Method and apparatus for removing a checkerboard-like noise artifact from a captured composite NTSC video frame

• US 6,768,518 B1
• Filed: 12/16/1998
• Issued: 07/27/2004
• Est. Priority Date: 12/16/1998
• Status: Expired due to Fees

• Current Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Bozdagi, Gozde
• Primary Examiner(s): Lee, Michael H.
• Assistant Examiner(s): DESIR, JEAN WICEL
• Application Number:
  US09/213,649
• Time in Patent Office: 2,050 Days
• Field of Search: 348/615, 348/607, 348/608, 348/627, 348/398.1, 348/397.1, 348/437.1, 348/438.1, 348/400.1, 348/403.1, 348/408.1, 382/275, 382/276, 382/277, 382/281, 358/447, 375/240.19, 375/240.18
• US Class Current: 348/615
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/20064 Wavelet transform [DWT]
• CPC Class Codes: G06T 2207/20192 Edge enhancement; Edge preservation
• CPC Class Codes: G06T 5/10 using non-spatial domain filtering
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: H04N 5/21 Circuitry for suppressing or minimising disturbance, e.g. moiré or halo
• CPC Class Codes: H04N 5/4448 for frame-grabbing

Abstract

A wavelet domain filtering technique is disclosed that removes or at least minimizes checkerboard-like noise artifacts that result from capturing a composite NTSC video signal. A Haar basis filter bank is used to obtain a two-level wavelet decomposition. Low-pass filtering is performed on the high-high (HH) and the high-low (HL) bands of the wavelet decomposition since the checkerboard-like noise artifact is primarily visible on the vertical and the diagonal edges of the captured composite NTSC video frame.

Claims

1. A method of processing a video frame captured from a moving video signal, comprising the steps of:

transforming the video frame into the wavelet domain to form a transformed video frame;

detecting the transformed video frame for checkerboard noise;

low-pass filtering the transformed video frame where the checkerboard noise is detected; and

reconstructing the video frame from the transformed video frame to form a filtered video frame.

2. The method of claim 1, wherein the moving video signal is a Composite NTSC video signal.

3. The method of claim 1, wherein the transforming step includes the step of:

generating average and difference components of the video frame.

4. The method of claim 1, wherein the detecting includes determining that the checkerboard noise is present when:

5. The method of claim 4, wherein the low-pass filtering assigns an intensity Y₁ to pixel (S, P) in accordance with:

6. The method of claim 1, wherein:

the transforming step includes transforming the video frame into the wavelet domain using Haar basis functions; and

the reconstructing step includes reconstructing the video frame from the transformed video frame using inverse Haar basis functions.

7. The method of claim 1, wherein:

the video frame includes a first interlaced field and a second interlaced field,

the transforming step includes the step of transforming the first and second interlaced fields into the wavelet domain to form first and second transformed fields, respectively,

the filtering step includes the step of filtering out noise artifacts occurring in the first and second transformed fields, and

the reconstructing step includes the step of reconstructing the video frame from the first and second transformed fields to form the filtered video frame.

8. The method of claim 1, further including the step of:

downstream processing the filtered video frame.

9. The method of claim 8, wherein the downstream processing step includes at least one of the steps of:

printing the filtered video frame on a printer,

displaying the filtered video frame on a video monitor, and

inserting the filtered video frame into an electronic document.

10. The method of claim 1, wherein said low-pass filtering is not applied to the transformed video frame where the checkerboard noise is not detected.

11. A method of processing a video frame captured from a moving video signal, comprising:

transforming the video frame into the wavelet domain to form a transformed video frame, said transforming including executing a two-level wavelet decomposition to generate Low-Low, Low-High, High-Low, and High-High bands of the transformed video frame;

filtering out noise artifacts present in the transformed video frame, said filtering being carried out in less than all the bands; and

,

reconstructing the video frame from the transformed video frame to form a filtered video frame.

12. The method of claim 11, wherein the filtering is carried out in only the High-Low and High-High bands.

13. An image processing apparatus comprising:

a frame grabber that generates a video frame from a moving video signal;

a first processor that transforms the video frame into the wavelet domain to form a decomposed video frame;

a filter that removes noise artifacts present in the decomposed video frame, said filter detecting checkerboard-like noise artifacts occurring in the decomposed video frame and low-pass filtering decomposed video frame pixel values associated with the checkerboard-like noise artifact; and

a second processor that reconstructs the video frame from the decomposed video frame to form a filtered video frame.

14. The apparatus of claim 13, wherein the moving video signal is a Composite NTSC video signal.

15. The apparatus of claim 13, wherein the first processor includes means for generating average and difference components of the video frame.

16. The apparatus of claim 13, wherein the filter determines that the checkerboard-like noise artifact is present when:

17. The image processing apparatus of claim 13, wherein:

the first processor transforms the video frame into the wavelet domain utilizing Haar basis functions; and

the second processor reconstructs the video frame from the decomposed video frame utilizing inverse Haar basis functions.

18. The apparatus of claim 16, wherein the low-pass filtering conducted by the filter assigns an intensity Y₁ to pixel (S, P) in accordance with:

19. The apparatus of claim 13, wherein the video frame includes a first interlaced field and a second interlaced field, the first processor transforms the first and second interlaced fields into first and second decomposed fields, respectively, the filter removes noise artifacts occurring in each of the first and second decomposed fields, and the second processor reconstructs the video frame from the first and second decomposed fields.

20. The apparatus of claim 13, further including at least one of a printer for printing the filtered video frame, a display monitor for displaying the filtered video frame, and means for inserting the filtered video frame into an electronic document.

21. The image processing apparatus of claim 13, wherein the filter does not perform filtering on decomposed video frame pixel values not associated with the checkerboard-like noise artifact.

22. An image processing apparatus comprising:

a frame grabber that generates a video frame from a moving video signal;

a first processor that transforms the video frame into the wavelet domain to form a decomposed video frame, said first processor including means for generating Low-Low, Low-High, High-Low, and High-high bands of the decomposed video frame;

a filter that removes noise artifacts present in the decomposed video frame, said filter acting on less than all the bands; and

,

a second processor that reconstructs the video frame from the decomposed video frame to form a filtered video frame.

23. The apparatus of claim 22, wherein the filter acts on only the High-Low and High-High bands of the decomposed video frame.