacent blocks due to quantization errors is added to the decoded image data. Block distortion (block noise), in which (block boundary) appears as distortion (tone difference), is likely to occur. Therefore, in order to reduce the block distortion, a block distortion reduction apparatus that conventionally applies a noise reduction filter to a block boundary where block distortion occurs has been proposed (see, for example, Patent Document 1 and Patent Document 2).

  FIG. 9 shows a block diagram of an example of a conventional block distortion reducing apparatus described in Patent Document 1 described above. In the figure, a decoded video signal which is an input video signal is supplied to an isolated differential point extraction circuit 1 where it is differentiated into a differential signal, and then the differential value of the adjacent pixel is compared to protrude from the adjacent pixel. Only the differentiated value is extracted as an isolated differential point. The isolated differential data output from the isolated differential point extraction circuit 1 is supplied to the block noise detection circuit 2, where block noise (block distortion) is utilized by utilizing the fact that the isolated differential points have high correlation in the horizontal and vertical directions. ) Is detected and supplied to the switching terminal of the switch 4.

  When the block noise detection signal indicating that there is a lot of block noise is supplied from the block noise detection circuit 2, the switch 4 selectively outputs the isolated differential data output from the isolated differential point extraction circuit 1 to the filter 5, and in units of frames. When there is little block noise and no block noise detection signal is input, a 0 level signal is selectively output to the filter 5. The filter 5 outputs a correction signal for correcting the signal level difference at the boundary between the rectangular block in which block noise is generated and the adjacent block.

  On the other hand, the input video signal is delayed by a predetermined time delay circuit 3 for adjusting the timing of the output correction signal of the filter 5 and then supplied to the adder 6 where it is added to the output correction signal of the filter 5. As a result, when block noise is generated from the adder 6, the gradation difference at the block boundary between adjacent blocks is smoothed and the block noise is hardly noticeable and is output to the output terminal 7. Is done. This avoids filtering at positions other than block noise, and suppresses adverse effects.

  Here, the isolated differential data is selectively output from the switch 4 for a frame with a high block noise, and a 0 level signal is output from the switch 4 for a frame with a low block noise. Block noise reduction processing is performed for a large number of frames, and block noise reduction processing is turned off for frames with little block noise so as to suppress image quality deterioration associated with the block noise reduction processing.

  FIG. 10 shows a block diagram of an example of a conventional block distortion reducing apparatus described in Patent Document 2. In the figure, an MPEG bit stream input from an input terminal 10 is decoded by a decoder 11 and then supplied to a block distortion detection unit 12 in units of pixels. For each block boundary, between the block boundary and its neighboring pixels. The presence / absence of block distortion and the distortion intensity are detected in pixel units using the pixel difference value.

The decoded image data output from the decoder 11 is supplied to the filter switch 13, and the edge storage selected by the filter switch 13 according to the detection result of the pixel unit block distortion detected by the pixel unit block distortion detection unit 12 is stored. After being smoothed by one of the type smoothing filters 14 _{1 to} 14 _n , it is output to the output terminal 15. As described above, in this conventional block distortion reducing device, the plurality of filters 14 _{1 to} 14 _n are switched and applied in accordance with the presence or absence and the strength of the block distortion, so that the block distortion portion and the portion close to the block distortion portion are applied. Even in a situation where the contour portion exists, the block distortion is accurately reduced while leaving the contour portion.

JP 2001-119695 A Japanese Patent Laying-Open No. 2005-079617

  However, the conventional block distortion reducing device described in Patent Document 2 described above is erroneously determined as block noise in an image contour portion or the like in an image with little block noise (block distortion), and performs unnecessary filtering. There is a problem of end.

  On the other hand, the conventional block distortion reduction apparatus described in Patent Document 1 tries to solve unnecessary filtering due to erroneous determination of the block noise, but only controls on / off of the block noise reduction processing. Therefore, there is a problem that visual unnaturalness is given at the time of switching. Further, since the determination is made based on the block distortion amount in one whole frame, as shown in FIG. 11, it is possible to cope with block noise (block distortion) generated at a local location 51 in the screen of one frame. It is insufficient.

  The present invention has been made in view of the above points, and an object of the present invention is to provide a block distortion reduction device that can suppress erroneous determination and reduce visual unnaturalness in an image with little block distortion.

  Another object of the present invention is to provide a block distortion reduction apparatus that can appropriately reduce block distortion even with respect to block distortion that occurs at a local location in a screen.

  To achieve the above object, the present invention decodes block-unit encoded image data obtained by compressing and encoding image data in units of blocks having a predetermined number of pixels, and inputs the decoded image data. A block distortion reduction apparatus that receives image data and reduces block distortion that occurs at the boundary between adjacent blocks of input image data, and a plurality of filters having different smoothing intensity characteristics according to the smoothing intensity. A group of multiple filters, a pixel difference absolute value between two pixels across the block boundary of the input image data, and a pixel difference absolute value between adjacent pixels in the same block with two or more pixels near the block boundary The average value of the values is calculated, and the block distortion level is detected from the difference between the absolute value of the pixel difference and the average absolute value of the pixel difference between adjacent pixels. Block distortion detection unit and block distortion that has distortion intensity greater than or equal to a predetermined value among all block boundaries existing in the divided area when one screen of input image data is divided into multiple areas larger than the block Based on the number of block boundaries where the block distortion occurs, the block distortion amount detection means for each area for detecting the block distortion amount, and the block distortion amount for each divided area detected by the block distortion amount detection means for each area After selecting a set of filters having smoothing intensity characteristics from a plurality of filter sets, the filter of the selected filter selected according to the block distortion detection level detected by the block distortion detection means in pixel units is used. And switch means for selecting one filter from the set and supplying input image data to the one filter.

  In this invention, after selecting one filter set having a smoothing intensity characteristic according to the block distortion amount for each divided region detected by the block distortion amount detection means for each region from the plurality of filter sets, Since one filter is selected from the set of filters selected according to the block distortion detection level detected by the block distortion detection means in pixel units, and the input image data is supplied to the one filter, 1 Select a filter with the optimal smoothing strength characteristic according to the block distortion amount of the divided area of the input image data based on the distortion amount of each area obtained by dividing the screen into a plurality of areas instead of judging by the block distortion amount of the entire frame. Thus, block distortion of input image data can be reduced.

  Further, in the present invention, filter switching is performed not only by the presence and intensity of block distortion using pixel change values in the vicinity of the block boundary, but also by adding block distortion amount information to the judgment material. Therefore, a filter having the optimum smoothing intensity characteristic can be selected from a plurality of filters.

  According to the present invention, the optimum smoothing according to the block distortion amount of the divided area of the input image data is not performed based on the block distortion amount of one whole frame but the distortion amount of each area obtained by dividing the screen into a plurality of areas. A filter with strength characteristics is selected to reduce block distortion in the input image data, so that an erroneous decision is made on an image with little block distortion without affecting the distortion reduction effect in an image with much block distortion. In addition, it is possible to obtain an effect of reducing visual unnaturalness and maintaining a reduction effect on local block distortion.

  In addition, according to the present invention, filter switching is performed not only by the presence and intensity of block distortion using pixel change values in the vicinity of the block boundary but also by adding block distortion amount information to the determination material. By performing the selection, a filter having the optimum smoothing intensity characteristic can be selected and applied from the plurality of filters, and the optimum block distortion reduction process can be performed on the decoded image data.

  Next, each embodiment of the present invention will be described with reference to the drawings.

(First embodiment)
FIG. 1 shows a block diagram of a first embodiment of a block distortion reducing apparatus according to the present invention. In the figure, the same components as those in FIG. In FIG. 1, a decoder 11 receives block-unit encoded data (for example, MPEG bit stream) input via an input terminal 10 as input, decodes it, and outputs image data. The block distortion amount detection units 21 _{1 to} 21 _m (m is an arbitrary natural number of 2 or more), the block distortion detection unit 22 in units of pixels, and the filter switch 23 are image data decoded by the decoder 11. Is entered. Decoded image data is input to only one filter selected by the filter switch 23 in n (n is an arbitrary natural number of 2 or more) filters 24 _{1 to} 24 _n .

Here, in the present embodiment, as shown in FIG. 2, it is assumed that one frame screen is divided into m areas, and the amount of block distortion is determined for each of the divided areas 1 to m. The block distortion amount detectors 21 _{1 to} 21 _{m shown} in FIG. Block distortion amount information detected for each region by the block distortion amount detection units 21 _{1 to} 21 _m is supplied to the filter switch 23.

The filter switch 23 includes block distortion amount information detected for each region by the block distortion amount detection units 21 _{1 to} 21 _m and block distortion amount information detected for each pixel by the block distortion detection unit 22 for each pixel. Is selected as a selection control signal, a selection operation is performed with priority given to block distortion amount information detected in units of areas, which will be described later, and decoded image data input to the common terminal is selected from among the _n filters 24 _{1 to} 24 _n Select and output to one specified filter.

When a general low-pass filter (LPF) is used for the noise reduction processing as the filters 24 _{1 to} 24 _n , there arises a problem that the erroneously detected edge is rounded and the resolution is lowered. Therefore, in the present embodiment, this problem is solved by using an edge-preserving smoothing filter as the above-described filters 24 _{1 to} 24 _n . The filters 24 _{1 to} 24 _n are edge-preserving smoothing filters for smoothing the pixel values around the block distortion, but have different smoothing strength characteristics.

For example, the filter 24 ₁ has the strongest smoothing intensity characteristic, and the number of pixels whose value is changed is the largest among the pixels around the block distortion position. The order of the filters 24 ₂ , 24 ₃ ,. strength of in smoothing gradually weakened, the strength characteristics of the filter 24 _n is smoothed weakest, there is a characteristic that left and right pixel values only near the position of the block distortion is changed. Accordingly, the values of pixels in a wide range of positions are changed so that the filter having stronger smoothing intensity characteristics becomes smoother with respect to the pixels around the block distortion position.

It is reported that edge-preserving smoothing filters used as the filters 24 _{1 to} 24 _n include an ε filter, a ripple correction filter, etc. (Takahiko Fukiuki, “Suppression of DCT quantization distortion by a ripple correction filter”, 2002, video) It is known by the Information Media Society Winter Conference).

Next, the operation of the present embodiment will be described. The block unit encoded data (for example, MPEG bit stream) input via the input terminal 10 is supplied to the decoder 11 to be decoded and output as image data. The decoded image data output from the decoder 11 is supplied to the common terminal of the filter switch 23, while being supplied to the block distortion detection unit 22 in pixel units and the block distortion amount detection units 21 _{1 to} 21 _m in each region. The

  The block distortion detection unit 22 for each pixel performs block distortion presence / absence detection and block distortion intensity detection on the input decoded image data using a difference value between the block boundary and neighboring pixels, and outputs the result. To do. The operation of the block distortion detection unit 22 in units of pixels will be described in more detail with reference to FIGS. FIG. 3 is a diagram showing the vicinity of a unit block boundary in block coding, and FIG. 4 is a flowchart for explaining the operation of the block distortion detection unit 22 in units of pixels.

  In block distortion detection, when block distortion detection is performed for the block boundary 31 shown in FIG. 3, detection is performed using a1, a2, a3, i1, i2, and i3 that are adjacent pixels of the block boundary 31. The pixel difference absolute value between pixels (a1-i1) sandwiching the block boundary 31 is bound0, and the pixel difference absolute value between adjacent pixels of the neighboring pixels of the block boundary 31 is diff1, diff2, diff3, diff4, etc. as shown in FIG. Represented by When these difference absolute values are large, it indicates that the pixel value changes drastically between the pixels.

  Therefore, as shown in the flowchart of FIG. 4, the block distortion detection unit 22 in units of pixels starts with a boundary pixel difference that is an absolute value of a pixel difference between pixels (a1−i1) sandwiching the block boundary 31 illustrated in FIG. 3. The absolute value bound0 is calculated (step S1), and then the pixel difference absolute values diff1 to diff4 of the neighboring pixels are calculated (step S2). In step S2, as shown in FIG. 3, the absolute difference value diff1 between the pixel a1 and the pixel a2 in the block on the left side of the block boundary 31 and the absolute difference value diff2 between the pixel a2 and the pixel a3 are calculated. At the same time, the absolute difference value diff3 between the pixel i1 and the pixel i2 in the block on the right side of the block boundary 31 and the absolute difference value diff4 between the pixel i2 and the pixel i3 are calculated.

  Next, using the pixel difference absolute values (diff1 to diff4) of the neighboring pixels calculated in step S2, the pixel difference absolute value average (AveDiff_a, AveDiff_i) near the block boundary is calculated by the following equation (step S3).

AveDiff_a = (diff1 + diff2 + 1) / 2 (1)
AveDiff_i = (diff3 + diff4 + 1) / 2 (2)
In the detection of block distortion, there is a possibility of erroneously detecting an edge that appears at the block boundary by chance, and this can minimize the influence of the erroneous detection.

  Finally, block distortion detection and distortion level detection results are obtained by comparing bound0 with AveDiff_a and AveDiff_i (step S4). That is, when the boundary pixel difference absolute value bound0 is larger than the pixel difference absolute value averages AveDiff_a and AveDiff_i in the vicinity of the block boundary, it indicates that the block boundary 31 has undergone a large change compared to the vicinity thereof. Since the block distortion is a noticeable distortion because the block boundary 31 is greatly changed compared to the adjacent portion, the presence or absence of block distortion is detected by detecting an isolated pixel value change between the block boundary pixels. be able to.

  Therefore, in step S4, first, it is determined whether or not the pixel difference absolute value bound0 is larger than at least one of AveDiff_a × 2 and AveDiff_i × 2, which is twice the average pixel difference absolute value near the block boundary (step S41). ). If it is larger, it is subsequently determined whether it is larger than at least one of AveDiff_a × 3 and AveDiff_i × 3, which is three times the average pixel difference absolute value in the vicinity of the block boundary (step S43). It is determined whether it is larger than at least one of AveDiff_a × 4 and AveDiff_i × 4, which is four times the average pixel difference absolute value in the vicinity (step S45).

  If it is determined in step S41 that the pixel difference absolute value bound0 is not larger than the first threshold value (AveDiff_a × 2, AveDiff_i × 2) having the smallest value, the change in the block boundary is small compared to the adjacent location. Therefore, it is considered that there is no block distortion, and the detection level DistLevel is set to “1” (step S42).

  If it is determined in step S43 that the pixel difference absolute value bound0 is not greater than the second threshold value (AveDiff_a × 3, AveDiff_i × 3) having the second smallest value, the block boundary is compared with the adjacent location. It is assumed that the change is slightly small, and the detection level DistLevel is set to “2” (step S44).

  Similarly, when it is determined in step S45 that the pixel difference absolute value bound0 is not larger than the third threshold value (AveDiff_a × 4, AveDiff_i × 4) having the largest value, the block boundary changes compared to the adjacent portion. The detection level DistLevel is set to “3” (step S46), and the pixel difference absolute value bound0 is larger than the third threshold value (AveDiff_a × 4, AveDiff_i × 4) having the largest value in step S45. When the determination is made, it is considered that the block boundary is largely changed compared to the adjacent portion, and the detection level DistLevel is set to “4” (step S47).

  As described above, the block distortion detection unit 22 in units of pixels performs level division in four stages by comparing the threshold values obtained from bound0 and AveDiff_a and AveDiff_i in step S4, and the difference between bound0 and AveDiff_a and AveDiff_i is the largest. When DistLevel = 1 is small, the block boundary is considered to have a small change compared to the adjacent portion, and it is considered that there is no block distortion. As the difference between bound0, AveDiff_a, and AveDiff_i increases, the value of DistLevel also increases. It is estimated that large block distortion has occurred. As described above, the block distortion detection unit 22 in units of pixels determines the block distortion level (intensity) based on the magnitude of the boundary pixel difference absolute value bound0 and the threshold value related to the average AveDiff_a and AveDiff_i of the adjacent pixel difference absolute values. ).

Referring back to FIG. 1 again, the block distortion amount detection units 21 _{1 to} 21 _m in each region are all blocks in the assigned region among the m divided regions of the input decoded image data. Of the boundaries, the number of block boundaries where the block distortion with a certain or greater distortion intensity is measured independently of each other, and the result of the block distortion detection is measured independently of each other. As a result, the filter switches 23 are supplied.

The filter switch 23 includes a detection level DistLevel_a1i1 detected in pixel units from the block distortion detection unit 22 in pixel units, and block distortion detected for each region from the block distortion amount detection units 21 _{1 to} 21 _m in each region. As a selection control signal, a rough filter selection range for selecting which filter among the filters 24 _{1 to} 24 _n is determined by the block distortion amount for each region. For example, as this filter selection range, when n = 9, a set of filters 24 _{1 to} 24 _3, a set of filters 24 _{4 to} 24 _{6, and} a set of filters 24 _{7 to} 24 ₉ are respectively set as rough filter selection ranges, One filter group (filter selection range) is selected (determined) from this according to the block distortion amount for each region.

At this time, when the block distortion amount for each region is large, a filter set having a strong smoothing strength characteristic (a set of filters 24 _{1 to} 24 ₃ ) is selected, and when the block distortion amount for each region is small, smoothing is performed. If a group of filters having a weak smoothing characteristic is selected (a group of filters 24 _{7 to} 24 ₉ ) and the amount of block distortion for each region is medium, a group of filters having a moderate smoothing characteristic (filter 24 _{4 to} 24 ₆ ).

Subsequently, the filter switch 23 configures a plurality of filters constituting one filter group selected (determined) previously according to the value of the detection level DistLevel_a1i1 detected in pixel units from the block distortion detection unit 22 in pixel units. Among the filters, one filter is selected (determined). Therefore, when the block distortion amount for each region is small, the filter switch 23 uses a set of filters having a weak smoothing intensity characteristic to be used in that region, regardless of how large the block distortion amount for each pixel is detected. One filter corresponding to the block distortion amount in units of pixels is selected from the group of 24 _{7 to} 24 ₉ ).

In this way, the filter switch 23 supplies the image data decoded by the decoder 11 to one finally selected filter among the filters 24 _{1 to} 24 _n . Of the filters 24 _{1 to} 24 _n , one filter to which the decoded image data is supplied from the filter switch 23 performs edge preserving smoothing processing, and outputs image data with reduced block distortion as an output terminal. To 25.

Here, an edge-preserving smoothing process when the ε filter, which is a nonlinear filter, for example, is used as the filters 24 _{1 to} 24 _n will be described with reference to the flowchart of FIG. First, the pixel to be processed is A, the adjacent pixel of the pixel A is a (x), and these values are captured (step S11), and the difference value diff (= a (x) −A) is calculated (step S11). S12). Subsequently, it is compared whether or not the difference value diff is larger than a predetermined first threshold value THR (step S13). When diff> THR, the difference value diff is set to a first threshold value THR that is an upper limit value. (Step S14).

  On the other hand, when the comparison result of diff ≦ THR is obtained in step S3, it is compared whether or not the difference value diff is smaller than the second threshold value −THR (step S15), and if diff <−THR, the difference is calculated. The value diff is set to the second threshold value -THR which is the lower limit value (step S16). Further, when a comparison result of diff ≧ −THR is obtained in step S15, that is, when the difference value diff is between THR and −THR, or when the upper limit value is set in step S14, or in step S16. When the lower limit value is set, the difference value diff at that time is substituted into the variable sum_diff (step S17), and the processing of steps S11 to S17 is repeated for all the target adjacent pixels (step S18).

  In this way, a value after_filter is calculated by adding the value A of the target pixel to the average value obtained by dividing the variable sum_diff obtained for all of the target adjacent pixels by the number of adjacent pixels n (step S9). Is output as a filter output.

  Thereby, for example, among the pixels a1 to a5 arranged in the horizontal direction shown in FIG. 6, a 5-tap ε filter whose processing target is the central pixel a3 is obtained by the following equation.

上記のεフィルタは図７に示すように、入力Ｉｎに対し、出力Ｏｕｔが非線形となり、数１に示した画素差分絶対値ＡＢＳ（ａｉ−ａ３）が閾値ＴＨＲより小さい場合に、平滑化の対象となる。すなわち、閾値ＴＨＲによりフィルタ強度を制御できる。よって、各領域でのブロック歪み量検出部２１_１〜２１_ｍからの領域毎に検出されたブロック歪み量に応じて、ブロック歪み量が大きい場合にはエッジ保存型平滑化フィルタ強度を強く、適用範囲を広くするようにフィルタの切り替え制御を行うことで、切り替え時の不自然さの少ないブロック歪みの軽減を行うことができる。

As shown in FIG. 7, the ε filter described above is subject to smoothing when the output Out becomes nonlinear with respect to the input In, and the pixel difference absolute value ABS (ai−a3) shown in Equation 1 is smaller than the threshold value THR. It becomes. That is, the filter strength can be controlled by the threshold value THR. Therefore, according to the block distortion amount detected for each region from the block distortion amount detection units 21 _{1 to} 21 _m in each region, the edge preserving smoothing filter strength is increased and applied when the block distortion amount is large. By performing filter switching control so as to widen the range, it is possible to reduce block distortion with little unnaturalness during switching.

Thus, according to the present embodiment, not only on / off of one filter as in the conventional block distortion reducing device described in Patent Document 1, but also by switching the filters 24 _{1 to} 24 _n having several levels of strength. Smooth filtering with reduced visual effects is possible. And since it can judge not with the block distortion amount of the whole frame like patent document 1 but with the distortion amount for every area | region which divided | segmented the screen into multiple pieces, as shown in FIG. It is possible to select a strong filter for the local block distortion occurring in the area, and only the local block distortion without affecting the area where the block distortion does not occur. Can be reduced.

  Further, as in the conventional block distortion reduction device described in Patent Document 2, the filter is switched only by information from a block distortion detection unit that detects the presence and intensity of block distortion using a pixel change value near the block boundary. Instead, the information on the amount of distortion is added to the determination material and the filter is switched, so that the optimum selection application from a plurality of filters can be performed.

(Second Embodiment)
Next, a second embodiment of the present invention will be described. FIG. 8 shows a block diagram of a second embodiment of a block distortion reducing apparatus according to the present invention. In the figure, the same components as those in FIG. In the second embodiment shown in FIG. 8, value holding units 311 to 31k, 321 to 32k, 331 to 33k, and average value calculation units 341 to 34m are added to the configuration of the first embodiment. The filter switch 35 is characterized in that switching control is performed in accordance with the block distortion amount average value for each region from the average value calculation units 341 to 34m and the block distortion amount from the block distortion detection unit 22 in units of pixels. .

8, the divided regions each region block distortion amount detection unit 21 1 of the _unit, 21 _2, ..., respectively output distortion amount detection signal from the 21 _m (detection level DistLevel) via an input terminal 30 , 33 ₁ are held in the value holding units 31 ₁ , 32 ₁ ,..., 33 ₁ at the timing of the frame pulse input at every frame switching, and the value holding units 31 ₁ , 32 ₁ , 33 values held to _1, the holding portion _{31 2,} 32 2 of the next stage values, ... are held shifted by 33 _2. Similarly, the value held in the previous value holding unit is shifted to the next stage every time the frame pulse is input.

Accordingly, k value holding units 31 _{1 to} 31 k, 32 _{1 to} 32 k,..., 33 ₁ to 33 k are sequentially assigned to k frames (k is an arbitrary natural number) at the frame pulse timing. The block distortion amount is held, and the block distortion amount for each area in units of frames from the holding units 31 _{1 to} 31 k, 32 _{1 to} 32 k,..., 33 ₁ to 33 k is calculated as an average value calculation unit 34 ₁ . 34 ₂ ,..., 34 _m are supplied in parallel to calculate an average value, and the averaged distortion amount for each region is supplied to the filter switch 35.

The filter switch 35 is based on the average value of the block distortion amount for each area detected by the average value calculation units 34 ₁ , 34 ₂ ,..., 34 _m for each frame from the first frame to the kth frame. A filter group is selected from among the filters 24 _{1 to} 24 _n , and one filter is selected from among a plurality of filters in the group in accordance with the detected amount of block distortion in units of pixels. The image data of the (k + 1) th frame from the decoder 11 is supplied. Thus, according to the present embodiment, frequent filter switching can be avoided and visual unnaturalness can be suppressed even when the block distortion amount changes greatly for each frame. Further, similarly to the first embodiment, it is possible to perform a reduction operation only on locally generated block distortion.

The present invention is not limited to the above embodiments, for example, the average value calculating unit 34 ₁ to 34C _m may be replaced with an intermediate value selecting section for selecting an intermediate value. Further, in the above embodiment, the description has been made so that the image data encoded by the block unit is applied to the decoded image data after decoding. However, the present invention decodes the image data encoded by the block unit. It can also be applied at the stage. In the above embodiment, the description has been made on the pixel column in the horizontal direction. However, if similar processing is performed on the pixel column in the vertical direction, it can be applied to reduce adjacent block distortion in the vertical direction. . Furthermore, the present invention can also be realized by a computer program. In this case, the program can be taken into a computer via a recording medium or a communication line.

It is a block diagram of a 1st embodiment of the present invention. It is a figure which shows an example of the m area | region which divided the screen of 1 frame into m in this invention. It is a figure shown about the unit block boundary periphery in block coding. 2 is a flowchart for explaining the operation of a block distortion detection unit in units of pixels in FIG. 1. It is a flowchart for demonstrating an example of the process by an epsilon filter. It is a figure which shows the pixel located in a line with a horizontal direction. It is a figure which shows the input-output characteristic of an epsilon filter. It is a block diagram of the 2nd Embodiment of this invention. It is a block diagram of an example of the conventional block distortion reduction apparatus of patent document 1. It is a block diagram of an example of the conventional block distortion reduction apparatus of patent document 2. FIG. 10 is a diagram in which local block noise is generated in a screen, which is a problem in Patent Document 1.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Input terminal of image data encoded by block unit encoding 11 Decoder 21 _{1 to} 21 Block distortion amount detection unit in _m region 22 Block distortion detection unit in units of pixels 23 and 35 Filter switch 24 _{1 to} 24 _n filter 25 Output terminal 30 Frame pulse input terminal 31 _{1 to} 31 k, 32 _{1 to} 32 k,..., 33 _{1 to} 33 k Value holding unit 34 _{1 to} 34 _m Average value calculating unit

Claims (1)

The block-unit encoded image data obtained by compressing and encoding the image data in units of blocks having a predetermined number of pixels is decoded, the decoded image data is received as input image data, and adjacent to the input image data In a block distortion reducing apparatus for reducing block distortion generated at a boundary between the blocks,
A plurality of filter sets obtained by grouping a plurality of filters having different smoothing intensity characteristics according to the smoothing intensity;
A pixel difference absolute value between two pixels sandwiching a block boundary of the input image data and an average value of pixel difference absolute values between adjacent pixels in the same block at two or more pixels in the vicinity of the block boundary are calculated. , Block distortion detection means in pixel units for detecting the level of block distortion from the difference between the pixel difference absolute value and the average value of the pixel difference absolute values between the adjacent pixels;
When one screen of the input image data is divided into a plurality of areas larger than the block, among all the block boundaries existing in the divided area, a block having a block distortion having a distortion intensity equal to or greater than a predetermined value A block distortion amount detecting unit for each area for detecting the block distortion amount based on the number of boundaries;
After selecting one filter set having a smoothing intensity characteristic according to the block distortion amount for each of the divided regions detected by the block distortion amount detection unit for each region from the plurality of filter sets, One filter is selected from the set of the one filters selected according to the block distortion detection level detected by the block distortion detection means in pixel units, and the input image data is supplied to the one filter. And a block distortion reducing device.

Images