Method and apparatus for reducing coding artifacts of block-based image encoding and object-based image encoding

• US 6,115,503 A
• Filed: 07/03/1997
• Issued: 09/05/2000
• Est. Priority Date: 07/04/1996
• Status: Expired due to Term

• Current Assignee: Siemens AG
• Original Assignee: Siemens AG
• Inventors: Kaup, Andre
• Primary Examiner(s): Tran, Phuoc
• Assistant Examiner(s): Mariam, Daniel G.
• Application Number:
  US08/887,577
• Time in Patent Office: 1,160 Days
• Field of Search: 382/166, 382/262, 382/268, 382/275, 382/254, 348/606, 348/607, 348/610, 358/433
• US Class Current: 382/268
• CPC Class Codes: H04N 19/20 using video object coding
• CPC Class Codes: H04N 19/60 using transform coding
• CPC Class Codes: H04N 19/70 characterised by syntax aspects related to video coding, e.g. related to compression standards
• CPC Class Codes: H04N 19/80 Details of filtering operations specially adapted for video compression, e.g. for pixel interpolation H04N19/635, H04N19/86 take precedence
• CPC Class Codes: H04N 19/86 involving reduction of coding artifacts, e.g. of blockiness

Abstract

A method and apparatus for enhancing decoded images which have been encoded checks the neighboring picture elements of a picture element which are located in the region of the picture element to see whether the brightness values and/or the color values of the neighboring picture elements is similar enough to brightness values and/or color values of the picture element in question according to a similarity criterion such as a threshold. When the similarity is within the threshold, then the corresponding neighboring picture element is taken into consideration in determining a new brightness value for the picture element in question. Neighboring picture elements are taken into consideration when the neighboring picture element and the picture element are located in different image blocks of the image which has been decoded from a block-based encoding process. Neighboring picture elements are not taken into consideration when the neighboring picture element and the picture element are located in different objects of the image which has been decoded from an object-based image encoding process.

Claims

1. A method for reducing coding artifacts of block-based image encoding methods, using a computer, comprising the steps of:

implementing the following steps for picture elements of at least a part of a digitized, decoded image;

checking for neighboring picture elements in said digitized decoded image at a same processing level without intervening filtering that are located in an environment of a picture element under investigation to see whether an amount of a difference between at least one of a brightness value and a color value of the picture element under investigation and a respective one of a brightness value and a color value of a neighboring picture element is below a predetermined threshold;

taking the at least one of the brightness value and the color value of the neighboring picture element into consideration in determining at least one of a new brightness value and of a new color value of the picture element when one of the following conditions is met;

an amount of the difference is lower than the predetermined threshold, and

the picture element under investigation and the neighboring picture element are located in different image blocks of the image; and

deriving the at least one of the new brightness value and the new color value of the picture element by weighted averaging of at least the at least one of the brightness values and color values of the neighboring picture elements.

4. A method according to claim 1, further comprising the step of:

taking into consideration at least one of the brightness value and the color value of the picture element under investigation in determining the at least one of the new brightness value and the new color value of the picture element under investigation.

6. A method according to claims 1, wherein only picture elements directly adjoining the picture element under investigation are employed as the neighboring picture elements.

8. A method according to claim 1, wherein said method is implemented row-by-row for the picture elements under investigation within the image.

9. A method as claimed in claim 1, wherein said method is implemented column-by-column for the picture elements under investigation within the image.

12. A method according to claim 1, further comprising the steps of:

performing a first iteration of the method in a sequence for the picture elements under investigation along a first direction; and

performing a second iteration of the method in a sequence for the picture elements under investigation along a second direction.

15. A method according to claim 12, wherein only picture elements that are located in an environment of the picture element under investigation along a first direction are employed as the neighboring picture elements in the first iteration;

and

only picture elements that are located in an environment of the picture element under investigation along a second direction are employed as neighboring picture elements in the second iteration.

16. A method according to claim 1, further comprising the steps of:

repeatedly applying the method to an image;

determining a quality criterion with respect to the image after each iteration; and

implementing a new iteration for the image when the determined quality criterion lie below a predetermined quality threshold.

18. A method according to claim 1, further comprising the step of:

applying the method for image post-processing.

20. A method according to claim 1, further comprising the step of:

predicting the image in the framework of image encoding and/or image decoding.

2. A method for the reducing coding artifacts of object-based image encoding methods, using a computer, comprising the steps of:

implementing the following steps for picture elements of at least a part of a digitized, decoded image;

carrying out a check for neighboring picture elements that are located in an environment of a picture element under investigation to see whether an amount of a difference between at least one of a brightness value and a color value of the picture element under investigation and a respective one of brightness value and a color value of a neighboring picture element is below a predetermined threshold;

taking the at least one of the brightness value and the color value of the neighboring picture element into consideration in determining at least one of a new brightness value and a new color value of the picture element when both of the following conditions are met;

the amount of the difference is lower than the predetermined threshold, and

the picture element under investigation and the neighboring picture element are not located in different image objects of the image; and

deriving the at least one of the new brightness value and the new color value of the picture element by weighted averaging of at least the respective brightness values and color values of the neighboring picture elements.

3. A method according to claim 2, further comprising the steps of:

when a block-based image encoding method is additionally employed in the object-based image encoding method and an image object includes image blocks, checking to see whether the picture element under investigation and the neighboring picture element which are both located in a same image object are located in different image blocks of the image object; and

taking the at least one of the brightness value and the color value of a neighboring picture element into consideration in determining the at least one of the new brightness value and the new color value of the picture element when the picture element and the neighboring picture element are located in different image blocks of the image object.

5. A method according to claim 2, further comprising the step of:

taking into consideration at least one of the brightness value and the color value of the picture element under investigation in determining the at least one of the new brightness value and the new color value of the picture element under investigation.

7. A method according to claims 2, wherein only picture elements directly adjoining the picture element under investigation are employed as the neighboring picture elements.

10. A method according to claim 2, wherein said method is implemented row-by-row for the picture elements under investigation within the image.

11. A method as claimed in claim 2, wherein said method is implemented column-by-column for the picture elements under investigation within the image.

13. A method according to claim 2, further comprising the steps of:

performing a first iteration of the method in a sequence for the picture elements under investigation along a first direction; and

performing a second iteration of the method in a sequence for the picture elements under investigation along a second direction.

14. A method according to claim 13, wherein only picture elements that are located in an environment of the picture element under investigation along a first direction are employed as the neighboring picture elements in the first iteration;

and

only picture elements that are located in an environment of the picture element under investigation along a second direction are employed as neighboring picture elements in the second iteration.

17. A method according to claim 2, further comprising the steps of:

repeatedly applying the method to an image;

determining a quality criterion with respect to the image after each iteration; and

implementing a new iteration for the image when the determined quality criterion lie below a predetermined quality threshold.

19. A method according to claim 2, further comprising the step of:

applying the method for image post-processing.

21. An image encoding unit for encoding digitized images, comprising:

a first means for transformation coding of a digital image including providing transformation coefficients;

a second means for quantizing transformation coefficients that were formed by the first means in the transformation encoding;

a third means for the inverse quantization of the quantized transformation coefficients;

a fourth means for the inverse transformation coding of the quantized transformation coefficients;

an image storage for storing image data;

a fifth means for predicting image data of two chronologically successive images;

a subtraction unit to which, first, image data of an image to be encoded and, second image data of a chronologically preceding image are supplied, so that only difference image data of two successive images are supplied to the first means and encoded; and

a sixth means for eliminating coding artifacts from the digital image, said sixth means considering a brightness value and/or color value of an adjacent image point in said encoded digitized image at a same process level without intervening filtering upon acquisition of a new brightness value or new color value of an image point of the digital image if

an amount of a difference between the brightness value and/or color value of the image point and a brightness value and/or color value of the adjacent image point is smaller than a limit, or

the image point and the adjacent image point are situated in different image blocks and/or different image objects of the digital image.

22. An image encoding unit according to claim 21, further comprising:

a seventh means for segmenting the image into at least one of image blocks and image objects.

23. An image decoding unit for decoding digitized images, comprising:

a means for the inverse quantizating of quantized transformation coefficients;

a means for the inverse transformation coding;

an image storage for storing image data;

a means for predicting image data of two chronologically successive images;

an adder unit to which, first, the quantized transformation coefficients of an image to be decoded and, second, image data of a chronologically preceding, decoded image are supplied, so that predicted image data of two successive images are supplied to the inverse quantizating means and decoded; and

a means for eliminating coding artifacts in the digital image, said means for eliminating coding artifacts considering a brightness value and/or color value of an adjacent image point in said digitized decoded image at a same processing level without intervening filtering upon acquisition of a new brightness value or new color value of an image point of the digital image if;

an amount of a difference between the brightness value and/or color value of the image point and a brightness value and/or color value of the adjacent image point is smaller than a limit, or

the image point and the adjacent image point are situated in different image blocks and/or different image objects of the digital image.