JPH11275584A - Block distortion reduction circuit for image signal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a block distortion reduction circuit for an image signal that realizes suppression of block distortion in matching with a visual characteristic by a simple configuration. SOLUTION: A decoding section 1 outputs an image signal series VS, motion vector information MV, and de-blocking filter operation area information BAR based on an input signal BTS. A motion vector conversion section 2 generates a motion vector BV per one frame of an image signal series VS based on the motion vector information MV. A characteristic setting section 3 generates a selection control signal SL based on a motion speed detected as a scalar quantity of the motion vector BV and the de-blocking filter operation area information EAR. A de-blocking filter section 4 obtains a signal VP resulting from application of optimum de-blocking processing from a still area up to a moving image area according to the selection control signal SL.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a circuit for suppressing block distortion generated in discrete cosine transform (hereinafter abbreviated as DCT) coding, and more particularly to a circuit for changing the characteristics of a deblocking filter according to the speed of image motion. The present invention relates to an image signal block distortion reduction circuit suitable for effectively suppressing block distortion in signal processing to be performed.

[0002]

2. Description of the Related Art In DCT coding, when the bit rate is low, the quantization accuracy of DCT coefficients becomes coarse, and block distortion occurs in a decoded image. In order to suppress this, it is known to pass through a low-pass filter called a deblocking filter. However, there is a problem that sharpness is reduced.

In general, block distortion in DCT coding tends to occur frequently in a motion part. For this reason, a method of detecting the amount of motion of an image and passing it through a narrow low-pass filter in a region where movement is strong and a wide low-pass filter in a region where movement is gentle is disclosed in the 1997 Annual Conference of the Institute of Image Information and Television Engineers, 13- 5, pages 156 to 157. Here, in order to simplify the signal processing, the gradient method is used for detecting the amount of motion.

As another method for accurately detecting the amount of motion, a method of detecting a motion vector by a block matching method is known, as described in, for example, Japanese Patent Application Laid-Open No. 7-170496.

[0005]

SUMMARY OF THE INVENTION In the prior art,
In the method using the gradient method, the gradient direction component of the movement amount is obtained from the gradient at a certain position of the image and the frame difference.
There is a problem that a component orthogonal to the gradient direction cannot be detected. In addition, the method of detecting a motion vector by the block matching method requires a huge amount of calculation for searching for a motion vector, and has a problem that the apparatus becomes complicated and costs increase.

SUMMARY OF THE INVENTION The present invention has been made in consideration of the above problems, and has as its object to provide a block distortion reduction circuit for an image signal, which can suppress block distortion matched to visual characteristics with a simple configuration.

[0007]

An object of the present invention is to provide a decoding unit for decoding an image signal sequence and motion vector information from an input signal, and a motion for generating a motion vector per frame of the image signal sequence from the motion vector information. This is achieved by a circuit including a vector conversion unit and a filter unit that filters an image signal sequence using a deblocking filter having a frequency characteristic corresponding to a scalar amount of a motion vector per frame.

In addition, this circuit is provided with a prediction error calculating section for calculating a motion correction error occurring in a motion vector per frame of the image signal sequence, and performs filtering processing of the image signal sequence in consideration of the amount of the motion correction error. By doing so, it is possible to suppress block distortion limited to a conspicuous area.

Here, the frequency characteristics of the deblocking filter can be constituted by one-dimensional frequency characteristics in the horizontal direction or two-dimensional frequency characteristics in the horizontal and vertical directions. The operation of the deblocking filter is performed in a region around the boundary of each block in the decoded image signal sequence. The input signal to the decoding unit is a signal compliant with the MPEG video coding of the international standard.

Further, the image decoding apparatus of the present invention converts an image signal sequence and motion vector information from a video coded bit stream signal obtained by coding an image signal by a combination of motion compensation predictive coding and discrete cosine transform coding. The image signal sequence is configured to be decoded and filtered according to the scalar amount of the motion vector per frame of the image signal sequence generated from the motion vector information. Here, when the motion correction error generated in the motion vector per frame of the image signal sequence is less than the threshold, the filtering process is performed with the through characteristic.

The image decoding method according to the present invention further comprises the steps of: decoding an image signal sequence and motion vector information from an input signal; generating a motion vector per frame of the image signal sequence from the motion vector information; The method includes a step of selecting a desired filter from a plurality of types of filters having different frequency characteristics in accordance with a scalar amount of a winning motion vector, and a step of performing a filtering process on an image signal sequence using the selected filter. Here, as the filter, when the scalar amount of the motion vector is large, a filter having a narrow low-pass characteristic is selected, and as the scalar amount becomes small, a filter having a gradually loosened low-pass characteristic is selected.

As a result, highly accurate detection of the amount of motion can be performed with extremely simple signal processing, and suppression of block distortion more consistent with visual characteristics can be realized at low cost.

[0013]

FIG. 1 is a block diagram showing a first embodiment of the present invention. In the figure, a decoding unit 1 performs a predetermined decoding process on a video coded bit stream signal BTS obtained by coding an image signal by a combination of motion compensation prediction coding and discrete cosine transform coding, and The signal sequence VS, the motion vector information MV, and the deblocking filter operation area information BAR (0 is canceled,
1 is an operation).

The motion vector conversion unit 2 generates a motion vector BV per frame of an image signal sequence by performing a process of calculating a motion vector for each block of the motion vector information MV. The characteristic setting unit 3 generates the selection control signal SL based on the motion speed detected by the scalar amount of the motion vector BV and the deblocking filter operation area information BAR. The deblocking filter unit 4 performs a filtering process with characteristics determined by the selection control signal SL, and obtains a signal VP that has been subjected to an optimal deblocking process from a still region to a moving image region.

The configuration of each unit in an image decoding apparatus having such an image signal block distortion reduction circuit will be described below.
The operation will be described in detail. FIG. 2 shows a motion vector conversion unit 2
1 is a configuration example. In the P picture vector conversion unit 5,
1 from motion vector MVp used for decoding P-picture
Generate a motion vector in a frame period. FIG. 3 shows an outline of this operation. The P picture indicated by the symbol P in the image signal sequence obtained from the decoding unit is subjected to one-way motion compensation prediction coding and DCT coding using a motion vector MVp between n frames. Therefore, this motion vector M
Vp (MVp1, MVp2, MVp3 in the figure) is divided by the number of frames n between P pictures (n = 3 in the figure) to be converted into a motion vector for one frame, and a motion vector BVp
(MVp1 / 3, MVp2 / 3, MVp3 / 3 in the figure)
Generate

The B picture vector conversion unit 6 generates a motion vector in one frame period from the motion vector MVb used for decoding a B picture. This operation is schematically shown in FIG. The B picture indicated by the symbol B in the image signal sequence obtained from the decoding unit is subjected to bidirectional motion compensation prediction coding and DCT coding using the motion vector MVb.
Therefore, among the motion vectors MVb, the motion vectors for one frame (MVb11, MVb12, MVb12 in the figure)
b21, MVb22, MVb31, MVb32) to generate a motion vector BVb.

The control section 7 generates a control signal CT required for the operation of each section based on the coding parameters (information of I, P, B pictures, etc.) of the motion vector information MV. The selecting unit 8 calculates a motion correction error value for each block with respect to the motion vectors BVp and BVb, and outputs a vector having a small value to the block unit motion vector BV.

FIGS. 5A to 5C are diagrams for explaining the operation of the characteristic setting section 3. FIG. FIG. 6A shows the relationship between the motion vector BV and the frequency characteristics of the deblocking filter. As shown in the figure, the scalar amount of the motion vector | B
In a region of a large movement where V | is large, a narrow low-pass characteristic 3 is selected. In the region of | BV | where the movement is normal, and in the region where | BV | is small and slowly moving, the low-pass characteristics 2 and 1 that are gradually loosened are selected.
In the still region where | BV | = 0, the through characteristic is selected. The state of the scalar quantity | BV | and the selection control signal S
The relationship of L is shown in FIG.

FIG. 2C shows an operation area of the deblocking filter. In the drawing, a rectangle surrounded by a solid line corresponds to an area of one block. An area around the block boundary indicated by oblique lines indicates an operation area of the deblocking filter. This operation area is set by deblocking filter operation area information BAR (0 indicates stop, 1 indicates operation).

In the operation area where the deblocking filter operation area information BAR is 1, the selection control signal S shown in FIG.
L is output. On the other hand, in the operation suspension area where BAR is 0, the through characteristic is output to the selection control signal SL.

FIG. 6 is a diagram showing an example of the configuration of the deblocking filter unit 4. Vertical deblocking filter 9
, 9-2,..., 9-n perform filtering processing of the characteristic 1, characteristic 2,. Horizontal deblocking filters 11-1, 11-2,
, 11-n respectively perform filtering processing of the characteristic 1, characteristic 2, ..., characteristic n of the horizontal low-pass. Selector 1
0, 12 select and output signals of through, characteristic 1,..., Characteristic n according to the selection control signal SL, respectively.

FIG. 7 is a diagram showing an example of an impulse response of the deblocking filter. The slowest low-pass characteristic 1 is realized with an impulse response of 1, 2, 1 (× １ ／). Characteristic 2,..., Characteristic 4
A filter having an impulse response of 1,1 (× １ ／) is realized in a cascade connection.

As described above, according to the present embodiment, a block distortion reduction circuit for an image signal for effectively reducing block distortion can be realized at a low cost with a simple configuration, and DCT coding is performed. A remarkable effect can be obtained for improving the image quality of an image.

Next, a second embodiment of the present invention will be described with reference to the block diagram of FIG. This embodiment is suitable for suppressing block distortion only in a conspicuous area. The decoding unit 1, the motion vector conversion unit 2, and the deblocking filter unit 4 shown in FIG.
This is the same as that of the embodiment.

The prediction error calculator 13 calculates the motion vector BV
ER is calculated. Motion vector BV
, The horizontal component is BVx, the vertical component is BVy, and the signals of the current frame and the previous frame of the image signal sequence are VSct and VSp.
r, if the position of the pixel in the block is (x, y), the motion compensation error is ER = Σ | VSct (x, y) −VSpr
(X + BVx, y + BVy) |. Here, Σ indicates addition of all pixels in the block.

The characteristic setting section 14 sets the selection control signal SL in a relationship as shown in FIGS. 9 (a) and 9 (b). That is, as shown in FIG. 7A, a block in which the motion correction error ER is equal to or larger than the threshold value TH is determined to be a region where the quantization characteristics of DCT coding are coarse and block distortion is conspicuous. Then, as in the first embodiment, the scalar amount of the motion vector |
The selection control signal SL is set so as to select a filter having such a characteristic that the low-pass characteristic becomes narrower as BV | becomes larger. On the other hand, as shown in FIG.
A block in which R is less than the threshold value TH is determined to be a region where the quantization characteristics of DCT coding are fine and block distortion does not occur. Then, the selection control signal SL is set to a through characteristic.

As described above, according to this embodiment, a block distortion reduction circuit for an image signal that suppresses block distortion limited to a conspicuous region can be realized with a simple configuration at low cost. A remarkable effect can be obtained for improving the image quality of a DCT-coded image.

Next, FIG. 10 shows a block configuration of a third embodiment of the present invention. In the present embodiment, the operation area of the deblocking filter is performed for all pixels in the block. The decoding unit 1, the motion vector conversion unit 2,
The deblocking filter unit 4 is the same as that of the first embodiment shown in FIG. Characteristic setting unit 15 shown in FIG.
Sets the selection control signal SL and performs the operation when the signal of the deblocking filter operation area information BAR in the first embodiment of FIG. 1 is 1.

Next, FIG. 11 shows a block configuration of a fourth embodiment of the present invention. In this embodiment as well, the operation area of the deblocking filter is performed for all the pixels in the block. The decoding unit 1, the motion vector conversion unit 2,
The deblocking filter unit 4 is the same as that of the first embodiment shown in FIG. 1, and the prediction error calculation unit 13 is the same as that of the second embodiment shown in FIG. The characteristic setting unit 16 shown in FIG.
In the second embodiment, the operation is performed when the signal of the deblocking filter operation area information BAR is 1.

As described above, according to the third and fourth embodiments, a circuit for reducing block distortion of an image signal can be realized by simpler signal processing.

As described above, according to the present invention, an image signal block distortion reduction circuit and an image decoding device for suppressing block distortion generated by DCT coding in a form matched with the visual characteristics can be realized with a simple configuration and at low cost. can do.
In addition, a remarkable effect can be obtained in improving the quality of a received image such as a digital broadcast in which high-efficiency coding based on DCT coding is performed.

[0032]

According to the present invention, it is possible to provide a circuit for reducing block distortion of an image signal, which can suppress block distortion matched to visual characteristics with a simple configuration.

[Brief description of the drawings]

FIG. 1 is a diagram showing a block configuration of a first embodiment of the present invention.

FIG. 2 is a diagram illustrating a configuration example of a motion vector conversion unit.

FIG. 3 is a diagram illustrating an operation of a P picture vector conversion unit.

FIG. 4 is a diagram for explaining an operation of a B picture vector conversion unit.

FIGS. 5A to 5C are diagrams for explaining the operation of a characteristic setting unit;

FIG. 6 is a diagram illustrating a configuration example of a deblocking filter unit.

FIGS. 7A to 7D are diagrams each showing an example of an impulse response of a deblocking filter.

FIG. 8 is a diagram showing a block configuration of a second embodiment of the present invention.

FIG. 9 is a diagram for explaining an operation of a characteristic setting unit.

FIG. 10 is a diagram showing a block configuration of a third embodiment of the present invention.

FIG. 11 is a diagram showing a block configuration of a fourth embodiment of the present invention.

[Explanation of symbols]

 Reference Signs List 1 decoding unit 2 motion vector conversion unit 3, 14, 15, 16 characteristic setting unit 4 deblocking filter unit 5 P picture vector conversion unit 6 B picture vector conversion unit 7 control unit 8, 10, 12 selection unit 9 vertical deblocking Filter 11 Horizontal deblocking filter 13 Prediction error calculation unit

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Masato Sugiyama 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Inside the Multimedia Systems Development Division of Hitachi, Ltd. (72) Inventor Mitsuo Nakajima Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa 292 Hitachi Multimedia Systems Development Division

Claims (10)

[Claims] A decoding unit configured to decode an image signal sequence and motion vector information from an input signal; a motion vector conversion unit configured to generate a motion vector per frame of the image signal sequence from the motion vector information; A filtering unit for filtering the image signal sequence using a deblocking filter having a frequency characteristic according to a scalar amount of a motion vector per block. 2. A decoding unit for decoding an image signal sequence and motion vector information from an input signal, a motion vector conversion unit for generating a motion vector per frame of the image signal sequence from the motion vector information, A prediction error calculation unit for calculating a motion correction error occurring in a motion vector per frame of a signal sequence; and a deblocking filter having a scalar amount of the motion vector per frame and a frequency characteristic corresponding to the calculated motion correction error amount. And a filter unit for filtering the image signal sequence by using the image signal sequence. 3. The circuit according to claim 1, wherein the frequency characteristics of the deblocking filter are one-dimensional frequency characteristics in a horizontal direction. 4. The circuit according to claim 1, wherein the frequency characteristics of the deblocking filter are two-dimensional frequency characteristics in horizontal and vertical directions. 5. The image signal block according to claim 1, wherein the operation area of the deblocking filter is an area around a boundary of each block in the decoded image signal sequence. Distortion reduction circuit. 6. The method according to claim 1, wherein the input signal is an MPE of an international standard.
The circuit according to any one of claims 1 to 5, wherein the signal is a signal compliant with G video encoding. 7. An image signal sequence and motion vector information are decoded from a video coded bit stream signal obtained by encoding an image signal by a combination of motion compensation predictive coding and discrete cosine transform coding. An image decoding device, characterized in that the image signal sequence is filtered in accordance with a scalar amount of a motion vector per frame of the generated image signal sequence. 8. The apparatus according to claim 7, wherein when the motion correction error generated in the motion vector per frame of the image signal sequence is less than a threshold value, the filtering process is performed with a through characteristic. Image decoding device. 9. Decoding an image signal sequence and motion vector information from an input signal; generating a motion vector per frame of the image signal sequence from the motion vector information; Selecting a desired filter from a plurality of types of filters having different frequency characteristics according to the scalar amount; and filtering the image signal sequence using the selected filter. Image decoding method. 10. The step of selecting the filter includes selecting a filter having a narrow low-pass characteristic when the scalar amount of the motion vector is large, and gradually reducing the filter with a low-pass characteristic as the scalar amount decreases. 10. The image decoding method according to claim 9, wherein the method is selected.