KR0174441B1 - Full motion image encoder by using adaptive motion compensation - Google Patents

Abstract

The present invention relates to a video encoding apparatus that selectively uses a large block and a sub-block of an image frame during ultra low-speed transmission. To solve this problem, the present invention performs encoding between frames based on the large block of a video frame. First video encoding means for providing motion displacement with respect to the frame, second video encoding means for providing motion displacement with respect to the current frame by performing inter-frame encoding based on the subcompartment block, and first and second video encoding means. By including a motion displacement output means for outputting only one of the motion displacements for the current frame encoded by the motion displacement, and a frame memory for storing the previous frame based on the motion displacement output by the motion displacement output means, Efficient encoding can be performed.

Description

Video Coding Device Using Adaptive Motion Compensation

1 is a diagram of an encoding apparatus for compressing a conventional video signal.

2 is a video encoding apparatus using adaptive motion compensation according to the present invention.

3 is a block diagram for explaining in detail the motion predictor of FIG.

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

* Explanation of symbols for main parts of the drawings

200: first video encoding means 300: second video encoding means

320: motion estimation unit 400: motion displacement output means

410,420: first and second bit counter 430: comparator

440,450: switch stage

The present invention relates to a video encoding apparatus using adaptive motion compensation. More particularly, the present invention relates to a video encoding apparatus using adaptive motion compensation by selectively using a giant block and a subcombination block of an image frame.

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.

Therefore, various data compressor methods have been developed, and the double motion compensation inter coding method is known as one of the effective compressor methods.

This encoding technique uses temporal redundancy between two adjacent frames. For example, the current frame predicts motion based on motion estimation from a previous frame.

The predicted motion may be represented as a motion displacement of pixels between the previous frame and the current frame.

Thus, when describing a block matching algorithm (BMA), one of motion displacement estimation techniques proposed in the art, the current frame is divided into a number of search blocks of the same size, and the size of the search block Typically has a 16 × 16 pixel range.

That is, a similar calculation is performed between the search block of the current frame and each of a plurality of candidate blocks having the same size as the search block to determine the motion displacement with respect to the search block in the current frame.

The candidate block is included in the search area in the previous frame, and the size of the search area is generally larger than the search block.

In this case, an error function such as an absolute mean error (MSE) or a square mean error (MSE) is used to measure the similarity between the search block of the current frame and each candidate block in the search area.

And, by definition, the motion displacement represents the displacement of the motion between the search block and the optimal matching block, that is, the candidate block causing the minimum error or difference.

In such motion estimation, it is desirable to find only one minimum absolute mean error in all search areas corresponding to the search block, but many have the same minimum error during block matching.

FIG. 1 is a coding apparatus for data compression of an existing image using a large block block (16 × 16). The encoding apparatus includes a subtractor 100 and a subtractor 100 that calculate a difference between motions between a current frame and a previous frame. The quantizer (Q, 104) which quantizes the DCT transform coefficient from the DCT 102 by using a cosine function and the DCT transform coefficient from the DCT 102 according to the step size. A dequantizer (IQ) 106 for restoring the quantized transform coefficients to the original transform coefficients, an inverse DCT for inverse discrete cosine transforming the inversely transformed transform coefficients, and a frame memory for storing previous frames ), A motion estimated by the motion estimator 112 and the motion estimator 112 for predicting the motion between the input current frame and the previous frame stored in the frame memory 110 using a 16 × 16 macroblock. Based on displacement A motion compensator 114 for compensating the previous frame stored in the frame memory 110, an adder 116 for adding the compensated motion displacement and the inverse DCT data to the frame memory 110, and a quantizer An entropy encoder 108 for encoding the transform coefficients quantized at 104 using a variable length encoding method or the like and outputting a motion displacement to a transmitter.

When the conventional block-based motion estimation technique is applied as described above, the search block of the current frame is determined to determine the motion vector (displacement) of the large block of the current frame (or the search block, 16 × 16). The optimal matching block is determined through similar calculation among a plurality of candidate blocks having the same size as the search block, and the interframe displacement vector of the entire block is measured with respect to the current frame.

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 using the large block block (16 × 16) in the case of a very slow motion of 64 Kbps, the effective motion compensation is not effective. Remarkably falls.

Accordingly, an object of the present invention is to provide a video encoding apparatus using adaptive motion compensation, which encodes a frame of an image signal into a large block block and a sub block block and selectively transmits the encoded information bits according to a counter result. It is.

According to the present invention for achieving the above object, there is provided a video encoding apparatus comprising: first video encoding means (200) for performing motion coding for a current frame by performing inter-frame encoding on the basis of a large block block; Second image encoding means (300) for performing inter-frame encoding based on the subcompartment block to provide a motion displacement with respect to the current frame; Motion displacement output means (400) for outputting only one motion displacement information of the motion displacements with respect to the current frame encoded by the first and second image encoding means (200, 300); It characterized in that it comprises a frame memory 500 for storing the previous frame based on the movement displacement output by the movement displacement output means 400.

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

2 is a diagram of a video encoding apparatus using adaptive motion compensation according to the present invention.

Referring to FIG. 2, the video encoding apparatus using the adaptive motion compensation according to the present invention includes a first image encoding unit 200, a second image encoding unit 300, a motion displacement output unit 400, and a frame memory 500. ).

The first image encoding unit 2000 includes a subtractor 202, a DCT 204, a quantizer 206, an inverse quantizer 208, an inverse DCT 210, a motion estimator 212, and a motion compensator 214. , An adder 216 and an entropy encoder 218.

The second image encoding means 300 includes a subtractor 302, a DCT 304, a quantizer 306, an inverse quantizer 308, an inverse DCT 310, an adder 312, an entropy encoder 314, and a motion. The prediction unit 320 is configured.

As shown in FIG. 3, the motion predictor 320 includes a block separation stage 322, a motion estimation stage 324, a motion compensation stage 326, and a block integration stage 328.

The block separation stage 322 is configured to separate the sub-blocks of the current frame to be input into the A, B, C, and D regions as shown in FIG.

The motion estimation stage 324 is based on the respective A, B, C, D regions separated from the block separation stage 322 and the previous frame provided from the frame memory 500. And a western block block estimator (3241-3244) for providing the estimated motion displacement to the entropy encoder (314) and the motion compensation stage (326).

The motion compensator 326 may perform motion compensation on the basis of the motion displacement of each of the areas A, B, C, and D estimated by the motion estimator 324 and the previous frame of the frame memory 500. And a subcompartment block compensating stage 3221-3264 for providing a frame formed in the C, D regions to the block integrating stage 328.

The block integrator 328 integrates the subcompartment block compensated by the motion compensator 326 into the original macrocompartment block to be provided to the subtractor 302 and the adder 312 of the second image encoding means 300. It is composed.

The motion displacement output means 400 includes first and second bit counters 410 and 420, a comparator 430, and first and second switches 440 and 450.

The first bit counter 410 is configured to counter the bit for the motion displacement information provided from the entropy encoder 218 of the first image encoding unit 200 to be provided to one side of the comparator 430.

The second bit counter 420 is configured to counter bits of the motion displacement information provided from the entropy encoder 420 of the second image encoder 300 to be provided to the other end of the comparator 430.

The comparator 430 switches the first and second switching signals for selecting any one of the encoding means 200 and 300 whose counters have a small number of information bits of the motion displacement counted from the first and second bit counters 410 and 420. And configured to be output to the stages 440 and 450.

The switch stage 440 is switched so that the encoded previous frame may be stored in the frame memory 500 based on the first and second switching signals of the comparator 430. Based on the first and second switching signals, any one of the encoded motion displacement information is switched to be output to the transmitter.

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

Referring to FIG. 2, the frame data is also stored in the frame memory 500 in the same manner as the frame stored in the frame memory of the decoder (not shown).

Subsequently, the current frame of the input image signal is provided to the subtractor 202 and the motion estimation unit 212 of the first image encoding unit 200. At the same time, the current frame is provided to the motion predictor 320 and the subtractor 302 of the second image encoding unit 300, respectively.

For example, in the first image encoding unit 200, the current frame and the motion displacement compensated by the motion estimator 212, the motion compensator 214, and the frame memory 500 are subtracted by the subtractor 202. Find the difference value.

The difference value indicates that the motion displacement encoded by the DCT 204, the quantizer 206, and the entropy encoder 218 is the contact A of the switch 450 of the motion displacement output means 400 and the motion displacement output means ( To the first bit counter 410 of 400.

The first bit counter 410 counts the number of bits of the motion displacement information and then provides it to one side of the comparator 430.

In addition, the DCT and the quantized transform coefficient are inversely quantized and inversely dictated by the inverse quantizer 208 and the inverse DCT 210, and then provided to the adder 216, and at the same time, the motion estimation unit 212 and the motion compensator. The movement displacements compensated by 214 and frame memory 216 are provided to adder 216.

The adder 216 provides a frame formed by combining the two signals to the contact point A of the switch 440 of the motion displacement output means 400.

At this time, in the motion estimation unit 212, the motion compensator 214, and the frame memory 500, a large block of blocks (search blocks) composed of 16 x 16 pixels in the current frame is compared with the blocks of the previous frame. The displacement of motion with respect to the registration block is obtained.

On the contrary, the second image encoding unit 300 obtains the motion displacement compensated on the basis of the motion predictor 320 and the frame memory 500, and obtains the difference value between the frames by the subtractor 302. The operation of the unit 320 will be described later.

In other words, the difference value is the motion displacement coded by the DCT 304, the quantizer 306, and the entropy encoder 314. The second bit counter 420 of the motion displacement output means 400 contacts the switch 450. Provided by (B).

The second bit counter 420 counts the number of information bits of the movement displacement and then provides the counter to the other end of the comparator 430.

The DCT and the quantized transform coefficients are inversely quantized and inversely dictated by the inverse quantizer 308 and the inverse DCT 310, and then provided to the adder 312 and compensated by the motion predictor 320. The motion displacement is provided to the adder 216 above.

The adder 216 combines the two signals to provide the movement displacement to the switch B of the output means 400.

At this time, the operation process of the motion predictor 320 and the frame memory 500 will be described in detail. An optimal matching block is compared with the blocks of the previous frame by comparing a subcompartment block (search block) consisting of 8 × 8 pixels in the current frame. Then we find the motion displacement with respect to the motion displacement.

Referring to FIG. 3 in more detail, the current frame is provided to the block separating end 322, and the block separating end 322 shows the subcompartment block (search area) of the current frame as shown in FIG. It is divided into A, B, C, and D areas and provided to the motion estimation stage 324.

Here, the motion estimation unit 324 estimates the motion based on the A region of the separated frame and the previous frame of the frame memory 500, and then provides the motion to the subcompartment block compensator 3331, B, C, D. The region is also estimated based on the previous frame of the frame memory 500 and then provided to the subcompartment block compensation stage 3262-3264.

The subcompartment block compensating stage 3221-3264 compensates for the motion based on the estimated motion displacement and the previous frame of the frame memory 500, and then subblocks the subcompartment blocks of the A, B, C, and D regions. 328).

In the block integration stage 328, the frame implemented for the motion displacements of the compensated A, B, C, and D regions is integrated into a large block block so that the adder 312 and the subtractor 302 of the second image encoding means 300 are used. To be provided.

Referring to FIG. 1, the comparator 430 of the motion displacement output means 400 compares the number of bits for the motion displacement provided from the first and second bit counters 410 and 420.

As a result of the comparison, when the number of motion displacement bits encoded by the first image encoding means 200 is less than the number of motion displacement bits encoded by the second image encoding means 300, the comparator 430 switches the switch stage 440. 450 to provide the first control signal.

At this time, since the switch of the switch stage 440 is switched to the contact point A, the image frames combined by the adder 212 are stored as the previous frame data in the frame memory 216.

Therefore, in the next encoding, the previous frame is used.

In addition, since the switch of the switch stage 450 is switched to the contact point A, the motion displacement encoded by the entropy encoder 218 of the first image encoding unit 200 is output to the transmitter through the switch stage 450. will be.

On the contrary, when the number of bits of the motion displacement encoded by the second image encoding means 300 is less than the number of bits of the motion displacement encoded by the first image encoding means 200, the comparator 430 is switched. 440 and 450 provide a second control signal.

At this time, since the switch of the switch stage 430 is switched to the contact point B, the image frames combined by the adder 312 are stored as previous frame data in the frame memory 216. Therefore, as described above, in the next encoding, the previous frame is used.

In addition, since the switch of the switch stage 450 is switched to the contact point B by the second control signal, the motion displacement encoded by the entropy encoder 314 of the second image encoding means 300 is switched. It is output to the transmitter through.

As described above, the present invention has an advantage of enabling efficient encoding by selectively using a large block block and a sub block block in ultra low speed transmission.

Although the present invention has been described with reference to specific embodiments, it will be understood by those skilled in the art that various changes may be made without departing from the spirit and scope of the appended claims.

Claims (2)

An image encoding apparatus, comprising: first image encoding means (200) for performing inter-frame encoding based on a large block block to provide a motion displacement with respect to a current frame; Second image encoding means (300) for performing inter-frame encoding based on the subcompartment block to provide a motion displacement with respect to the current frame; Motion displacement output means (400) for outputting only one motion displacement information of the motion displacements with respect to the current frame encoded by the first and second image encoding means (200, 300); And a frame memory (500) for storing the previous frame based on the motion displacement output by the motion displacement output means (400). According to claim 1, wherein the movement displacement output means (400); First and second bit counters (410, 420) for countering the number of bits for the motion displacement information provided from the first and second image encoding means (200, 300); A first selecting one of the encoding means from among the first and second image encoding means (200, 300) based on a comparison of the number of bits of the motion displacement counted from the first and second bit counters (410, 420). A comparator 430 for providing a switching control signal; A switch stage 440 for switching switching so that the encoded previous frame may be stored based on a first switching signal of the comparator 430, and the encoded motion based on a second switching signal of the comparator 430 The video encoding apparatus using the adaptive motion compensation, characterized in that the switch stage for switching switching so that any one of the displacement information can be output.