KR0174448B1 - Method and apparatus for encoding a video signal using region based motion vectors - Google Patents

Abstract

The present invention relates to a motion compensation video signal encoding apparatus for determining a predicted current frame based on a current frame and a previous frame of a discrete video signal, comprising: dividing means for dividing a previous frame into a plurality of regions; Area unit displacement estimation means for providing a plurality of displacement vectors representing the motion of each of the plurality of divided regions between a current frame and a previous frame; Prediction means for moving the plurality of regions by a corresponding displacement vector to form an intermediate prediction frame; Using the plurality of displacement vectors, a motion vector for each pixel included in an intermediate prediction frame is provided, and a pixel value of a previous frame corresponding to the motion vector for each pixel of the intermediate prediction frame is generated as the corresponding pixel value and predicted. And post-processing means for determining the current frame.

Description

Apparatus and method for encoding motion compensation image signal using region-based motion vector, apparatus and method for decoding motion compensation image signal

1 is an overall block diagram of a video encoder showing an embodiment of the present invention.

2 is a detailed block diagram illustrating a motion compensator of the first diagram.

3 is a view for explaining a process performed in the motion compensation unit a to c.

* Explanation of symbols for main parts of the drawings

105: video signal encoder 107: entropy encoder

113: video signal decoder 124: frame memory

150: motion compensation unit 20: region division unit

30: area unit displacement estimation unit 40: prediction unit

50: post-processing unit

The present invention relates to an apparatus and method for encoding an image signal, and to an apparatus and method for decoding an image signal. In particular, the present invention relates to an apparatus and method for encoding an image signal using a region-based motion vector. The present invention relates to a motion compensation video signal decoding apparatus and method for reconstructing a video signal encoded using an area unit motion vector to an original video signal.

In general, when transmitting a video signal, it is well known that digital transmission can maintain a much better image quality than analog transmission. When an image is represented in digital form, a large amount of digital data must be transmitted.

However, in the conventional transmission channel, the usable frequency range is limited. Therefore, in order to transmit a large amount of digital data, it is necessary to compress the transmitted data and reduce the amount thereof. Hybrid coding combined with temporal and spatial compression is known to be the most efficient.

Most hybrid coding techniques use motion compensated DCPM (Differential Pulse Code Modulation), 2-D Discrete Cosine Transform (DCT), quantization of DCT coefficients, VLC (variable length coding), etc. The motion compensation DPCM uses Staffan Ericsson's Fixed and Adaptive Predictors for Hybrid Predictive / to determine the difference between the current frame and the predicted value. Transform Coding, IEEE Transactions on Communication, COM-33, NO.12 (1985, December), or Ninomiy and Ohtsuka's A motion Compensated Interframe Coding Scheme for Television Pictures, IEEE Transactions on Communication, COM-30, NO.1 (January, 1982).

In addition, two-dimensional DCT reduces or eliminates spatial redundancy between image data by converting one block of digital image data, for example, a block of 8x8 pixels into a group of transform coefficient data.

This two-dimensional DCT technique is described in Chen and Pratt's Scene Adaptive Coder, IEEE Transaction on Communications, COM-32, No. 3 (March 1984), and quantizer and zigzag scanning of the transform coefficient data described above. By processing by scanning, VLC, etc., data to be transmitted can be efficiently compressed.

In the motion compensation DPCM, the motion between the current frame and the previous frame is estimated to predict the current frame from the previous frame. The motion estimation is represented as a two-dimensional vector representing the displacement between the previous frame and the current frame.

Meanwhile, various methods for estimating the displacement of an object in an image signal have been proposed. Conventional motion vector estimation methods have been proposed in two types, namely, a pixel repetition algorithm and a block matching algorithm (JR Jain et al., Displacement Measurement and Its Application in Interframe Image Coding). , IEEE Transactions on Communications COM-29, NO.12 (see December 1188, December).

In the more widely used block matching algorithm, the current frame is divided into a number of search blocks, and the size of each search block is generally in the range of 8x8 to 32x32 pixels to determine the motion vector of the search block of the current frame. To calculate the similarity between each candidate block included in the search area of the previous frame and the search block of the current frame having the same size, the average absolute error or the average is calculated for the similarity between the search block of the current frame and the candidate block of the search area Error functions such as square error are used.

In addition, the motion vector is defined as the displacement between the search block and the candidate block generating the smallest error function value.

In an actual image, an object or region is a unit of movement, rather than a block of a predetermined size. Therefore, in the motion compensation process using the conventional block matching algorithm, a blocking effect occurs at the boundary of the block, which may degrade the image quality.

Meanwhile, when an image is encoded in an area unit and transmitted to the decoder, information about an area boundary is necessary for the decoding process in the decoder.

Accordingly, an object of the present invention is to provide an apparatus and method for encoding and decoding a video signal using an area unit method, without additional information on an area boundary.

According to an aspect of the present invention, there is provided a motion compensation video signal encoding apparatus configured to determine a predicted current frame based on a current frame and a previous frame of a discrete video signal, including: dividing means for dividing a previous frame into a plurality of regions; ; Area unit displacement estimation means for providing a plurality of displacement vectors representing the motion of each of the plurality of divided regions between a current frame and a previous frame; Prediction means for moving the plurality of divided regions by a corresponding displacement vector to form an intermediate prediction frame; Using the plurality of displacement vectors, a motion vector for each pixel included in an intermediate prediction frame is provided, and a pixel value of a previous frame corresponding to the motion vector for each pixel of the intermediate prediction frame is generated as a corresponding pixel value. And post-processing means for determining the current frame.

Hereinafter, with reference to the accompanying drawings will be described in detail the present invention.

1 is a block diagram of an image encoding apparatus using the motion compensation unit 150 of the present invention.

The video signal is input to the adder 102 and the motion compensator 150 of the present invention.

In the motion compensator 150 of the image encoding apparatus according to the preferred embodiment of the present invention, the previous frame signal is received through the line L130 and reconstructed through the line L132 from the frame memory 124. After estimating the displacement vector representing the displacement between the region of the previous frame and the optimal matching region included in the current frame, the displacement vector is provided to the entropy encoder 107 through the line L134. Each area is defined as a group of pixels with similar properties.

In addition, the motion compensator 150 also determines the prediction signal for the current frame using the displacement vector and provides the subtractor 102 and the adder 115 through the line L160. The prediction signal is also predicted from the previous frame in the same manner as described with reference to FIG. 2 using the displacement vector.

The subtractor 102 subtracts the prediction signal received from the motion compensator 150 and the current frame signal, and inputs the result signal, that is, an error signal indicating a difference pixel value, to the image signal encoder 105, among which the error signal One block is encoded into a series of quantized transform coefficients by a discrete cosine transform (DCT) and the like.

After this, the quantized transform coefficients are transmitted through two paths, one of which is a combination of line length coding and variable length coding, together with a displacement vector input to the entropy encoder 107 and input through the line L134. The signal is encoded and transmitted to a transmitter (not shown) for transmission, and the other is input to the image signal decoder 113 and converted into a differential error signal reconstructed through inverse quantization and inverse transformation. The recovery of the error signal is necessary for the encoder to track the operation of the decoder of the receiver to prevent the signal recovered in the decoder from diverging from the current frame signal.

The error signal reconstructed by the image decoder 113 and the prediction signal of the motion compensator 150 are combined in the adder 115 to be a current frame signal reconstructed and stored in the frame memory 124.

2 is a detailed block diagram illustrating a motion compensator 150 of an image encoding apparatus according to an exemplary embodiment of the present invention. In the present invention, a region-based displacement vector is used for encoding and decoding processes. In other words, the displacement vector is determined in units of regions of the image. In order to reconstruct an image encoded by using the displacement vector of the region, information on the region is required in the corresponding decoder. Since the amount of data transferred from the encoder to the decoder is limited, it is more efficient to recover at the decoder than to transmit information about the region from the encoder.

The previous frame stored in the frame memory 124 of FIG. 1 is the same as the frame reconstructed by the corresponding decoder. The region obtained by dividing the restored previous frame is also the same as in the corresponding decoder. Therefore, if the current frame is encoded using a displacement vector corresponding to each region, it is possible to recover the current frame without transmitting information about the region from the encoder to the decoder.

FIG. 3 (a) shows a restored previous frame including a plurality of regions from R0 to R7.

The previous frame stored in the frame memory 124 shown in FIG. 1 is input to the area divider 20. In the region dividing unit 20, the previous frame is divided to determine a plurality of regions, for example, R7 through R0 in FIG. Various methods can be used for the segmentation process (see A.K. Jain's Fundamentals of Digital Image Processing, 1989, Prectice-Hall International).

The area information is input to the area unit displacement estimator 30, where the area information includes the location of the area and included pixels. The current frame signal represented by the input discrete image signal is also input to the area unit displacement estimator 30.

In the area unit displacement estimation unit 30, the displacement vector between the area of the current frame and the previous frame is estimated. FIG. 3 (b) shows the displacement vectors M0 to M7 between the previous frame and the current frame, and each displacement vector represents the movement of the corresponding region of the previous frame shown in FIG. For convenience, the displacement vectors having a value of 0 are assigned to the regions R0, R1, R2, and R7 representing the background; It is assumed that the displacement vectors M3 to M6 have the same value.

The displacement vector is provided through the line L134 to the predictor 40 and the entropy encoder 107 shown in FIG. The region information determined by the region divider 20 is also input to the predictor 40. The predictor 40 determines the intermediate prediction frame by moving each region of FIG. 3 (a) by the corresponding motion vector of FIG. 3 (b). In FIG. 3C, an intermediate prediction frame is shown. In FIG. 3C, an empty area, that is, an area filled with dots R11 represents a pixel without a pixel corresponding to the previous frame, and an overlapped area, that is, a shaded area R12, represents a pixel corresponding to the previous frame. Two or more pixels are shown. The remaining portion of the intermediate prediction block other than the region overlapping with the empty region indicates pixels corresponding to one pixel of the previous frame. The intermediate prediction frame often includes an empty region or an overlapping region because the intermediate prediction frame is determined by moving each region of the previous frame by the displacement vector of each region of the previous frame.

The intermediate prediction frame from the predictor 40 and the displacement vector from the region-based displacement estimator 30 are input to the post processor 50. In the post-processing unit 50, the motion vector for each pixel included in the area overlapping with the empty area is determined as described below, and then transmitted to the decoder, where the motion vector is determined by the intermediate prediction block determined in relation to the present invention. The displacement between the pixel and the pixel of the previous block is displayed.

As shown in FIG. 3 (c), in order to find the current frame predicted from the intermediate prediction frame, the movement of corresponding pixels whose displacement vector from the region dividing unit 20 does not belong to the empty region R11 or the overlapping region R12. Determined by a vector. Then, a motion vector for each pixel belonging to the empty region R11 or the overlapped region R12 is determined. In order to find a motion vector for the pixel P0 of the empty area R11 of FIG. 3C, all adjacent regions overlapping the circle C0 are determined. The circle C0 is centered on the pixel P0 and has a predetermined radius r. In this case, the area R6 and the area R7 are adjacent areas. The motion vector of the pixel P0 is determined as an average of the motion vectors of adjacent areas. Motion vectors for other pixels included in region R11 or region R12 are also determined in a similar manner. Then, the pixel value of the previous frame corresponding to the motion vector of each pixel of the intermediate prediction frame is determined as the predicted angle pixel value of the current frame.

Referring to FIG. 2 again, as described above, the pixel value of the previous frame is extracted from the frame memory 124 through the line L132 based on the motion vector determined by the post-processing unit 50. You can get it.

In the decoding apparatus corresponding to the encoding apparatus according to the embodiment of the present invention, the structure of the motion compensation block is similar to that of FIG. 2, but there is no area unit displacement estimator 30 of FIG. 2 because the displacement vector is transmitted from the encoder. Because it is provided. The motion compensator includes a region divider, a predictor, and a post processor. The function is the same as that described in the encoder.

In detail, the previous frame signal from the frame memory of the decoding apparatus according to the embodiment of the present invention is the same as the encoder described. The previous frame is input to an area divider and divided into a plurality of areas. The prediction unit determines the intermediate prediction frame using the region information from the region dividing unit and the displacement vector transmitted from the encoder (see FIG. 2). The post processor provides a predicted current frame such as an encoder. The predicted current frame is further processed by the decoder to reconstruct the current frame similar to the original video signal.

As described above, the present invention provides an apparatus and method for encoding and decoding a video signal using a region-based method without additional information on region boundaries.

Claims (12)

A motion compensation video signal encoding apparatus for determining a predicted current frame based on a current frame and a previous frame of a discrete video signal, comprising: dividing means for dividing the previous frame into a plurality of regions; Area unit displacement estimation means for providing a plurality of displacement vectors representing the movement of each of the plurality of divided regions between the current frame and the previous frame; Prediction means for moving the plurality of regions by a corresponding displacement vector to form an intermediate prediction frame; Using the plurality of displacement vectors, a motion vector for each pixel included in the intermediate prediction frame is provided, and a pixel value of the previous frame corresponding to the motion vector for each pixel of the intermediate prediction frame is generated as the corresponding pixel value and predicted. And a post-processing means for determining the current frame to which the motion compensation image signal is encoded. The intermediate image processing apparatus according to claim 2, wherein the post-processing means comprises: assigning the displacement vector as a motion vector for a corresponding pixel of a first region of the intermediate prediction frame and based on the motion vector of an adjacent pixel of the first region. Means for determining a motion vector of a pixel included in the second and third regions of the prediction frame, wherein the first region indicates a pixel corresponding to one pixel of the previous frame, and the second region indicates a previous frame. And a pixel that does not correspond to any of the pixels, and the third region indicates a pixel corresponding to two or more pixels of the previous frame. 3. The method according to claim 2, wherein the post-processing means comprises: averaging motion vectors for pixels of the first group of regions that intersect a circle of a predetermined radius centering on the pixel for which the motion vector is to be determined; Motion compensation image signal encoding apparatus using a region-by-unit motion vector, characterized by determining a motion vector for each pixel included in three regions. A motion compensation video signal encoding method using a video signal encoding apparatus for determining a predicted current frame based on a current frame and a previous frame of a discrete video signal, comprising: a first step of dividing the previous frame into a plurality of regions; Providing a plurality of displacement vectors representing the motion of each of the plurality of divided regions between the current frame and the previous frame; A third step of constructing an intermediate prediction frame by moving the plurality of regions by a corresponding displacement vector; Using the plurality of displacement vectors, a motion vector for each pixel included in the intermediate prediction frame is provided, and a pixel value of a previous frame corresponding to the motion vector for each pixel of the intermediate prediction frame is generated as a corresponding pixel value. And a fourth step of determining a predicted current frame. 5. The method of claim 4, wherein the fourth step comprises: assigning the displacement vector as a motion vector for a corresponding pixel of a first region of the intermediate prediction frame, and performing an intermediate prediction based on a motion vector of an adjacent pixel of the first region. And determining a motion vector of the pixels included in the second and third regions of the frame, wherein the first region indicates a pixel corresponding to one pixel of the previous frame, and the second region indicates a pixel corresponding to the previous frame. And a third region indicates a pixel to which no pixel corresponds, and the third region indicates a pixel to which two or more pixels of the previous frame correspond. 6. The method of claim 5, wherein the fourth step comprises: averaging motion vectors for pixels of a first group of regions that intersect a circle of a predetermined radius centering on a pixel for which the motion vector is to be determined; A motion compensation video signal encoding method comprising determining a motion vector for each pixel included in three regions. Motion compensation image signal encoding comprising splitting means for dividing a previous frame into a plurality of regions, and region-based displacement estimation means for providing a plurality of displacement vectors representing the motion of each of the plurality of divided regions between a current frame and a previous frame A motion compensation video signal decoding apparatus for determining a predicted current frame based on a plurality of motion vectors transmitted from a device and previous frames of an image signal, comprising: moving the plurality of regions by a corresponding displacement vector to form an intermediate prediction frame Prediction means; Provides a motion vector for each pixel included in the intermediate prediction frame using one or more of the displacement vectors, and generates a pixel value of the previous frame corresponding to the motion vector for each pixel of the intermediate prediction frame as the corresponding pixel value. And post-processing means for determining the predicted current frame. 8. The intermediate processing according to claim 7, wherein the post-processing means comprises: assigning the displacement vector as a motion vector for a corresponding pixel of a first region of the intermediate prediction frame, and performing intermediate prediction based on a motion vector of an adjacent pixel of the first region. And means for determining a motion vector of a pixel included in the second and third regions of the frame, wherein the first region indicates a pixel corresponding to one pixel of the previous frame, and the second region indicates a pixel corresponding to the previous frame. And a third region indicates a pixel to which no pixel corresponds, and the third region displays a pixel to which two or more pixels of the previous frame correspond. 8. The method of claim 7, wherein the post-processing means comprises: a second and a second means by averaging motion vectors for pixels in a first group of regions that intersect a circle of a predetermined radius centering on a pixel for which the motion vector is to be determined. A motion compensation video signal decoding apparatus characterized by determining a motion vector for each pixel included in three regions. A motion compensation image signal comprising splitting means for dividing a previous frame into a plurality of regions, and area-based displacement estimation means for providing a plurality of displacement vectors representing the movement of each of the plurality of divided regions between a current frame and a previous frame A motion compensation video signal decoding method for determining a predicted current frame based on a plurality of motion vectors transmitted from an encoding apparatus and previous frames of an image signal, comprising: moving the plurality of regions by a corresponding displacement vector to form an intermediate prediction frame A first step of doing; By using the plurality of displacement vectors to provide a motion vector for each pixel included in the intermediate prediction frame, and to generate the pixel value of the previous frame corresponding to the motion vector for each pixel of the intermediate prediction frame as a corresponding pixel value And a second step of determining the measured current frame. 12. The method of claim 10, wherein the second step comprises: assigning the displacement vector as a motion vector for a corresponding pixel in a first region of the intermediate prediction frame, and performing an intermediate prediction based on a motion vector of an adjacent pixel of the first region. And determining a motion vector of the pixels included in the second and third regions of the frame, wherein the first region indicates a pixel corresponding to one pixel of the previous frame, and the second region indicates a pixel corresponding to the previous frame. And a third region indicates a pixel to which no pixel corresponds, and the third region indicates a pixel to which two or more pixels of the previous frame correspond. 12. The method of claim 10, wherein the second step comprises: averaging motion vectors for pixels in a first group of regions that intersect a circle of a predetermined radius centering on a pixel for which the motion vector is to be determined; A motion compensation video signal decoding method comprising determining a motion vector for each pixel included in three regions.