KR20000002907A - Loss section restoring mehtod using projection - Google Patents

Abstract

PURPOSE: A method to restore a loss section is provided to normally restore a loss section without degrading of screen quality of an image encoding system using projection. CONSTITUTION: A loss section restoring comprises the steps of: projecting to produce the average value of the image elements existed at the predetermined projection direction by inputting the image signal of one frame; restoring to restore the image elements in a loss section existed at each predetermined projection direction by each average value produced according to the projection direction; supposing direction of edge to suppose the direction of edge of the loss section by operating the original image element value a(i,j) and the projected-restored image element value b(i,j); and selectively outputting of an image signal to selectively output the restored image signal toward the supposed edge direction.

Description

Loss Compartment Restoration Using Projection

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image encoding system, and more particularly, to a method for restoring loss division using projections effective for extracting and expressing edge components present in an image signal.

Most still image and video encoding standards currently used internationally use partition unit processing and variable length coding to obtain high data compression efficiency. These systems are very sensitive to errors that occur when retrieving or transmitting compressed data. This is because an error of one bit in the coded bitstream can affect the entire partition without being limited to one pixel in the reconstructed image data, and may damage several partitions until resynchronization is performed. Therefore, the systems for restoring the loss compartment generally include a part for detecting the loss compartment and a part for restoring the detected loss compartment.

Common techniques for restoring loss compartments require complex computational and iterative operations, such as techniques for predicting discrete cosine transform dc coefficients, minimizing POCS (Projections onto Convex Sets) theory, or minimizing specific objective functions. It can be divided into techniques.

FIG. 1 illustrates an example of a screen partition for explaining a method of restoring a loss compartment using a DCT dc coefficient. In Fig. 1, the A and B compartments represent the error-free compartments (blocks or macroblocks), and the D compartments represent loss compartments. For example, the DC coefficient of the loss section D2 is restored by adding the dc coefficient of the A2 section and the dc coefficient of the B2 section by dividing by two by using the DCT dc coefficient. The reconstruction algorithm used in the DCT dc coefficient prediction technique is simple, but has a disadvantage in that the reconstruction quality of the loss compartment is very degraded. On the other hand, the technique using POCS or the objective function improves the quality of the reconstructed loss block, but it requires a very large amount of computation and iterative operation. In addition, nonstationarity of an image signal such as an edge is not effective when a POCS or an objective function is introduced.

Accordingly, an object of the present invention is to provide a loss partition restoration method using a projection that can normally restore loss partitions without deterioration of image quality in an image encoding system.

It is still another object of the present invention to provide a loss compartment restoration method using a projection that can restore a loss compartment by a simple calculation process when a loss compartment occurs due to a transmission error in an image encoding system.

In the present invention for achieving the above object in the loss compartment restoration method using a projection,

A projection process of inputting an image signal of one frame to calculate an average value of pixel values existing in a preset projection direction;

A restoration process of restoring the pixels in the loss compartment existing in each projection direction with respective average values calculated according to the projection directions;

An edge direction estimation process of estimating the edge direction of the loss compartment by computing the original pixel value a (i, j) and the projected-restored pixel value b (i, j);

And a video signal selection output process of selecting and outputting a video signal reconstructed in the estimated edge direction.

1 is an exemplary block diagram illustrating a loss compartment restoration method using a DCT dc coefficient.

2 is a partial configuration of the loss compartment recovery system according to an embodiment of the present invention.

3 is an exemplary view illustrating a screen partition for explaining a loss compartment restoration method using projection according to an exemplary embodiment of the present invention.

4 is a series of loss block processing flow chart performed in the loss block recovery unit 300 of FIG.

Hereinafter, an operation according to an exemplary embodiment of the present invention will be described in detail with reference to the accompanying drawings. In the following description and the annexed drawings, numerous specific details are set forth in order to provide a more general understanding of the invention, such as the number of projection directions, loss compartments, and size of normal compartments. It will be apparent to those skilled in the art that the present invention may be practiced without these specific details. And detailed descriptions of well-known functions and configurations that may unnecessarily obscure the subject matter of the present invention will be omitted.

2 is a diagram illustrating a part of a loss compartment restoration system according to an exemplary embodiment of the present invention, and FIG. 3 is a diagram illustrating an example of screen division for explaining a loss compartment restoration method using projection according to an embodiment of the present invention. It is. 4 illustrates a series of loss compartment processing flowcharts performed by the loss compartment restoration unit 300 of FIG. 2.

Hereinafter, the loss compartment restoration method using projection according to an embodiment of the present invention will be described in detail with reference to FIGS. 2 to 4. First, in FIG. 2, the image decoder 100 inputs an encoded bit string to perform an inverse process. The original screen image is restored and output by decoding. The error detector 200 detects whether an error occurs in the coded data transmitted by inputting the coded bit string and outputs the detected error. When the error detection signal is input from the error detector 200, the loss block restoration unit 300 reads an image signal of one frame from the frame memory 400 and projects and restores the loss block pixels in the edge direction. Output to memory 400. The decoded image data output from the image decoder 100 is stored in the frame memory 400.

Hereinafter, the projection-restoration processing of the loss compartment performed by the loss compartment restoration unit 300 will be described in detail with reference to FIGS. 3 and 4.

First, projection is a process of obtaining an average value of pixel values existing along a specific direction, and reconstruction is a process of replacing an average value calculated by the projection with pixel values of pixels in a loss compartment. That is, the projection is to obtain the average value by averaging the values of the pixels existing along the solid line as shown in FIGS. 3A and 3B, and the reconstruction is the loss compartment L1 as shown in FIG. ) Is a process of replacing the values of the pixels present in the image with the average value. If there is an edge component within a given n × n partition (defined as 8 × 8 in the embodiment of the present invention), the result of the projection-reconstruction along the edge direction will be the one closest to the original partition. Therefore, by comparing the result of projection-restore for any section with the original pixel values, the direction having the closest value can be estimated as the edge direction for the section.

In the embodiment of the present invention, the original pixel value is a (i, j) And the projection-restored pixel value b (i, j) By the following equation, an approximation of the original section and the restoration section is estimated by the following equation (1). Therefore, in the loss compartment restoration method according to the embodiment of the present invention, the loss compartment may be restored only by the operation of calculating integer addition and absolute value.

Hereinafter, a method of restoring a loss compartment by using Equation 1 will be described with reference to FIG. 4. First, when the error detection signal is input from the error detector 200, the loss compartment restoration unit 300 receives the frame memory ( 400, a video signal of one frame is read. In operation 510, the loss compartment restoration unit 300 calculates an average value of pixel values existing in the set projection direction. The " set projection direction " refers to a projection direction that is set in advance in order to restore the loss compartment, and includes a plurality of projection directions ranging from 0 ° to 180 ° such as 0 °, 22.5 °, 45 °, and 67.5 °. Can be set and used. As described above, the loss compartment restoring unit 300, which calculates an average value of the pixels existing in the preset projection direction, restores the pixels of the loss compartment L1 to the average value calculated according to the projection direction in step 520. In operation 530, the loss compartment restoration unit 300 estimates the edge direction of the loss compartment L1 by comparing the values obtained by Equation 1 above. That is, the loss compartment restoration unit 300 compares the sums obtained by substituting the projection-restored pixel values b (i, j) calculated for each set projection direction into Equation 1 and having the smallest projection direction. Is estimated in the edge direction. The loss compartment restoring unit 300 estimates the edge direction and then selects only the image signal restored in the edge direction estimated in step 540 and outputs the image signal to the frame memory 400. In this way, by estimating the edge direction and projecting and restoring the loss compartment, the loss compartment closest to the original compartment can be obtained.

As described above, the present invention uses the projection-reconstruction technique that considers the sensitivity of human vision, and estimates the edge direction, and restores the pixels in the loss compartment in the estimated edge direction, thereby maintaining a good image quality. In addition, since the edge of the loss compartment can be restored with a low calculation amount, there is an advantage of simplifying the loss compartment restoration algorithm.

Claims (4)

In the loss compartment restoration method using the projection, An edge direction estimating process for estimating the edge direction of the loss segment in one frame image signal by using projection each time an error of the transmitted image signal is detected; And calculating a mean value of the pixels present in the estimated edge direction, and restoring the pixel value in the loss compartment by using the calculated mean value. The method of claim 1, wherein the edge direction estimation process sets the projection direction when the sum calculated by Equation 2 has the smallest value to the edge direction. a (i, j) = original pixel value, b (i, j) = projected-restored pixel value. In the loss segment restoration method using the projection in the image encoding system, A projection process of inputting an image signal of one frame to calculate an average value of pixel values existing in a preset projection direction; A restoration process of restoring the pixels in the loss compartment existing in each projection direction with respective average values calculated according to the projection directions; An edge direction estimation process of estimating the edge direction of the loss compartment by computing the original pixel value a (i, j) and the projected-restored pixel value b (i, j); And a video signal selection output process of selecting and outputting an image signal reconstructed in the estimated edge direction. The method of claim 3, wherein the edge direction estimation process sets the projection direction when the sum calculated by Equation 3 has the smallest value to the edge direction.