Method and apparatus for filtering chrominance signals of image

• US 6,577,352 B1
• Filed: 06/18/1999
• Issued: 06/10/2003
• Est. Priority Date: 06/30/1998
• Status: Expired due to Term

• Current Assignee: Pantech Incorporated
• Original Assignee: SK Hynix Inc.
• Inventors: Kim, Hae Kwang, Moon, Joo Hee, Park, Cheol Soo
• Primary Examiner(s): Miller, John
• Assistant Examiner(s): YENKE, BRIAN P
• Application Number:
  US09/335,636
• Time in Patent Office: 1,453 Days
• Field of Search: H04/N.523, 348/624, 348/630, 348/644, 348/645, 348/646, 348/647, 348/649, 348/713, 348/587-591, 348/592, 348/617, 348/621, 348/631, 348/627, 348/576, 348/577, 382/162, 382/164, 382/165, 382/167, 382/260, 382/261, 345/607, 358/464, 358/453, 358/462, 358/512, 358/516, 358/526, 358/538
• US Class Current: 348/624
• CPC Class Codes: G06T 9/20 Contour coding, e.g. using detection of edges
• CPC Class Codes: H04N 19/117 Filters, e.g. for pre-processing or post-processing sub-band filter banks H04N19/635
• CPC Class Codes: H04N 19/17 the unit being an image region, e.g. an object
• CPC Class Codes: H04N 19/21 with binary alpha-plane coding for video objects, e.g. context-based arithmetic encoding [CAE]

Abstract

An image chrominance signal filtering method and apparatus in which object chrominance information and background chrominance information are extracted in an image format conversion process, based on shape information of luminance signals, and then decimation-filtered or interpolation-filtered. In accordance with the method and apparatus of the present invention, it is possible to avoid a color bleeding phenomenon occurring at the boundaries of objects in an image. Accordingly, there is an advantage in that a degradation in picture quality is prevented.

Claims

1. A method for filtering chrominance signals of an image using a decimation filter, comprising the steps of:

(A) if the number of source object chrominance samples existing in taps of said decimation filter is a positive integer less than the number of said filter taps, then determining whether a pixel, to be filtered, of an input image is an object chrominance pixel or a background chrominance pixel, based on shape information of luminance signals of said image;

(B) if it is determined at said step (A) that said pixel is a background chrominance pixel, then marking object chrominance samples existing in said filter taps, and reproducing background chrominance samples for said marked object chrominance samples, respectively;

(C) if it is determined at said step (A) that said pixel is an object chrominance pixel, then marking background chrominance samples existing in said filter taps, and reproducing object chrominance samples for said marked background chrominance samples, respectively; and

(D) computing pixel values of all chrominance samples including said reproduced samples obtained after the completion of said step (B) or said step (C) using respective filter coefficients of said filter taps.

2. The method as claimed in claim 1, wherein:

said step (B) comprises the step of reproducing, as said background chrominance samples to be filtered, source background chrominance samples spatially nearest all source object chrominance samples existing in said filter taps; and

said step (C) comprises the step of reproducing, as said object chrominance samples to be filtered, source object chrominance samples spatially nearest all source background chrominance samples existing in said filter taps.

3. The method as claimed in claim 2, wherein:

said step (B) comprises the step of if there are a plurality of source background chrominance samples spaced by the same distance from each of said source object chrominance samples, then reproducing one of said source background chrominance samples, which is spatially nearest a decimated object chrominance sample position, as each of said background chrominance samples to be filtered; and

said step (C) comprises the step of if there are a plurality of source object chrominance samples spaced by the same distance from each of said source background chrominance samples, then reproducing one of said source object chrominance samples, which is spatially nearest a decimated background chrominance sample position, as each of said object chrominance samples to be filtered.

4. The method as claimed in claim 1, wherein

said step (B) comprises the step of reproducing, as said background chrominance samples to be filtered, an average value of two source background chrominance samples spatially nearest all source object chrominance samples existing in said filter taps;

and

said step (C) comprises the step of reproducing, as said object chrominance samples to be filtered, an average value of two source object chrominance samples spatially nearest all source background chrominance samples existing in said filter taps.

5. The method as claimed in claim 4, wherein

said step (B) comprises the step of if there is only one source background chrominance sample, which is spatially nearest all source object chrominance samples existing in said filter taps, then reproducing said source background chrominance sample as each of said background chrominance samples to be filtered;

and

said step (C) comprises the step of if there is only one source object chrominance sample, which is spatially nearest all source background chrominance samples existing in said filter taps, then reproducing said source object chrominance sample as each of said object chrominance samples to be filtered.

6. The method as claimed in claim 1, wherein:

said step (B) comprises the step of reproducing an average value of all source background chrominance samples existing in said filter taps for each of all source object chrominance samples existing in said filter taps; and

said step (C) comprises the step of reproducing an average value of all source object chrominance samples existing in said filter taps for each of all source background chrominance samples existing in said filter taps.

7. A method for filtering chrominance signals of an image using a decimation filter, comprising the steps of:

(A) determining whether or not there are, in taps of said filter, chrominance samples associated with an object region different from that of chrominance samples to be filtered;

(B) if there are chrominance samples, to be filtered, associated with a different object region from that of other chrominance samples to be filtered, then marking said chrominance samples associated with said different object region, and reproducing, for said marked chrominance samples, a chrominance sample associated with the same object region as that of said other chrominance samples to be filtered; and

(C) said step comprising the step of reproducing one of said other chrominance samples, which is spatially nearest all source chrominance samples existing in said filter taps while being associated with said different object region, as a chrominance sample to be filtered in replace of each of said marked chrominance samples.

8. The method as claimed in claim 7 wherein said step (B) comprises the step of if said other chrominance samples include a plurality of chrominance samples spaced by the same distance from each of said source chrominance samples associated with said different object region, then reproducing a source chrominance sample, which is spatially nearest a position of said source chrominance sample while being associated with said object region of said other chrominance samples, as a chrominance sample to be filtered in replace of each of said marked chrominance samples.

9. The method as claimed in claim 7, wherein said step (B) comprises the step of reproducing an average value of two source chrominance samples, which are associated with said object region of said other chrominance samples and spatially nearest all source chrominance samples existing in said filter taps while being associated with said different object region, as a chrominance sample to be filtered in replace of each of said marked chrominance samples.

10. The method as claimed in claim 9, wherein said step (B) comprises the step of if there is only one source chrominance sample, which is spatially nearest a position of each of said source chrominance samples while being associated with said object region of said other chrominance samples, then reproducing said source chrominance sample as a chrominance sample to be filtered in replace of each of said marked chrominance samples.

11. The method as claimed in claim 7, wherein said step (B) comprises the step of reproducing an average value of all source chrominance samples associated with said different object region, as a chrominance sample to be filtered in replace of all source chrominance samples existing in said filter taps while being associated with said different object region.

12. An apparatus for filtering chrominance signals of an image comprising a decimation filter adapted to compute pixel values using respective filter coefficients of filter taps thereof, said apparatus further comprising:

a background chrominance pixel reproducer for determining whether a pixel, to be filtered in accordance with a filtering operation, of an input image is an object chrominance pixel or a background chrominance pixel, based on chrominance shape information of said image, said background chrominance pixel reproducer serving to mark background chrominance pixel values existing in said filter taps, when it is determined that said pixel is an object chrominance pixel, and reproducing object chrominance pixel values for said marked background chrominance pixel values, respectively; and

an object chrominance pixel reproducer for marking object chrominance pixel values existing in said filter taps, when said pixel is a background chrominance pixel, and reproducing background chrominance pixel values for said marked object chrominance pixel values, respectively.

13. The apparatus as claimed in claim 12, wherein:

said object chrominance pixel reproducer comprises an object chrominance pixel marker for marking said object chrominance pixel values existing in said filter taps, based on said chrominance shape information, so as to allow a reproduction for said object chrominance pixel values, and a marked-pixel value reproducer for reproducing a background chrominance pixel value for each of said pixel values marked by said object chrominance pixel marker; and

said background chrominance pixel reproducer comprises a background chrominance pixel marker for marking said background chrominance pixel values existing in said filter taps, based on said chrominance shape information, so as to allow a reproduction for said background chrominance pixel values, and a marked-pixel value reproducer for reproducing an object chrominance pixel value for each of said pixel values marked by said background chrominance pixel marker.

14. The apparatus as claimed in claim 13, wherein said marked-pixel value reproducer comprises means for reproducing the pixel value of a chrominance signal, which is not marked and spatially nearest each of said marked pixel values, for said marked pixel value.

15. The apparatus as claimed in claim 13, wherein said marked-pixel value reproducer comprises means for reproducing an average value of two chrominance signals, which are not marked and spatially nearest each of said marked pixel values, for said marked pixel value.

16. The apparatus as claimed in claim 13, wherein said marked-pixel value reproducer comprises means for reproducing an average value of all chrominance signals, which are not marked while existing in said filter taps, for said marked pixel value.

17. A method for filtering chrominance signals of an image using an interpolation filter, comprising the steps of:

(A) if the number of decimated object chrominance samples existing in taps of said interpolation filter is a positive integer less than the number of said filter taps, then determining whether a pixel, to be filtered, of an input image is a decimated object chrominance pixel or a decimated background chrominance pixel, based on shape information of luminance signals of said image;

(B) if it is determined at said step (A) that said pixel is a decimated background chrominance pixel, then marking decimated object chrominance samples existing in said filter taps, and reproducing decimated background chrominance samples for said marked decimated object chrominance samples, respectively;

(C) if it is determined at said step (A) that said pixel is an decimated object chrominance pixel, then marking decimated background chrominance samples existing in said filter taps, and reproducing decimated object chrominance samples for said marked decimated background chrominance samples, respectively; and

(D) computing pixel values of all chrominance samples including said reproduced samples obtained after the completion of said step (B) or said step (C) using respective filter coefficients of said filter taps.

18. The method as claimed in claim 17, wherein:

said step (B) comprises the step of if there are a plurality of decimated background chrominance samples spaced by the same distance from each of said decimated object chrominance samples, then reproducing one of said decimated background chrominance samples, which is spatially nearest a decimated object chrominance sample position, as each of said decimated background chrominance samples to be filtered; and

said step (C) comprises the step of if there are a plurality of decimated object chrominance samples spaced by the same distance from each of said decimated background chrominance samples, then reproducing one of said decimated object chrominance samples, which is spatially nearest a decimated background chrominance sample position, as each of said decimated object chrominance samples to be filtered.

19. The method as claimed in claim 17, wherein

said step (B) comprises the step of reproducing, as said decimated background chrominance samples to be filtered, an average value of two decimated background chrominance samples spatially nearest all decimated object chrominance samples existing in said filter taps;

and

said step (C) comprises the step of reproducing, as said decimated object chrominance samples to be filtered, an average value of two decimated object chrominance samples spatially nearest all decimated background chrominance samples existing in said filter taps.

20. The method as claimed in claim 19, wherein

said step (B) comprises the step of if there is only one decimated background chrominance sample, which is spatially nearest all decimated object chrominance samples existing in said filter taps, then reproducing said decimated background chrominance sample as each of said decimated background chrominance samples to be filtered;

and

said step (C) comprises the step of if there is only one decimated object chrominance sample, which is spatially nearest all decimated background chrominance samples existing in said filter taps, then reproducing said decimated object chrominance sample as each of said decimated object chrominance samples to be filtered.

21. The method as claimed in claim 17, wherein:

said step (B) comprises the step of reproducing an average value of all decimated background chrominance samples existing in said filter taps for each of all decimated object chrominance samples existing in said filter taps; and

said step (C) comprises the step of reproducing an average value of all decimated object chrominance samples existing in said filter taps for each of all decimated background chrominance samples existing in said filter taps.

22. A method for filtering chrominance signals of an image using an interpolation filter, comprising the steps of:

(A) determining whether or not there are, in taps of said filter, decimated chrominance samples associated with an object region different from that of decimated chrominance samples to be filtered; and

(B) if there are decimated chrominance samples, to be filtered, associated with a different object region from that of other decimated chrominance samples to be filtered, then marking said decimated chrominance samples associated with said different object region, and reproducing, for said marked decimated chrominance samples, a decimated chrominance sample associated with the same object region as that of said other decimated chrominance samples to be filtered; and

(C) computing pixel values of all decimated chrominance samples including said reproduced samples obtained after the completion of said step (B) using respective filter coefficients of said filter taps, thereby filtering said chrominance signals.

23. An apparatus for filtering chrominance signals of an image comprising an interpolation filter adapted to compute pixel values using respective filter coefficients of filter taps thereof, said apparatus further comprising:

a background chrominance pixel reproducer for determining whether a decimated chrominance pixel, to be filtered, of an input image is an object chrominance pixel or a background chrominance pixel, based on shape information of said image, said background chrominance pixel reproducer serving to mark background chrominance pixel values existing in said interpolation filter taps, when it is determined that said pixel is an object chrominance pixel, and reproducing object chrominance pixel values for said marked background chrominance pixel values, respectively; and

an object chrominance pixel reproducer for marking object chrominance pixel values existing in said interpolation filter taps, when said pixel is a background chrominance pixel, and reproducing background chrominance pixel values for said marked object chrominance pixel values, respectively.

24. The apparatus as claimed in claim 23, wherein:

said object chrominance pixel reproducer comprises an object chrominance pixel marker for marking said object chrominance pixel values existing in said interpolation filter taps, based on said chrominance shape information, so as to allow a reproduction for said object chrominance pixel values, and a marked-pixel value reproducer for reproducing a background chrominance pixel value for each of said pixel values marked by said object chrominance pixel marker; and

said background chrominance pixel reproducer comprises a background chrominance pixel marker for marking said background chrominance pixel values existing in said interpolation filter taps, based on said chrominance shape information, so as to allow a reproduction for said background chrominance pixel values, and a marked-pixel value reproducer for reproducing an object chrominance pixel value for each of said pixel values marked by said background chrominance pixel marker.