Block boundary artifact reduction for block-based image compression

• US 6,529,638 B1
• Filed: 02/01/1999
• Issued: 03/04/2003
• Est. Priority Date: 02/01/1999
• Status: Expired due to Term

• Current Assignee: Rakuten, Inc. (Rakuten Group, Inc.)
• Original Assignee: Sharp Laboratories of America Incorporated (Hon Hai Precision Industry Co., Ltd.)
• Inventors: Westerman, Larry Alan
• Primary Examiner(s): LEE, TOMMY D
• Application Number:
  US09/241,829
• Time in Patent Office: 1,492 Days
• Field of Search: 382/254, 382/268, 382/275, 382/232, 382/233, 382/294, 358/426, 358/432, 358/433
• US Class Current: 382/275
• CPC Class Codes: H04N 19/60 using transform coding
• CPC Class Codes: H04N 19/86 involving reduction of coding artifacts, e.g. of blockiness

Abstract

Block boundary artifact reduction in decompressed digital images is accomplished by filtering the intensity of pixels in the vicinity of the block boundary. The filter utilizes filter coefficients selected from tables on the basis of the distribution of a scalar quantity describing the pixels neighboring the pixel to which the intensity adjustment is to be applied. The intensities of pixels at the boundary and one pixel removed from the boundary are adjusted. If interpolation pixels in each of neighboring blocks are of relatively constant intensity, the intensities of additional pixels, more remote from the boundary, are adjusted. Filter coefficients can be selected from different arrays for pixels on the boundary or removed from the boundary or if the pixel intensity adjustment is based on the intensities of pixels in a horizontal row or vertical column of interpolation pixels.

Claims

1. A method of reducing an artifact at a block boundary in a decompressed digital image comprising the steps of:

(a) identifying a plurality of interpolation pixels including at least one pixel on each of a first side and a second side of said block boundary;

(b) identifying one said interpolation pixel on said first side of said block boundary as a subject pixel;

(c) computing the mean of an intensity said interpolation pixels;

(d) computing an intensity difference between said mean intensity and said intensity of each of said interpolation pixels; and

(e) quantizing said intensity differences;

(f) constructing an intensity distribution index from said quantized intensity differentials;

(d) selecting a plurality of filter coefficients correlated to said intensity distribution index;

(e) modifying said intensity of said subject pixel by applying a filter producing said modified intensity as a function of said filter coefficients.

2. The method of claim 1 wherein said plurality of interpolation pixels comprise:

(a) said subject pixel;

(b) a first pixel adjoining a side of said subject pixel remote from said block boundary;

(c) a second pixel on said second side of said block boundary and adjoining a side of said subject pixel nearest said block boundary; and

(d) a third pixel adjoining a side of said second pixel remote from said subject pixel.

3. The method of claim 2 further comprising the steps of:

(a) selecting a second plurality of filter coefficients correlated to a distribution of said scalar characteristic of said first pixel, said subject pixel, a second pixel on said second side of said block boundary and adjoining said subject pixel, and a third pixel adjoining a side of said first pixel remote from said block boundary; and

(b) modifying an intensity of said first pixel by applying a filter producing said modified intensity of said first pixel as a function of said second plurality of filter coefficients.

4. The method of claim 1 wherein said plurality of interpolation pixels comprise:

(a) said subject pixel;

(b) a first pixel adjoining a side of said subject pixel remote from said block boundary;

(c) a second pixel adjoining a side of said subject pixel nearest said block boundary; and

(d) a third pixel on said second side of said block boundary adjoining said second pixel.

5. A method of reducing an artifact at a block boundary in a decompressed digital image comprising the steps of:

(a) identifying a plurality of interpolation pixels including at least one pixel on each of a first side and a second side of said block boundary;

(b) identifying one said interpolation pixel on said first side of said block boundary as a subject pixel;

(c) computing a statistical measure of an intensity said interpolation pixels;

(d) computing an intensity difference between said statistical measure intensity and said intensity of each of said interpolation pixels; and

(e) quantizing said intensity differences;

(f) constructing an intensity distribution index from said quantized intensity differentials;

(g) selecting a plurality of filter coefficients correlated to said intensity distribution index;

(h) modifying said intensity of said subject pixel by applying a filter producing said modified intensity as a function of said filter coefficients.

6. The method of claim 5 wherein said plurality of interpolation pixels comprise:

(a) said subject pixel;

(b) a first pixel adjoining a side of said subject pixel remote from said block boundary;

(c) a second pixel on said second side of said block boundary and adjoining a side of said subject pixel nearest said block boundary; and

(d) a third pixel adjoining a side of said second pixel remote from said subject pixel.

7. The method of claim 6 further comprising the steps of:

(a) selecting a second plurality of filter coefficients correlated to a distribution of said scalar characteristic of said first pixel, said subject pixel, a second pixel on said second side of said block boundary and adjoining said subject pixel, and a third pixel adjoining a side of said first pixel remote from said block boundary; and

(b) modifying an intensity cf said first pixel by applying a filter producing said modified intensity of said first pixel as a function of said second plurality of filter coefficients.

8. The method of claim 5 wherein said plurality of interpolation pixels comprise:

(a) said subject pixel;

(b) a first pixel adjoining a side of said subject pixel remote from said block boundary;

(c) a second pixel adjoining a side of said subject pixel nearest said block boundary; and

(d) a third pixel on said second side of said block boundary adjoining said second pixel.