Method of image filtering based on comparison of difference between selected pixels

Abstract

A method and apparatus to remove a block effect and a ring effect appearing in a compression-coded image is disclosed. The present invention is especially applicable to an image compression-coded at a low bit rate. In particular, the present invention includes a variety of masks for the removal of the block/ring effect. Thus, one mask is select for a pixel to be filtered. Moreover, candidate pixels to be averaged with the pixel to be filtered is selected from the pixels of the selected mask to better maintain the details of the image, and a weight is applied to improve the calculation speed of the averaging operation. More particularly, the masks according to the present invention have longer tabs toward adjacent blocks at boundaries of the given block to remove the block and ring effect. In one embodiment, a pixel of an image is filtered according to a filtering methodology. The filtering methodology includes selecting at least four pixels of successive pixels according to a position of a pixel to be filtered. The selected at least four pixels include the pixel to be filtered, at least one of the selected at least four pixels is in a first block, and three of the selected at least four pixels are in a second block adjacent to the first block. The filtering methodology includes at least one comparison. The comparison compares a determined value and a threshold. The threshold is based on a quantization information of at least a portion of the image including the pixel to be filtered, the determined value is based on a difference value, and the difference value is based on two of the successive pixels.

Description

Claims

1. A filtering method for a pixel P of a block B in a reconstructed image, comprising:

selecting one of a plurality of filtering masks based upon a position of said pixel P in said block B; and

averaging said pixel P and candidate pixels within the selected filtering mask.

2. A method of claim 1, wherein each of the plurality of filtering masks has 8 tabs.

3. A method of claim 1, wherein a filtering mask with more tabs toward a block adjacent said block B is selected.

4. A method of claim 1, further comprising selecting a pixel within the selected filtering mask P_m as a candidate pixel if the value of the pixel P_m meets a predetermined condition.

5. A method of claim 4, wherein the pixel P_m is selected as a candidate pixel if the absolute value of the difference between said pixel P and pixel P_m is less than a threshold value.

6. A method of claim 5, wherein the threshold value is calculated by an equation below,

δ=k×q

7. A method of claim 6, wherein the value of k is 1.0 for filtering boundary pixels of said block B and 0.6 for filtering pixels within the boundary pixels of said block B.

8. A method of claim 4, further comprising adding a weight value to said pixel P prior to the averaging, wherein said weight value is based upon a number of pixels P_m not selected as candidate pixels.

9. A method of claim 8, wherein said weight value is the number of pixels P_m not selected multiplied by the value of said pixel P.

10. A filtering apparatus to filter a pixel P of a block B in a reconstructed image, comprising:

a filtering masking unit selecting one of a plurality filtering masks based upon a position of said pixel P in said block B; and

an averaging unit averaging said pixel P and candidate pixels within the selected mask.

11. An apparatus of claim 10, wherein each of the plurality of filtering masks has 8 tabs.

12. An apparatus of claim 11, wherein the plurality of filtereing masks are modified 3×3 mask forms including:

a filtering mask in which eight tabs are selected from the 3×3 mask form, discarding one corner tab;

a filtering mask in which more tabs are selected in a vertical direction than a horizontal direction, and more vertical lower tabs are selected than vertical upper tabs;

a filtering mask in which more tabs are selected in the vertical direction than the horizontal direction, and more vertical upper tabs are selected than vertical lower tabs;

a filtering mask in which more tabs are selected in the horizontal direction than the vertical direction, and more horizontal left tabs are selected than horizontal right tabs; and

a filtering mask in which more tabs are selected in the horizontal direction than the vertical direction, and more horizontal right tabs are selected than horizontal left tabs.

13. An apparatus of claim 10, wherein a filtering mask with more tabs toward a block adjacent said block B is selected.

14. An apparatus of claim 10, further comprising:

a comparison unit selecting a pixel within the selected mask P_m as a candidate pixel if the value of the pixel P_m meets a predetermined condition.

15. An apparatus of claim 14, wherein the pixel P_m is selected as a candidate pixel if the absolute value of the difference between said pixel P and pixel P_m is less than a threshold value.

16. An apparatus of claim 15, wherein the threshold value is calculated by an equation below,

δ=k×q

17. An apparatus of claim 16, wherein the value of k is 1.0 for filtering boundary pixels of said block B and 0.6 for filtering pixels within the boundary pixels of said block.

18. An apparatus of claim 14, wherein the averaging unit adds a weight value to said pixel P prior to the averaging, wherein said weight value is based upon a number of pixels P_m not selected as candiate pixels.

19. An apparatus of claim 18, wherein said weight value is the number of pixels P_m not selected multiplied by the value of said pixel P.

20. A coding and decoding method comprising:

a discrete cosine transform (DCT) unit performing a DCT operation with respect to divided blocks of an image to generate DCT coefficients;

a quantization unit quantizing the DCT coefficients and transmitting the DCT coefficients in a form of a bit stream through a transmission channel;

a dequantization unit dequantizing the DCT coefficients received through the transmission channel;

an inverse DCT unit performing an inverse DCT operation with respect to the dequantized DCT coefficients to form a reconstructed image, and

a filtering process unit filtering each pixel of each block of the reconstructed image by selecting one of a plurality of filtering masks based upon a position of said pixel in said block; and averaging said pixel and candidate pixels within the selected filtering mask.

21. A method of filtering an image, comprising:

filtering, with a filtering unit, a pixel of an image according to a filtering methodology, the filtering methodology including,

selecting at least four pixels of successive pixels according to a position of a pixel to be filtered, the selected at least four pixels including the pixel to be filtered,

determining a difference value between two of the at least four successive pixels, the two of the selected at least four pixels including non-adjacent pixels,

comparing the difference value and a threshold, the threshold being based on quantization information of at least a portion of the image including the pixel to be filtered, and

applying a mathematical operation to the at least two of the selected at least four pixels based on the comparing step.

22. The method of claim 21, wherein the difference value is an absolute value of the difference between the pixel to be filtered and the another of the selected at least four pixels.

23. The method of claim 21, wherein the filtering methodology filters the pixel based on a result of the comparison.

24. The method of claim 23, wherein the filtering methodology filters the pixel using the pixel to be filtered and another of the selected at least four pixels based on a result of the comparison.

25. The method of claim 24, wherein if the determined value is less than the threshold, the filtering methodology filters the pixel using the pixel to be filtered and another of the selected at least four pixels.

26. The method of claim 25, wherein the difference value is an absolute value.

27. The method of claim 25, wherein the determined value is the difference value.

28. The method of claim 23, wherein if the determined value is less than the threshold, the filtering methodology filters the pixel based on the difference value.

29. The method of claim 21, wherein the determined value is the difference value.

30. The method of claim 21, wherein the determined value is an absolute value.

31. The method of claim 21, wherein the difference value is an absolute value.

32. The method of claim 31, wherein the difference is based on the pixel to be filtered and a nearby pixel.

33. The method of claim 21, wherein one of the two successive pixels upon which the difference is based is the pixel to be filtered.

34. The method of claim 21, wherein the quantization information includes a quantization parameter.