KR20070053402A - Method for compensating motion of video compression system - Google Patents

Abstract

The present invention relates to a motion compensation method of an image compression system. To this end, unlike the conventional method of compensating only planar motion by using motion vectors according to longitudinal and transverse motions in an image compression system, an input image After extracting the magnified and reduced image data of the current frame macroblock from the signal, and comparing the extracted enlarged and reduced image with the image of the previous and subsequent frames, Compute the motion vectors and weights that contain the information about the length, width, and extent of the expansion, reduction, and inverse of the motion vectors and weights, and calculate the inverse difference with the macroblock of the current frame. By performing motion compensation, in addition to the planar motion compensation in the image compression system, That will be done to compensate for the jikim.

Image compression system, motion compensation, motion vector, macroblock

Description

Motion compensation method of video compression system {METHOD FOR COMPENSATING MOTION OF VIDEO COMPRESSION SYSTEM}

1 is a block diagram of an image compression system suitable for compensating three-dimensional motion according to the present invention;

2 is a flowchart illustrating a process of compensating three-dimensional motion in an image compression system according to the present invention;

3 is a view illustrating an enlarged and reduced image of a macroblock according to the present invention;

4 is a diagram illustrating a motion vector for an enlarged and reduced image according to the present invention.

<Description of the symbols for the main parts of the drawings>

102: subtractor 104: DCT unit

106: quantizer 108: inverse quantizer

110: reverse DCT 112: adder

114: frame memory device 116: motion estimation unit

118: motion compensation unit

The present invention relates to a motion compensation method of an image compression system, and more particularly, to a motion compensation method of an image compression system suitable for compensating not only planar but three-dimensional motion in an image compression system.

As is well known, three types of techniques are generally used to compress an image using a compression scheme such as MPEG. That is, a lossless compression method such as variable length coding (VLC) that allocates smaller bits to frequently used pattern data, a method of reducing data volume using IDCT according to spatial continuity, and movement using temporal continuity A motion compensation method is used.

Among them, the motion compensation method is a method of reducing the amount of data using only the information and the difference because the current time image is similar to the previous or subsequent time image. The conventional motion compensation method compares a macroblock of a current frame with a previous or subsequent frame, detects an image having the most similar size, and transmits the position information (motion vector) and the difference.

In this case, since the comparison unit of the image compares only the image having the same size as the original macroblock, this method compensates only when an arbitrary object moves planarly on the screen, and when an arbitrary object moves three-dimensionally on the screen Even if the motion compensation is performed, it is a factor that causes an error in the three-dimensional part.

Accordingly, an object of the present invention is to provide a motion compensation method of an image compression system capable of compensating not only planar motion but also stereoscopic motion in an image compression system such as MPEG. .

In order to achieve the above object, the present invention provides a motion compensation method of a video compression system that compensates and compresses a motion of an input video signal, extracting enlarged and reduced image data of a current frame macroblock from the input video signal. And comparing the extracted enlarged and reduced images with images of previous and subsequent frames, and longitudinal, transverse distance information, and enlarged and reduced information with the most similar previous and subsequent frame images for the extracted enlarged and reduced images. Calculating a motion vector and weights including information about the degree, and inversely calculating a motion vector and weights including information about the length, width, and distance of the previous and subsequent frame images And performing motion compensation by calculating the macroblock of the current frame and the inversely calculated difference. It provides a method of motion-compensated video compression system including a system.

The above and other objects and various advantages of the present invention will become more apparent from the preferred embodiments of the present invention described below with reference to the accompanying drawings by those skilled in the art.

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

A key technical aspect of the present invention is that in an image compression system, unlike the conventional method of compensating only planar motion by using motion vectors according to longitudinal and transverse motion, enlarged and reduced image data of a current frame macroblock from an input video signal. After extracting and comparing the extracted enlarged and reduced images with the images of the previous and subsequent frames, the longitudinal and horizontal distance information and the extent of the enlargement and reduction with the most similar previous and subsequent frame images for the extracted enlarged and reduced images By calculating the motion vectors and weights that contain information about, inversely calculating these motion vectors and weights, and performing the motion compensation by calculating the inverse difference between the macroblocks of the current frame. The object of the present invention can be easily achieved.

1 is a block diagram of an image compression system suitable for compensating three-dimensional motion according to the present invention, which includes a subtractor 102, a DCT unit 104, a quantizer 106, an inverse quantizer 108, and an inverse DCT unit ( 110, an adder 112, a frame memory 114, a motion estimator 116, and a motion compensator 118.

Referring to FIG. 1, an input video signal composed of one frame is grouped into predetermined blocks (for example, 8 * 8 macroblocks, etc.) from an input memory or a frame buffer (not shown), and sequentially read out to be a motion estimation unit ( 116 and subtractor 102. At the same time, an image signal one frame before the current frame is read from the frame memory 114 and provided to the motion estimation unit 116.

In addition, the motion estimation unit 116 indicates a motion vector indicating in which direction the current frame has moved compared to the previous frame by comparing the previous frame signal read from the input memory and the frame memory 114 with the current frame, respectively. Detect.

Herein, the process of the motion estimator 116 will be described in more detail. The motion estimator 116 enlarges and reduces the current frame macroblock (for example, an 8 * 8 macroblock) from an input video signal. After extracting the image data and comparing the extracted enlarged or reduced image with the image of the previous frame and the subsequent frame from the frame storage unit 114, the motion vector and the weight of the most similar previous / post-frame image are calculated. In addition, the motion vector including the information on the longitudinal, transverse distance, and information about the degree of enlargement and reduction and the weight of the motion vector and the corresponding weight are inversely calculated, and the difference between the macroblock and the inverse information of the current frame is calculated, and the DCT is used. Then, motion compensation is performed through the quantization process.

Next, the subtractor 102 calculates a difference between the input image signal block and the image signal block provided from the motion compensator 118, and the signal output from the subtractor 102 is obtained by the DCT unit 104 in the frequency domain. By transforming the transform coefficients, the correlation in the frame is removed, and the quantizer 106 adaptively quantizes the signal output from the DCT unit 104.

The inverse quantizer 108 inversely quantizes the signal output from the quantizer 106, and inverse-DCT-converted through the inverse DCT unit 110 to restore the original signal. The adder 112 combines the output signal of the inverse DCT unit 110 and the signal output from the motion compensator 118, and the combined signal is output and recorded in the frame memory 114.

Next, in the image compression system having the above-described configuration, the enlarged and reduced image data of the current frame macroblock is extracted, the images of the previous frame and the subsequent frame are compared, and the motion vector of the most similar previous / post frame image. After the weights are calculated, a process of performing the motion compensation by performing the inverse calculation of the enlargement / reduction, the motion vectors and the weights, and storing the difference with the current frame macroblock will be described.

2 is a flowchart illustrating a process of compensating three-dimensional motion in an image compression system according to the present invention. The motion compensation method according to the present invention will be described with reference to these drawings.

Referring to FIG. 2, the motion estimation unit 116 extracts enlarged and reduced image data of a current frame macroblock (eg, an 8 * 8 macroblock) from an input video signal (step 202). For example, to obtain an enlarged or reduced image as illustrated in FIG. 3, a nearest neighbor interpolation (NNI) matching a value of the nearest coordinate in the transformed coordinate, or a bilinear interpolation (BI) using three adjacent coordinate information. This can be used and the appropriate method can be chosen taking into account the complexity by accuracy and computation.

The motion estimation unit 116 compares the enlarged and reduced image extracted in the step 202 with the image of the previous frame and the subsequent frame from the frame memory 114 (step 204). Here, the comparison range may be selected within the comparison range of the predetermined area in consideration of the amount of calculation and the like.

Next, the motion estimation unit 116 calculates a motion vector and a weight with the most similar previous / post frame image after the comparison in the step 204 (step 206). Here, the motion vector is recorded in addition to the longitudinal and transverse distance information, including information on the degree of enlargement and reduction. As an example, FIG. 4 is a diagram illustrating an example of a motion vector including information about the degree of enlargement and reduction. In the current macro block, a previous 7 * 7 sized image moves -2 vertically and -1 horizontally. , Which is enlarged by 1 pixel. In addition, if the motion estimation is referenced in two or more frames before or after, weight information for the motion is calculated.

The motion estimator 116 inversely calculates the motion vector including the information about the length, the horizontal distance, and the information about the enlargement and reduction, and the weight thereof (step 208). This inverse operation information is stored as information of one macroblock.

In addition, the motion estimation unit 116 calculates a difference between the macroblock of the current frame and the information inversely calculated in step 208 and stores the difference (step 210). Afterwards, the motion compensator 118 performs the motion compensation through the DCT and the quantization process using the information (step 212). In this case, the header of the macroblock having the above-described processes may include motion compensation information considering three-dimensional motion.

Therefore, in addition to the planar motion as well as the three-dimensional motion according to the motion vector and the weight according to the video signal input from the image compression system and the information on the length, width and distance from the previous and subsequent frame images, and information on the degree of enlargement and reduction, the image is compensated for. Can be compressed.

In the foregoing description, the present invention has been described with reference to preferred embodiments, but the present invention is not necessarily limited thereto. Those skilled in the art will appreciate that the present invention may be modified without departing from the spirit of the present invention. It will be readily appreciated that branch substitutions, modifications and variations are possible.

As described above, the present invention is different from the conventional method of compensating only planar motion by using motion vectors according to longitudinal and transverse motion in an image compression system, and enlarges and reduces image data of a current frame macroblock from an input image signal. After extracting and comparing the extracted enlarged and reduced images with the images of the previous and subsequent frames, the longitudinal and horizontal distance information and the extent of the enlargement and reduction with the most similar previous and subsequent frame images for the extracted enlarged and reduced images By calculating the motion vectors and weights including the information about, inversely computing these motion vectors and weights, and performing the motion compensation by calculating the inverse difference with the macroblock of the current frame, Three-dimensional motion compensation is performed by using motion vectors including cross-lateral information and zooming information. Can be done.

Claims (3)

A motion compensation method of an image compression system for compensating and compressing a motion of an input video signal, Extracting magnified and reduced image data of a current frame macroblock from the input video signal; Comparing the extracted enlarged or reduced image with an image of a previous or subsequent frame; Calculating a motion vector and a weight including information on longitudinal, transverse distance information, and information on the extent of the enlargement and reduction of the extracted enlarged and reduced image, which are most similar to previous and subsequent frame images; Inversely calculating a motion vector and weights including information on the length, width, and scale of the previous and subsequent frame images; Performing motion compensation by calculating a macroblock of the current frame and the inversely calculated difference Motion compensation method of the image compression system comprising a. The method of claim 1, The extracting of the enlarged and reduced image data is performed by Nearest Neighbor Interpolation (NNI) or Bilinear Interpolation (BI). The method of claim 1, The motion compensation information is included in the header of the macroblock, the motion compensation method of the image compression system.

Images