KR0174444B1 - Motion compensated apparatus for very low speed transmission - Google Patents

Abstract

The present invention relates to a motion compensation device for ultra low-speed transmission that compensates for motion by selectively using a large block block and a sub-block block when a motion is severe during video encoding. First motion prediction means for obtaining a motion displacement and a signal-to-noise ratio between the current frame and the previous frame; Second motion prediction means for obtaining a motion displacement and a signal-to-noise ratio between the current frame and the previous frame based on the subcompartment block; A post-processing means for providing a compensated motion displacement based on the signal-to-noise ratio of the first and second motion prediction means, and if the motion-compensating unit has too much motion, by performing motion compensation using a subcompartment block. Better motion compensation

Description

Motion compensation device for ultra low speed transmission

1 is a code and device diagram for compressing a conventional video signal.

2 is a motion compensation device for ultra low speed transmission according to the present invention.

FIG. 3 is a block diagram for explaining in detail the motion prediction coding unit of FIG.

4 is a diagram showing a large block and a sub block for explaining the present invention.

* Explanation of symbols for main parts of the drawings

100: first motion prediction means 120: motion estimation unit

140: motion compensation unit 160,240: PSNR calculation unit

220: motion prediction encoder 300: post-processing means

320: comparator 340: switching unit

The present invention relates to a motion compensation device for ultra-low speed transmission, and particularly a motion compensation device for ultra-low speed transmission that compensates for the motion by selectively using a large block block and a sub-block block when the motion is severe during video encoding. It is about.

In general, when a video signal is to be transmitted as a digital signal such as a video telephone or a high-definition television (HDTV), the digital data transmitted by using a high rate data compression method must be encoded in order to reduce the enormous amount of data.

In such compression encoding, an apparatus for encoding an image signal substantially reduces an amount of data to be transmitted. In particular, encoding is performed by using spatial and temporal correlation of the image signal during transmission.

That is, the encoding apparatus removes spatial redundancy of the video signal by using transform coding such as DCT (Discrete Cosine Transform) and removes temporal redundancy of the video signal by using differential encoding through motion estimation and prediction. Therefore, the video signal can be compressed efficiently.

As a method for compressing massive data as described above, there are various compression schemes. In particular, a hybrid coding scheme combining probabilistic coding and temporal and spatial compression is known to be the most efficient.

The hybrid coding technique uses motion compensated DCPM (differential pulse code modulation), two-dimensional DCT (discrete cosine transform), quantization of DCT coefficients, variable length coding, and the like.

In motion compensation DPCM (Differential Pulse Code Modulation), a differential signal is generated according to the motion of the object between the current frame and the previous frame.The current frame is predicted as the previous frame according to the motion of the object estimated between the current frame and the previous frame. This predicted motion displacement is represented by a two-dimensional motion displacement, that is, a motion vector.

On the other hand, the method of estimating the displacement of a pixel of an object includes a block unit method and a pixel unit motion estimation method. A pixel unit motion estimation method is used for a representative pixel of a pixel that can represent adjacent pixels in a set (frame). The motion vector is sent to the transmitter. That is, a feature point is selected from all the pixels included in the previous frame, and a motion vector for the selected feature point is determined. In this case, each motion vector is a displacement between one feature point of the previous frame and a corresponding matching point of the current frame.

In detail, the matching point for each feature point is searched in the search area of the current frame, wherein the search area is defined as a preset area including the location of the search point.

On the contrary, in the block-based motion estimation, the optimal matched block is determined by comparing the blocks of the previous frame using the blocks of the current frame (16 × 16), and the interframe displacement vector of the entire block for the current frame is determined. Measured.

1 is a block diagram of a conventional apparatus for compressing an existing image by using a large block block, including: a subtractor 10, a DCT 12, a quantizer (Q, 14), an inverse quantizer (IQ, 16), and an inverse DCT (IDCT 18). And a frame memory 20, a motion estimator 22, a motion compensator 24, an adder 26, and an entropy encoder 28.

As described above, in the conventional block-based motion estimation, an optimal matching block is determined by comparing the blocks of the previous frame using the large block block (16 × 16) of the current frame and at the same time, the entire block for the current frame is determined. The interframe displacement vector is measured.

Based on the measured displacement between frames and the difference between the current frames, the encoder is encoded through a DCT, quantization, and entropy encoder, and then output to the transmitter.

However, when the motion is compensated by the large block block (16 × 16), for example, if the block unit (the large block block) to compensate is too large, effective motion compensation is not effective. The signal is of poor quality.

Accordingly, the present invention has been made to solve the above disadvantages, and an object of the present invention is to divide a frame of an image signal into a large block block and a small block block, and to use ultra-low speed transmission to selectively compensate for the motion by using these blocks. It relates to a motion compensation device for.

According to the present invention for achieving the above object, an encoding apparatus comprising means for providing a frame of a video signal reconstructed by encoding a frame of a video signal:

First motion predicting means for obtaining a motion displacement and a signal-to-noise ratio between the current frame and the encoded previous frame based on a large block block; Second motion predicting means for obtaining a motion displacement and a signal-to-noise ratio between the current frame and the encoded previous frame based on the subcompartment block; And post-processing means for providing the compensated motion displacement based on the signal-to-noise ratios of the first and second motion prediction means.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the illustrated drawings.

2 is a block diagram showing a motion compensation device for ultra low speed transmission according to the present invention.

Referring to FIG. 2, the motion compensation apparatus for ultra-low speed transmission of the present invention includes a first motion predicting means 100, a second motion predicting means 200, and a post-processing means 300.

The first motion predictor 100 calculates a motion displacement and a signal-to-noise ratio between the current frame and the encoded previous frame based on the large block block, the motion estimator 120, the motion compensator 140, and the PSNR calculator 160. It consists of.

In more detail, the motion estimation unit 120 determines the current frame from the previous frame based on a large block block (by 16 × 16 pixels) between the current frame to be input and the previous frame of the frame memory (reference numeral 20 in FIG. 1). The motion is estimated and provided to the motion compensator 140 to be described later.

The motion compensator 140 includes a PSNR calculator 160 and post-processing means which describe the motion displacement compensated by the motion estimator 120 and the motion displacement compensated between the previous frame of the frame memory (reference numeral 20 in FIG. 1). Provided to the switching unit 340 of (300).

The PSNR calculation unit 160 provides a result value of the motion displacement compensated by the motion compensation unit 140 and the signal-to-noise ratio (PSNR) between the current frames to one input terminal of the comparator 330 of the post-processing means 300 which will be described later. do.

The second motion predicting means 200 includes a motion prediction encoder 220 and a PSNR calculator 240 for obtaining a motion displacement and a signal-to-noise ratio between the current frame and the encoded previous frame based on the subcompartment block.

The motion prediction encoder 220 includes a block separation stage 222, a subcompartment motion estimation unit 224, a subcompartment motion compensator 226, and a block integrator 228.

The block separation stage 222 divides the large partition block (search area) of the current frame to be divided into first, second, third, and fourth sub-blocks (search area) and provides the sub-compartment motion estimation step 224 to be described later. do.

The subcompartment motion estimation stage 224 estimates the motion based on each of the first, second, third, and fourth subcompartment blocks separated by the block separation stage 222 and the previous frame, and the subcompartment motion compensation stage described later ( 226).

The subcompartment motion compensator 226 compensates for the motion based on the motion displacement and the previous frame of each of the first, second, third, and fourth subcompartment blocks estimated by the subcompartment motion estimation stage 224, and will be described later. Provided to the integrated stage 228.

As shown in FIG. 4, the subcompartment motion estimator 224 includes first, second, third and fourth motion estimators 224a through 224d that estimate motions of the first, second, third and fourth regions of the large block block.

It is composed of

The sub-compartment motion compensation stage 226 compensates the motion based on the motion displacement estimated by the first, second, third, and fourth motion estimators 224a-224d and the previous frame, and provides the block integration unit 228 to be described later. And first, second, third and fourth motion compensators 224a through 224d.

The block integration stage 228 integrates each of the first, second, third, and fourth subcompartment blocks compensated by the subcompartment motion compensator 226 into the original macrocompartment block, and integrates the motion displacement into the PSNR calculator. 240 and the post-processing means 300 is provided to the switching unit 340.

The PSNR calculation unit 240 is the other input terminal of the comparator 330 of the post-processing means 300 which describes the result of the motion displacement compensated by the motion prediction encoding unit 220 and the signal-to-noise ratio (PSNR) between the current frames. to provide.

The post-processing means 300 includes a comparator 320 comparing the motion error values obtained by the first and second motion predicting means 100 and 200 and the first and second according to a comparison result of the comparator 320. It is composed of a switching unit 340 for selectively switching and outputting the motion displacement value output from the motion prediction means (100, 200).

Referring to the operation of the present invention configured as described above in detail as follows.

First, the current frame of the video signal input through the line 1 and the previous frame stored in the frame memory (20 in FIG. 1) are provided to the first and second motion predicting means 100 and 200, respectively.

The motion estimation unit 120 of the first motion predicting means 100 estimates the motion of the current frame input through the line 1 from the previous frame of the frame memory 20 and provides the motion compensation unit 140 to the motion compensation unit 140. do.

In the motion estimation process, an optimal matching block is obtained by comparing a large block of blocks (search blocks) consisting of 16 × 16 pixels in the current frame with blocks of a previous frame.

Subsequently, the motion compensator 140 converts the motion displacement compensated for the motion based on the estimated motion displacement and the previous frame of the frame memory 20 to the PSNR calculator 160 and the switching unit of the post-processing means 300. 340 is provided at one side end.

The PSNR calculator 160 calculates a motion displacement compensated by the motion compensator 140, a noise (noise) value for the current frame, and a signal to noise ratio (PSNR) as follows.

MxN is the block size of the frame; x, y is coordinate data of each pixel; I (x, y) is the frame coordinate of the original image; P (x, y) is the coordinate to find; 255 is defined as the level value of the pixel.

Here, the calculated signal-to-noise ratio first value is provided later as one terminal of the comparator 320 in order to select a coding scheme having a good coding rate.

In addition, the motion predictor encoding unit 220 of the second motion predictor 200 estimates the motion of the current frame input through the line 1 from the previous frame of the frame memory 20, and estimates the estimated motion. The compensation is provided to the PSNR calculation unit 240.

Referring to FIG. 3, the motion prediction encoder 220 will be described in more detail. The current frame (the same frame provided to the first motion prediction means 100) input through the line 1 is the block separation stage 222. Is provided.

In the block separation stage 222, the large partition block (search area) of the current frame is divided into first, second, third, and fourth sub-blocks (search area) as shown in FIG. Is provided to the motion estimators 224a-224d of the corresponding subcompartment motion estimation stage 224.

The motion estimators 224a-224d estimate motion based on the previous frames of each of the separated 1,2,3,4 subcompartment blocks (search area) and the frame memory (20 in FIG. 1). The motion displacement is provided to the motion compensators 224a-224d of the subcompartment motion compensator 226.

In the motion compensators 226a to 226d, the first, second, third, and fourth subcompartment blocks (search area) and the previous frames of the frame memory (20 of FIG. 1) are estimated. Compensation for the 2,3,4 subcompartment blocks is then provided to the block integration stage 228.

The block integration stage 228 integrates each of the first, second, third and fourth subcompartment blocks provided from the motion compensators 226a through 226d of the subcompartment motion compensator 226 into the original macrocompartment block. This movement displacement is provided to the PSNR calculator 240 and the switching unit 340.

The PSNR calculation unit 240 obtains the signal-to-noise ratio between the frame integrated by the block integrator 228 and the current frame by the above-described equation.

As such, the calculated signal to noise ratio second value is provided to the other terminal of the comparator 320.

Therefore, the comparator 320 performs the motion prediction so that the first and second values calculated by the first and second motion prediction means 100 and 200 are low, that is, the motion compensation error is low and the signal to noise ratio (PSNR) is high. The treatment means is selected.

For example, when the movement is not severe, the comparator 320 may provide a switching signal for switching the switch of the switching unit 340 to the contact point a.

At this time, the switch of the switching unit 340 is switched to the contact (a) and at the same time the movement displacement provided by the first motion predicting means 100 is provided to the subtractor 10 of FIG.

On the contrary, when the movement is severe, the comparator 320 provides a switching signal for switching the switch of the switching unit 340 to the contact point b.

The switch of the switching unit 340 is switched to the contact (b) and at the same time the movement displacement provided from the second motion predicting means 100 is provided to the subtractor 10 of FIG.

Accordingly, the subtractor 10 shown in FIG. 1 encodes the difference signal obtained by subtracting the current frame and the motion displacement applied through the switching unit 340 of the post-processing means 300 through the DCT and the quantization and entropy encoders. It will then be output to the transmitter.

As described above, in the present invention, when the motion compensation is performed by using the large compartment block, when the motion compensation is too large and the motion is severe, the motion compensation is performed by using the subcompartment block so that better motion compensation can be achieved. As a result, more data can be compressed.

In addition, this invention is not limited to the said Example, A deformation | transformation is possible in the range which does not deviate from the technical summary of this application.

Claims (7)

A coding apparatus comprising means for encoding a frame of an image signal to provide an encoded signal and means for encoding the encoded frame to provide a frame of an image signal, comprising: a current frame and the previous frame based on a large block block First motion predicting means for obtaining a motion displacement and a signal-to-noise ratio between the two; Second motion prediction means for obtaining a motion displacement and a signal-to-noise ratio between the current frame and the previous frame based on the subcompartment block; And post-processing means for providing the compensated motion displacement based on the signal-to-noise ratios of the first and second motion prediction means. The first motion predicting means 100 includes a motion estimator 120 for estimating motion based on the large block of the current frame and the previous frame, and a motion displacement based on the estimated displacement and the previous frame. And a PSNR calculator (160) for calculating a signal-to-noise ratio value between the compensated motion displacement and the current frame. The motion compensation device of claim 2, wherein the giant block is implemented in units of 16 x 16 pixels. The second motion predicting means includes: a motion prediction encoder 220 for obtaining a compensated motion displacement between the current frame and the previous frame based on the subcompartment block, and the motion displacement compensated by the motion prediction encoder 220. And a PSNR calculator 240 for calculating a signal-to-noise ratio value between the current frame and the current frame. The motion prediction encoding unit 220 includes a block separation stage 222 that separates the large block of the current frame into first, second, third, and fourth subblocks, and the separated 1,2,3,4 subblocks. Ultra-low-speed transmission comprising a sub-compartment motion compensation stage 226 for compensating for motion of each sub-compartment block based on a block and a previous frame, and a block integrating unit 226 integrating the compensated sub-compartment block. Motion Compensation Device. The motion compensation device of claim 5, wherein the sub-compartment block is implemented in 8 × 8 pixel units. The post-processing means 300 may include a comparator 320 comparing the signal-to-noise values obtained by the first and second motion predicting means 100 and 200 and the second comparator 320 according to a comparison result of the comparator 320. The motion compensation device for ultra-low speed transmission, characterized in that the switching unit 340 for selectively switching the movement displacement value output from the first and second movement prediction means (100, 200).