JP3725232B2 - Motion vector determination method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a motion compensation method for use in a video encoding system, and more particularly to an improved motion vector determination method capable of removing a blocking phenomenon appearing in a decoded video signal.
[0002]
[Prior art]
As is well known, when transmitting a digitized video signal, higher accuracy can be maintained than when transmitting it in analog. When a video signal composed of a series of video “frames” is expressed in digital form, a considerable amount of data is generated, particularly in the case of a high definition television (HDTV) system. However, since the effective frequency bandwidth of a normal transmission channel is limited, it is necessary to compress or reduce the amount of data in order to transmit a considerable amount of digital data. Of the various compression techniques, a hybrid coding technique combining a statistical coding technique with temporal and spatial compression techniques is known to be most effective.
[0003]
Most hybrid coding techniques use motion compensated DPCM (Differential Pulse Code Modulation), two-dimensional DCT (Discrete Cosine Transform), DCT coefficient quantization and VLC (Modulatable Coding). Motion compensated DPCM estimates the motion of the object between the current frame and the previous frame, predicts the current frame according to the motion of the object, and generates a difference signal that represents the difference between the current frame and the predicted value .
[0004]
A two-dimensional DCT that reduces or eliminates spatial redundancy between video data, such as motion compensated DPCM data, converts a block of digital video data, eg, an 8 × 8 pixel block, into a set of transform coefficient data.
[0005]
More specifically, in motion compensated DPCM, the current frame is predicted from the previous frame based on the motion of the object estimated between the current frame and the previous frame. The estimated motion can be represented by a two-dimensional motion vector representing the displacement of the pixel between the previous frame and the current frame.
[0006]
Several methods for estimating the displacement of an object have been proposed. Generally, these methods can be classified into two methods. One is a pixel sequential method, and the other is a block matching method. The present invention relates to the block matching method.
[0007]
In the block matching method, the current frame is divided into a plurality of search blocks. Search block sizes vary from 8x8 to 32x32. In order to determine the motion vector of the search block in the current frame, the similarity between the search block in the current frame and each of a plurality of candidate blocks of the same size included in a larger search area in the previous frame Perform the calculation. An error function such as mean absolute error (MAE) or mean square error (MSE) is used to perform similarity calculations between the search block of the current frame and each candidate block in the search area. The motion vector represents the displacement between the search block that generates the minimum error function and the candidate block.
[0008]
The encoded video data is transmitted through a normal transmission channel to a video signal decoder included in the video signal decoding system, and this video signal decoder performs a process reverse to the encoding process, Reconstruct the original video data. However, in this reconstructed video data, a blocking phenomenon occurs in which a block boundary line becomes noticeable at the reception stage. Such a blocking phenomenon occurs because a frame is encoded in units of blocks.
[0009]
[Problems to be solved by the invention]
Therefore, the main object of the present invention is to eliminate motion blocking at the video signal block boundary and improve the image quality of the video processed by the system for motion compensation for use in a video coding system. It is to provide a motion vector determination method.
[0010]
[Means for Solving the Problems]
In order to achieve the above object, the present invention determines an optimal motion vector between a current frame and a previous frame of a video signal, and the current frame is divided into a plurality of search blocks of the same size. A frame includes a corresponding number of search regions, and each search region has a plurality of candidate blocks of the same size. In the optimal motion vector determination method, the method uses a block matching method for each search block in the current frame. A first step of detecting a motion vector, a second step of dividing a search block into a central region located in the center of the search block and a boundary region located on the outer surface of the central region, and a motion vector of the search block in the central region Using the third process of determining the optimum motion vector and the motion vector of the search block and one or more search blocks adjacent thereto, the search is performed. In motion vector determination method which is characterized in that it comprises a fourth step of determining a plurality of optimal motion vectors of the block in the border region.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 shows a block diagram of a coding device 10 for digital video signal compression including a motion compensation device 150 according to the present invention. Here, it is assumed that the digital video signal includes a plurality of frames of the video signal.
[0012]
The encoding device 10 includes a first frame memory 100, a subtracter 102, a video signal encoder 105, an entropy encoder 107, a video signal decoder 113, an adder 115, a second frame memory 124, and a motion compensation device 150. .
[0013]
The current frame included in the input video signal is stored in the first frame memory 100 connected to the subtractor 102 via the line L9 and the motion compensator 150 via the line L10. The stored current frame is read out in units of blocks. Here, the size of the blocks varies from 8 × 8 to 32 × 32 pixels.
[0014]
The motion compensator 150 according to the present invention first detects a motion vector for each search block of the current frame using a normal block matching method, and the motion vector is detected from each search block included in the current frame and the second frame memory 124. Represents the spatial displacement between the most similar candidate blocks included in the corresponding search region in the previous frame supplied from the. In addition, the motion compensation device 150 determines an optimal motion vector of a pixel included in each search block using the motion vector of the search block and the search block adjacent thereto. Thereafter, the motion compensator 150 reads the corresponding pixel value of the previous frame from the second frame memory 124 using the determined optimum motion vector, and supplies the predicted current frame. The motion vector of the search block is supplied to the entropy encoder 107 via the line L20, and the predicted current frame signal is supplied to the subtractor 102 and the adder 115 via the line L30. The detailed operation of the motion compensation device 150 will be described in detail later with reference to FIGS.
[0015]
The predicted current frame signal supplied from the motion compensator 150 is subtracted from the current frame signal supplied via the line L9 in the subtractor 102, and the result data, that is, the current value predicted as the pixel value of the current frame. An error signal representing the difference between the pixel values of the frame is supplied to the video signal encoder 105, where the error signal is encoded into quantized transform coefficients using, for example, DCT and known quantization methods. The
[0016]
Thereafter, the quantized transform coefficients are transmitted to the entropy encoder 107 and the video signal decoder 113. The video signal decoder 113 converts the quantized transform coefficient supplied from the video signal encoder 105 into a reconstructed error signal using inverse quantization and inverse discrete cosine transform. In the adder 115, the reconstructed error signal supplied from the video signal decoder 113 and the predicted current frame signal supplied from the motion compensator 150 are combined, and the reconstructed current frame signal obtained therefrom is the first. It is stored in the 2-frame memory 124 as a previous frame signal.
[0017]
In the entropy encoder 107, the quantized transform coefficient supplied from the video signal encoder 105 and the motion vector supplied from the motion compensator 150 via the line L20 are encoded together by variable length encoding. The encoded signal is transmitted through a transmitter (not shown).
[0018]
FIG. 2 shows the motion compensation device 150 shown in FIG. 1 in detail. The motion compensator 150 includes a motion estimator 209, a memory 210, a region forming block 213, an optimal motion vector determination block 216, and a motion compensator 219.
[0019]
First, the motion estimator 209 searches the current frame signal from the first frame memory 100 and the previous frame signal from the second frame memory 124 and uses each block in the current frame and the previous frame signal using a normal block matching technique. A motion vector representing the spatial displacement between the most similar blocks in the frame is detected. The motion vector of each search region from the motion estimator 209 is supplied to the entropy encoder 107 and the memory 210 shown in FIG. 1, and the motion vector of the search block of the current frame is stored in the memory 210.
[0020]
At this time, the current frame signal is also supplied to the area forming block 213, where each search block is divided into a boundary area and a center area. FIG. 3 illustrates a region formation method performed in region formation block 213 in accordance with a preferred embodiment of the present invention. As shown in FIG. 3, each search block, for example, a search block SB5 of 16 × 16 pixels, is located along the outer surface of the central region CR5 and a central region CR5 of 12 × 12 pixels located at the center of the search block SB5 The boundary area is divided into pixel boundary areas, and the boundary areas include four edge areas ER5-1 to ER5-4 and four corner areas CR5-1 to CR5-4. Each edge area is composed of 12 × 2 pixels, The corner area is composed of 2 × 2 pixels. Region information representing the position of the center, edge and corner regions in the current frame is provided to the optimal motion vector determination block 316.
[0021]
The optimum motion vector determination block 216 receives the motion vector of the search block of the current frame from the memory 210, and according to the region information supplied from the region formation block 213, the motion vector of each search block and the search block adjacent thereto. Is used to determine the optimum motion vector for each search block region. More specifically, the optimal motion vector of an edge region (eg, ER5-1) located along the boundary between two search blocks (eg, SB5 and SB2) has two search blocks, ie, SB5 and SB2. It is obtained by taking the average of motion vectors. Similarly, the optimum motion vectors of the edge regions ER5-2 to ER5-4 are determined by taking the average of the motion vectors of the two search blocks SB5 and SB4, SB5 and SB6, and SB5 and SB8, respectively. On the other hand, the optimum motion vector of the corner area is obtained by taking the average of the motion vector of the search block including the corner area and the motion vector of the search block adjacent to the search block in the corner area. For example, the optimal motion vector of the corner area CR5-1 is obtained by averaging the motion vectors of the search blocks SB1, SB2, SB4, and SB5. Further, the optimum motion vectors of the corner areas CR5-2 to CR5-4 are determined based on the motion vectors of the search blocks SB2 to SB9. Similar to the optimal motion vector of the central region (eg, CR5), the motion vector of the search block (eg, SB5) including the central region CR5 is assigned as the optimal motion vector.
[0022]
In the above-described method of the present invention, adjacent corner regions (eg, CR1-1, CR2-1, CR4-1, and CR5-1) are searched blocks including these corner regions (eg, SB1, SB2, SB4, and Edge regions (e.g., ER2-1 and ER5-1) that have the same optimal motion vector determined by taking the average of the motion vectors of SB5), and that are adjacent to each other (for example, ER2-1 and ER5-1) For example, the optimal motion vector calculated by averaging the motion vectors of SB2 and SB5) is shared.
[0023]
The optimal motion vector of each search block in the current frame is supplied to the motion compensator 219.
The motion compensator 219 uses the optimal motion vector supplied from the optimal motion vector determination block 216 to read each corresponding pixel value of the previous frame from the second frame memory 124 shown in FIG. The signal is supplied to the subtracter 102 and the adder 115 shown in FIG.
[0024]
In the digital video signal decoding apparatus corresponding to the encoding apparatus 10 according to the present invention, the description has been made in relation to the region forming block 213 and the optimum motion vector determination block based on the motion vector transmitted from the encoding apparatus 10. By reconstructing the optimal motion vector in the same manner as described above, it is possible to reconstruct the current frame signal based on the transmission signal representing the error signal and the motion vector.
[0025]
While specific embodiments of the invention have been described above, it will be appreciated by those skilled in the art that various modifications may be made without departing from the scope of the claims set forth in the invention.
[0026]
【The invention's effect】
Therefore, according to the present invention, it is possible to remove the blocking phenomenon that occurs at the block boundary of the video signal and improve the image quality of the video processed by the system.
[Brief description of the drawings]
FIG. 1 is a block diagram of a video signal encoding system using a motion compensation apparatus of the present invention.
FIG. 2 is a detailed block diagram of the motion compensation device shown in FIG.
FIG. 3 is a diagram for explaining boundary region formation performed in the motion compensation apparatus of the present invention.
[Explanation of symbols]
100 first frame memory 102 differentiator 105 video signal encoder 107 entropy encoder 113 video signal decoder 115 adder 124 second frame memory 150 motion compensator 209 motion estimator 210 memory 213 region forming block 216 optimum motion vector determining block 219 Motion compensator

Claims (1)

An optimal motion vector is determined between the current frame and the previous frame of the video signal, the current frame is divided into a plurality of search blocks of the same size, and the previous frame includes a corresponding number of search areas, In the optimal motion vector determination method, the search region has a plurality of candidate blocks of the same size, the method:
A first step of detecting a motion vector of each search block in the current frame by a block matching method;
A second process of dividing the search block into a central area located at the center of the search block and a boundary area located on the outer surface of the central area;
A third step of determining the motion vector of the search block as the optimal motion vector of its central region;
Using the motion vector of the search block and one or more search blocks adjacent thereto, a fourth step of determining the optimum motion vector in the boundary area of the search block;
The A including motion vector determination method,
The boundary area of the search block includes four edge areas and four corner areas, each corner area is located at a corner of the search block, and each edge area is along an edge of the search block between the two corner areas. Each corner region has three neighboring search blocks, and each edge region has one neighboring search block,
The optimum motion vector of each corner area is determined by taking the average of four motion vectors of the search area and the three search blocks adjacent thereto, and the optimum motion vector of each edge area is adjacent to the search block. And determining a motion vector by averaging the two motion vectors of one search block .

Images