KR0178226B1 - Method for correcting distort factor in image encoder - Google Patents

Abstract

The present invention relates to a method of correcting a distortion value of an image encoder which encodes using a corrected distortion value between frames instead of quantization by a quantization step size level value, and compares a prediction error between a current frame and a predicted frame and a preset distortion value. A first step of doing; A second step of encoding by DCT and quantization and motion prediction based on the comparison result of the first step; when the prediction error between the frames is greater than or equal to the preset distortion value, the current frame and the predicted frame are zigzag scanned, respectively. The difference between the pixels of the same position is calculated based on the comparison result of the third and third steps of obtaining the difference value between the pixels at the same position between the current frame and the predicted frame, and summing the difference values. If the sum of the difference values is greater than or equal to the preset distortion value, a fourth step is performed in which the preset distortion value is reset to the sum value, thereby effectively utilizing the transmission channel capacity.

Description

Distortion value correction method of video encoder

1 is a schematic block diagram of a conventional video encoding apparatus.

2 is a distortion value correcting apparatus of a video encoder according to the present invention.

* Explanation of symbols for main parts of the drawings

220: distortion signal correction unit 240, 260: first and second switch unit

300: frame forming part 310: DCT

320, 440: quantization unit 330 inverse quantization unit

340: reverse DCT 350: the adder

370: motion estimation unit 380: motion compensation unit

500: VLC

The present invention relates to a distortion value correction method of an image encoder, and more particularly, to a distortion value correction method of an image encoder that encodes using a corrected distortion value between frames instead of quantization by a quantization step size level value.

In the case of high-definition television (HDTV), a video signal is represented in a digital form, thereby generating a large amount of data. Data compression is essential in order to transmit this generated amount of data through a limited frequency domain.

FIG. 1 is a schematic block diagram for data compression of a conventional image encoder. Referring to FIG. 1, a previous frame is stored in the frame memory 116, and a frame formed by the frame forming unit 102 is The subtractor 104 and the motion predictor 118 are provided.

The motion predictor 118 predicts the motion based on the frame formed by the frame forming unit 102 and the previous frame of the frame memory 116, and the predicted motion is moved from the previous frame by the motion compensator 120. The compensation is provided to the adder 114 and the subtractor 104.

The subtractor 104 obtains the difference value based on the frame provided from the frame forming unit 102 and the motion displacement provided from the motion compensating unit 120, and the DCT 106 and the quantizer 108 perform the DCT operation. Coefficients are quantized and provided to the dequantizer 110 and the VLC 122.

In this case, the inverse quantizer 110 inversely quantizes the coefficient quantized by the quantizer 108 and then converts the inverse DCT by the inverse DCT 112 to the adder 114.

The adder 114 combines the previously predicted frame and the inverse DCT data and stores the previous frame in the frame memory 110 as a previous frame.

The VLC 122 obtains the run value and the level value for the quantized coefficient and provides the variable length coded data to the output buffer 124. The output buffer 124 performs the variable length coded by the VLC 122. The data is output to the decoder according to the available channel capacity.

For example, when the amount of data to be output is large, the output buffer 124 increases the quantization step size (QP) level value and then provides it to the quantization unit 108.

Accordingly, the quantization unit 108 quantizes the DCT coefficients according to the applied quantization step size (QP) level.

However, when the inter-frame movement change is severe, a lot of data is temporarily provided to the output buffer 124, so that the output buffer 124 has an increased quantization step size (QP) level than the existing quantization step size (QP) level. As a result, the quantization unit 108 is interrupted.

In the video encoding apparatus configured as described above, encoding is performed based on a prediction error and a preset distortion value between the current frame and the predicted frame. For example, when the prediction error is smaller than the preset distortion value Th, Compressed data is transmitted via quantization and VLC.

On the contrary, if the prediction error between the current frame and the predicted frame is greater than or equal to the preset distortion value Th, the data encoded based on the quantization and the VLC are transmitted based on the difference signal between the restored frame and the current frame. do.

However, for example, in the case where the movement change in the frame is large, the predetermined distortion value is fixed, and thus there is a disadvantage in that more efficient encoding cannot be performed.

Accordingly, the present invention has been made to solve the above disadvantages, and an object of the present invention is a distortion value correction method of encoding an image signal based on a corrected distortion value between frames instead of quantization by a quantization step size level value. To provide.

According to the present invention for achieving the above object, there is provided an encoding apparatus for a video apparatus that encodes based on a difference signal between a frame predicted by a previous frame and a current frame and a frame currently input; Comparing a preset distortion value of a prediction error between the current frame and a predicted frame; A second step of performing encoding by DCT and quantization and motion prediction when the prediction error between the frames is smaller than a preset distortion value as a result of the comparison in the first step; As a result of the comparison of the first step, if the prediction error between the frames is greater than or equal to a preset distortion value, the pixels between the current frame and the predicted frame are located at the same position while zigzag scanning the current frame and the predicted frame, respectively. A third step of obtaining a difference value and comparing the calculated distortion value with the calculated difference value; As a result of the comparison in the third step, when the sum of the difference values between the pixels at the same position is smaller than the preset distortion value, the preset distortion value is maintained and the difference between the pixels at the same position is summed. If a value is greater than or equal to the preset distortion value, the preset distortion value may be reset to the summed value.

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

2 is a block diagram of a threshold correction apparatus of a video encoder of the present invention.

Referring to FIG. 2, the distortion signal correcting unit 220 describes a switch control signal for selecting one coding method based on a preset distortion value Th, which is a prediction error between a current frame and a predicted frame. It is configured to be provided to the first and second switching units 240 and 260.

The first switching unit 240 is configured as a switching device to provide or block the difference signal between the frames based on the control signal of the distortion signal correcting unit 220.

The second switching unit 260 is configured as a switching device for providing a differential signal between the current frame and the restored frame to the quantization unit 440 based on the control signal of the distortion signal correcting unit 220.

The frame forming unit 300 configures an image signal input in the form of a bit stream in units of frames to be provided to the distortion signal corrector 220, the subtractor 390, the motion predictor 370, and the subtractor 420. do.

The DCT 310 is configured to convert the difference value between the frames outputted through the first switching unit 240 into a DCT coefficient by using a cosine function to be provided to the quantization unit 320.

The quantization unit 320 is configured such that the coefficients DCT by the DCT 310 are quantized and provided to the inverse quantization unit 330 and the VLC 500.

The inverse quantization unit 330 is configured to inversely quantize data quantized by the quantization unit 320 and provide the inverse DCT 340, and the inverse DCT 340 is inverse quantized data by the inverse quantization unit 330. The reverse DCT is then configured to be provided to the adder 350.

The adder 350 is configured to combine the inverse DCT data from the inverse DCT 340 and the movement displacement whose motion is compensated to provide the frame memory 360 and the subtractor 420 as a previous frame.

The frame memory 360 stores a previous frame provided from the adder 350, and provides a frame reconstructed from the frame memory 360 to the motion predictor 370 and the motion compensator 380 for motion compensation. It is configured to be.

The motion predictor 370 is configured to predict the motion based on the current frame and the previous frame of the frame memory 360, and then provide the predicted data to the motion compensator 380.

The motion compensator 380 compensates for the motion based on the data predicted by the motion predictor 370 and the previous frame of the frame memory 360, and compensates for the compensated prediction frame by the distortion signal corrector 220 and the adder ( 350 and subtractor 390.

The subtractor 390 is configured to provide a difference signal between the current frame and the frame predicted by the motion compensator 380 to the DCT 310. The subtractor 420 is configured to obtain a difference signal between the current frame and the frame combined by the adder 350 and provide the difference signal to the quantizer 440.

The quantization unit 440 is configured to perform quantization on the difference signal between the frame (the current frame and the restored previous frame) and provide the quantized data to the VLC 500.

The VLC 500 includes a lookup table and is configured to provide a variable length coded data to the decoder based on the run value and the level value provided in the quantizers 320 and 440.

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

First, the encoded previous frame is stored in the frame memory 360, and the motion compensation unit 380 predicts the motion between the previous frame restored from the frame memory 360 and the currently input frame. ) Will be provided.

The motion compensator 380 provides the predicted frame to the subtractor 390 and the distortion signal corrector 220 based on the motion displacement of the motion predictor 370 and the previous frame of the frame memory 360.

The subtractor 390 obtains a difference value between a frame currently being input and a frame whose motion is predicted, and then selectively provides the difference signal to the DCT 310 according to a switching control signal provided from the distortion signal corrector 220 to be described later. Done.

That is, the distortion signal correcting unit 220 compares the prediction error between the current frame and the predicted frame provided from the frame forming unit 300, and the preset distortion value TH.

As a result of comparing the prediction error between the frames and the preset distortion value TH, if the prediction error between the frames is smaller than the preset distortion value TH, the first switching unit 240 and the second switching unit 260 are replaced. The switching control performs encoding by the above-described DCT and quantization, inverse quantization, motion prediction, and the like.

On the contrary, when the prediction error between the frames and the preset distortion value TH is compared, and the prediction error between the frames is greater than or equal to the preset distortion value TH, the current frame and the prediction provided from the frame forming unit 300 are predicted. While performing the zigzag scanning of each frame, the difference value between pixels at the same position between the current frame and the predicted frame is obtained.

Here, the obtained difference value is added and compared with the preset distortion value TH.

That is, when the sum of the difference values between the pixels at the same position is smaller than the preset distortion value, the preset distortion value is maintained, and the sum of the difference values between the pixels at the same position is the preset value. If greater than or equal to the distortion value, the predetermined distortion value is reset to the summed value.

The reset value thus reset is set to the next reference value.

For example, as a result of the comparison of the distortion signal correcting unit 220, when the sum of the difference values between the pixels at the same position is smaller than a preset distortion value, the distortion signal correcting unit 220 may generate a predetermined switching control signal in a first manner. , 2 switching unit 240, 260 to be provided.

Therefore, since the switches of the first and second switching units 240 and 260 are switched to the fixed contact a, the differential signal between the frames is DCT and quantized by the DCT 310 and the quantization unit 320 and then inversed. Provided to the quantization unit 330 and the VLC 500.

The inverse quantization unit 330 inversely quantizes the data quantized by the quantization unit 320 and provides the inverse DCT 340. In the inverse DCT 340, the inverse DCT is performed by inverse DCT to add an adder ( 350).

The adder 350 provides the motion displacement compensated from the previous frame and the inverse DCT data to the frame memory 360 and the subtractor 420.

In this case, since the switch of the second switching unit 260 is switched to the fixed contact a, the quantization unit 440 does not operate. Accordingly, the VLC 500 variable length encodes the quantized coefficients provided from the quantizer 320 and outputs the data to the decoder.

On the contrary, when the comparison result of the distortion signal correcting unit 220 indicates that the sum of the difference values between the pixels at the same position is greater than or equal to the preset distortion value TH, the distortion signal correcting unit 220 corresponds to this value. The switching control signal is provided to the first and second switching units 220 and 260.

That is, since the switch of the first switching unit 240 is switched to the fixed contact a, the predicted difference signal between the frames is DCT and quantized by the DCT 310 and the quantization unit 320 and then dequantized. Provided to section 330 and VLC 500.

At the same time, the switch of the first switching unit 440 is switched to the fixed contact b and at the same time, the next input frame data is not provided to the DCT 310, and the previously encoded image frame is transferred through the adder 350. To a subtractor 420.

At this time, since the switch of the second switching unit 260 is switched from the fixed contact (a) to the fixed contact (b), the adder 420 obtains a difference value between the current frame and the frame provided from the adder 350, and then The second switching unit 260 is provided to the quantization unit 440.

Here, the quantization unit 440 quantizes only the difference between the current frame and the reconstructed frame, and then provides the difference to the VLC 500.

Accordingly, the VLC 500 performs quantization only on the difference signal between the current frame and the reconstructed frame, and variably encodes the quantized data and outputs the quantized data to the decoder.

As described above, the present invention has an effect of increasing the signal-to-noise ratio between frames while efficiently using the transmission channel capacity by encoding using the corrected distortion value between frames instead of quantization by the quantization step size level value. .

Claims (1)

An encoding apparatus of a video apparatus that encodes based on a difference signal between a frame predicted between a previous frame and a current frame and a frame currently input; A first step of comparing a prediction error between the current frame and the predicted frame and a preset distortion value TH; A second step of performing encoding by DCT and quantization and motion prediction when the prediction error between the frames is smaller than a preset distortion value TH as a result of the comparison in the first step; As a result of the comparison in the first step, if the prediction error between the frames is greater than or equal to a preset distortion value TH, the same position between the current frame and the predicted frame is zigzag scanned while the current frame and the predicted frame are respectively zigzag scanned. A third step of obtaining a difference value between the pixels, and comparing the calculated difference value with the predetermined distortion value (TH) while adding up the difference value; As a result of the comparison in the third step, when the sum of the difference values between the pixels at the same position is smaller than the preset distortion value TH, the preset distortion value is maintained and the difference between the pixels at the same position is maintained. A fourth step of causing the predetermined distortion value TH to be reset to the summed value when the sum of the values is greater than or equal to the preset distortion value TH. Value correction method.