KR0162196B1 - Post processing method and device for removing blocking artifact - Google Patents

Abstract

The present invention relates to eliminating the blocking effect caused by quantization in image encoding using block transform, and particularly to blocking artifacts during image encoding suitable for image signal band compression and extension of a digital high-definition TV and digital VTR for USA. Post Processing method and apparatus for removing Blocking Artifact).

In particular, the present invention configures an adaptive low pass filter at the rear end of the inverse DCT and the dynamic compensation processing unit at the decoder side, and adaptively filters the blocking occurrence part according to the size of the quantization level, thereby eliminating blocking artifacts and thereby reducing image quality degradation. It is possible to maintain the high quality of the image at all times.

Description

Post-processing method and apparatus for removing blocking effect

1 is a block diagram of a transmitting unit of a general video data image bandwidth system.

2 is a block diagram of a receiver of a conventional video data update system.

3 is a block diagram of a receiver of an image data accommodating system according to the present invention;

4 is a detailed view of an adaptive low pass filter unit according to an embodiment of the present invention.

5 is a filter characteristic diagram.

6 is a detailed view of an adaptive low pass filter unit according to another embodiment of the present invention.

* Explanation of symbols for main parts of the drawings

21: variable length decoder 22: inverse quantization (Inverse Quantization)

23: reverse DCT (Discrete Cosine Transform) unit 24: delay unit

25: Motion Compensation 26: Refreshing

27: demultiplexer section 28: adaptive low pass filter section

281: decoding unit 282: filter switching unit

285: decoding and ROM address generator 286: multiplexer

287: filter coefficient storage ROM 288: filtering unit

The present invention relates to a method and an apparatus for removing a blocking effect caused by quantization in encoding of image data using a block transform in a decoder. In particular, the present invention relates to an American digital TV (ATV) and The present invention relates to a post processing method and apparatus in a decoder for removing blocking artifacts occurring during encoding of image data suitable for video signal band compression and extension of a digital VTR.

1 is a configuration diagram of a transmitter of a general video data image band companding system.

In the transmitter, which is an encoder of the image band compression system as shown in FIG. 1, video signals including Y, U, and V components are encoded by the DCT converter 12 and the quantizer 13, and the inverse quantizer 15 And decoded by the inverse DCT unit 16. The newly formed data passing through the delay unit 17 and the compensation unit 18 is input to the adder A1 and used to generate differential data. The buffer controller 19 generates a quantization level coefficient determined according to the amount of data stored in the buffer (not shown) and transmits the quantization level coefficient to the quantizer 13 and the receiver. The band compression coded data via the variable length coding unit 14 is inputted to the apparatus of FIG. 2 which performs band extension through channel coding (encoding). 2 is a configuration diagram of a receiver of a conventional video data update system, and is currently configured in an American high definition television (ATV). The variable length decoding unit 21 restores the variable length encoded input signal to its original length. The variable length coded data is restored by the dequantization unit 22 to the original data size. The data whose length and size are restored is reduced to the original image data by the inverse DCT unit 23, and a delay unit 24, a dynamic compensator 25, and a refreshing unit 26 to compensate for image quality deterioration in the case of a video. Then, the signal is added to the signal passing through the feedback loop composed of the < Desc / Clms Page number 5 >

However, in the conventional video data compression system as described above, since DCT and quantization are performed in a predetermined block unit (usually 8 pixels x 8 pixels), when the quantization is coarse due to a large amount of data input from the encoder side as a transmitter, the decoder restores the data. Blocking artifacts occur at such a block level, resulting in very serious image quality degradation.

An object of the present invention for solving the above problems is to provide a post-processing method to prevent the deterioration of image quality by removing the blocking artifacts.

Another object of the present invention is to provide a post-processing device that prevents deterioration of image quality by removing blocking artifacts.

An object of the present invention as described above is a post-processing method in a decoder for decompressing compressed image data in units of blocks of a predetermined size by using a quantization level coefficient used for data compression. Long-decoding, inversely quantizing the variable long-decoded data using the quantization level coefficient, restoring the inversely quantized data by using an inverse DCT transform, and performing quantization level coefficients transmitted from an encoder side. And a low pass filtering of the restored data with a filtering coefficient determined according to the present invention.

Another object of the present invention is to provide a decoding apparatus for receiving a quantization level coefficient used for data compression and decompressing compressed image data in units of a predetermined size, wherein the compressed image data is received and expanded to its original length. A variable long decoding unit connected to the variable long decoding unit, an inverse quantization unit for restoring data decoded according to the quantization level coefficient to its original size, and an inverse DCT for restoring the dequantized data to original image data information The filter coefficients corresponding to the magnitudes of the quantization level coefficients are stored, and the restored data is input using the filter coefficients corresponding to the magnitudes of the quantization level coefficients after receiving the data and the quantization level coefficients restored from the inverse DCT unit. The post-processing device for removing the blocking effect comprising an adaptive low-pass filter for adaptively low-pass filtering Is achieved.

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

3 is a configuration diagram of a receiver of an image data accommodating system according to the present invention, and the same components as those of the receiver of a conventional image data accommodating system are denoted by the same reference numerals as those of FIG. 2, and a detailed description thereof is omitted. As shown, the image data compression system of the present invention has an adaptive low pass filter unit at the rear end of the adder A2 that sums the output of the reverse DCT unit 23 and the output of the dynamic compensator 25 in the conventional image data compression system. (28) is further comprised.

4 is a detailed view of the adaptive low pass filter unit 28 according to an embodiment of the present invention. Adaptive Low Pass Filter (ALPF) 28 according to an embodiment of the present invention is a plurality of filters (FLT1, FLT2, ..., ..., FLTN) having different filter characteristics, respectively Any one filter of the low pass filter group according to the configured low pass filter group, the decoding unit 281 for detecting and decoding the quantization coefficient transmitted from the encoder side, and the quantization coefficient value output from the decoding unit 281. It consists of a filter switching unit 282 to select. In addition, when the quantization level coefficient value from the encoder side is less than or equal to the predetermined value (that is, if the quantization step size is small), since the original data information is almost maintained, the blocking effect does not occur, and thus bypasses the filter. It also includes a bypass line (e) in which the filter switching unit 282 is directly connected to the adder (A2).

5 is a filter characteristic diagram showing the characteristics of each filter of the adaptive low pass filter unit 28 according to the present invention.

The adaptive low pass filter 28 of the present invention configured as described above decodes the quantization level coefficient transmitted from the encoder side by the decoding unit 281 and applies the filter switching unit 282 to the filter switching unit 282. ) Selects either the bypass line e or the plurality of filters FLT1, FLT2, ..., ..., FLTN in accordance with the quantization level coefficient value. That is, since the blocking effect does not occur when the amount of data input from the buffer control unit 19 of FIG. 1 is small and the quantization level is reduced, the adaptive low pass filter unit 28 does not pass the quantized data through the filter. It is controlled to be input to the demultiplexer 27 as it is. On the other hand, if the amount of data input to the encoder control unit is large and the quantization level needs to be increased, the blocking effect is severe. Therefore, the quantized data is filtered by selecting a filter having appropriate characteristics according to the amount of data. do.

Therefore, the blocking artifact generation part, that is, the data difference between the block and the block displayed on the screen is reduced, so that the blocking artifact is eliminated. As a result, the adaptive low pass filter 28 selects a filter that is adaptively different in characteristics depending on the degree of data, and thus can always maintain a high picture quality.

6 is a detailed view of an adaptive low pass filter unit 28 according to another embodiment of the present invention. The ROM 287 stores filter coefficients and the ROM 287 by decoding the quantization level coefficients from the encoder side. A decoding and ROM address generating unit 285 for generating an address control signal of the control unit, a multiplexer for controlling the filtering unit 288 by reading filter coefficients stored in the filtering unit 288 and the ROM 287 for filtering the quantized data. 286. The adaptive low pass filter unit 28 configured as described above decodes the quantization coefficient transmitted from the encoder side to generate a signal capable of controlling the address of the ROM 287 in which the filter coefficients are stored and outputs the signal to the multiplexer 286. The multiplexer 286 generates an appropriate ROM address according to the decoded quantization coefficient value and reads out the filter coefficient. This filter coefficient is sent to the filter unit 288 through the multiplexer 286 and low pass filters the input data from the adder A2. In this case, if the quantization coefficient value is large, a filter coefficient having a gentle characteristic is selected, and an appropriate filter coefficient is selected according to the quantization coefficient value, thereby eliminating blocking artifacts, thereby maintaining a high image quality at all times.

As described above, the post-processing method and apparatus for removing the blocking effect according to the present invention has the effect of always maintaining the high quality of the image while eliminating the blocking artifacts to prevent image degradation.

Claims (7)

A post-processing method in a decoder which expands compressed image data in units of blocks of a predetermined size by using a quantization level coefficient used for data compression, comprising: variable length decoding the received image data; Dequantizing the variable-length decoded data using the quantization level coefficient; Restoring the dequantized data using an inverse DCT transform; And low-pass filtering the reconstructed data with a filtering coefficient determined according to a quantization level coefficient transmitted from an encoder side. 2. The method of claim 1, wherein the reconstructing step compensates the inverse DCT-converted data and uses the data of a predetermined block generated by the compensating to perform data inputted after differential pulse code modulation. Post-processing method for removing the blocking effect, characterized in that it further comprises the step of restoring. The method of claim 1, wherein the filtering comprises selecting one of a plurality of filters having different filter coefficients according to the magnitude of the quantization level coefficient value, and filtering the input data using the selected filter. Post-processing method for removing the blocking effect, characterized in that. A decoding apparatus for receiving a quantization level coefficient used for data compression and decompressing compressed image data in units of blocks of a predetermined size, comprising: a variable length decoding unit configured to receive the compressed image data and extend the original image data to an original length; An inverse quantization unit connected to the variable length decoding unit and restoring the decoded data to its original size according to the quantization level coefficient; An inverse DCT unit for restoring the dequantized data into original image data information; Stores filter coefficients corresponding to the magnitude of the quantization level coefficient, receives the recovered data and the quantization level coefficient from the inverse DCT unit, and adapts the restored data using the filter coefficient corresponding to the magnitude of the quantization level coefficient. And a post-processing device for removing a blocking effect including an adaptive low pass filter unit for low pass filtering. 5. The post-processing apparatus according to claim 4, further comprising a feedback loop means for receiving data from the inverse DCT unit and performing dynamic compensation and generating data before differential pulse code modulation. 6. The apparatus of claim 4 or 5, wherein the adaptive low pass filter unit is connected to an adder, each of the plurality of filters having different characteristics, a decoding unit for decoding a quantization level coefficient value, and a decoded quantization level coefficient of the decoding unit. When the value is less than or equal to the predetermined value, the inverse DCT data is bypassed to the demultiplexer, and when the quantization level coefficient value is greater than or equal to the predetermined value, the blocking effect is selected by selecting one of a plurality of filters according to the coefficient value. Post-processing unit for removal. 6. The apparatus of claim 4 or 5, wherein the adaptive low pass filter unit decodes a quantization level coefficient value from a transmitter, a filter coefficient storage unit storing a plurality of filter coefficients, and a decoded quantization level coefficient of the decoding unit. A filter coefficient selection unit for selecting any one of a plurality of filter coefficients stored in the filter coefficient storage unit according to the value, and a filter coefficient selection unit for filtering the inverse DCT and dynamic compensation image data according to the filter coefficient selected Post-processing device for removing the blocking effect, characterized in that consisting of a low pass filter.