SYSTEM AND METHOD FOR FEATURE EMPHASIS AND DE-EMPHASIS IN IMAGE PROCESSING

• US 20090279785A1
• Filed: 05/11/2009
• Published: 11/12/2009
• Est. Priority Date: 05/09/2008
• Status: Active Grant

• Current Assignee: Avago Technologies General IP PTE Limited (Broadcom, Inc.)
• Original Assignee: Broadcom Corporation (Broadcom, Inc.)
• Inventors: Wredenhagen, Gordon F.
• Granted Patent: US 8,417,051 B2
• Time in Patent Office: Days
• US Class Current: 382/190
• CPC Class Codes: G06T 2207/20192 Edge enhancement; Edge preservation
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06T 5/73 Deblurring; Sharpening

Abstract

A system and method for emphasizing and/or de-emphasizing features in a digital image. In an embodiment of the invention, the image is processed by a series of filters, which decompose the image into a series of sub-bands. Each sub-band is then processed by an emphasis circuit. In an embodiment of the invention, the emphasis circuit is connected to an input representing a particular sub-band. Each filter, therefore, is connected to its own respective emphasis circuit. The result is a series of emphasis circuit outputs, which are then combined. The result of the combination is the processed image, containing one or more emphasized or deemphasized features. In an alternative embodiment of the invention, the sub-bands are combined in a weighted fashion. The sum of the weighted sub-bands is then applied to a single emphasis circuit. The output of the emphasis circuit, in this embodiment, is a processed image having one or more emphasized or deemphasized features.

Claims

1. A system for feature emphasis and de-emphasis in processing an image, comprising:

a plurality of sub-band filters that are configured to receive an image, wherein each sub-band filter is configured to output a respective sub-band associated with said image; and

at least one emphasis circuit configured to receive information related to said sub-bands, perform edge orientation, detect a feature in said image, perform emphasis orientation, and output a processed image.

2. The system of claim 1, wherein said at least one emphasis circuit comprises an emphasis circuit for each sub-band filter, such that each sub-band filter is configured to output its sub-band to its respective emphasis circuit.

3. The system of claim 2, further comprising a combiner figured to combine an output from each emphasis circuit, and produce a processed image.

4. The system of claim 1, further comprising a combiner figured to receive said sub-bands, and output a weighted sum of said sub-bands.

5. The system of claim 4, wherein said at least one emphasis circuit comprises a single emphasis circuit configured to receive said weighted sum of said sub-bands.

6. The system of claim 1, wherein each of said at least one emphasis circuit comprises:

an edge orientation module configured to receive said information related to at least one of said sub-bands, and to output edge orientation information;

a feature detector module configured to receive said information related to at least one of said sub-bands and said edge orientation information, and to output feature detector information; and

an emphasis orientation module configured to receive said information related to at least one of said sub-bands and said edge orientation information, and to output emphasis orientation information.

7. The system of claim 6, wherein said feature detector module is configured to:

(a) determine a difference Δ

x_i between pixel data p_i and q_i for each of a plurality of pairs (p_i, q_i), where 0≦

i≦

n and where each p_i and q_i are pixels in said image on opposite sides of a feature;

(b) apply a function to each Δ

x_i; and

(c) calculate a weighted sum of f(Δ

x_i) for a plurality of values f(Δ

x_i).

8. The system of claim 7, wherein said function f is an absolute value function.

9. The system of claim 6, wherein said emphasis circuit comprises a first combiner and wherein said information related to at least one sub-band is combined with said feature detector information and said emphasis orientation information in the first combiner, to produce a first combiner output.

10. The system of claim 9, wherein said emphasis circuit comprises a second combiner and wherein said first combiner output is combined with a first user defined input in the second combiner, to produce a second combiner output.

11. The system of claim 10, wherein said emphasis circuit comprises a third combiner, and wherein a difference between said image and said second combiner output is determined, and combined with a second user defined input in the third combiner to produce a third combiner output.

12. The system of claim 11, wherein said emphasis circuit comprises a limiter configured to receive said third combiner output, to regulate the emphasis and de-emphasis applied to said feature, and to produce a limiter output.

13. The system of claim 12, wherein said emphasis circuit comprises:

a pattern injection circuit configured to apply a pattern to the limiter output and to produce a marked image; and

a multiplexer configured to receive said limiter output, said marked image, and a third user-defined input, and to output said one of said limiter output and marked image, chosen by said third user input as the processed image.

14. A method of emphasizing a feature in an image, comprising:

(a) receiving the image;

(b) decomposing the image into sub-bands;

(c) performing edge orientation, feature detection, and emphasis orientation on information related to the sub-bands to produce a first combiner output;

(d) combining the first combiner output with a first user-defined input to create a second combiner output;

(e) creating a difference between the image and the second combiner output;

(f) combining the difference with a second user-defined input to create a third combiner output;

(g) regulating the emphasis and de-emphasis in the third combiner output, creating a limiter output;

(h) applying a pattern to the limiter output to produce a marked image; and

(i) multiplexing the marked image with the limiter output on the basis of a third user-defined input to create an emphasis circuit output.

15. The method of claim 14, wherein said step (c) comprises:

(i) performing edge orientation on the information related to the sub-bands to produce edge orientation information;

(ii) performing feature detection on the information related to the sub-bands on the basis of the edge orientation information to produce feature detection information;

(iii) performing emphasis orientation on the information related to the sub-bands on the basis of the edge orientation information, to produce emphasis orientation information; and

(iv) combining the information related to the sub-bands, feature detection information, and emphasis orientation information into a first combiner to produce the first combiner output.

16. The method of claim 15, wherein said step (c)(ii) comprises:

A. for each of a plurality of pairs of pixels (p_i,q_i) determining a difference Δ

x_i where p_i and q_i are pixels on opposite sides of a feature;

B. calculating a function f(Δ

x_i) for each x_i; and

C. calculating a weighted sum of a plurality of values f(Δ

x_i) to produce feature detection information.

17. The method of claim 16, wherein the function f is the absolute value function.

18. The method of claim 14, wherein said step h comprises applying the pattern to the feature to thereby highlight the feature.

19. The method of claim 14, wherein the third user-defined input determines whether the limiter output or the marked image is output as the emphasis circuit output.

20. The method of claim 14, wherein the information related to the sub-bands comprises a weighted sum of the sub-bands.

21. The method of claim 20, wherein the emphasis circuit output comprises the processed image.

22. The method of claim 14, wherein said information related to the sub-bands comprises a particular sub-band.

23. The method of claim 22, wherein said steps (c) through (i) are performed for each of a plurality of sub-bands, resulting in a corresponding plurality of emphasis circuit outputs.

24. The method of claim 23, further comprising:

(j) combining the corresponding plurality of emphasis circuit outputs to produce the processed image.