Method and apparatus for texture analysis and replicability determination

• US 6,594,391 B1
• Filed: 09/03/1999
• Issued: 07/15/2003
• Est. Priority Date: 09/03/1999
• Status: Expired due to Fees

• Current Assignee: Lucent Technologies, Inc. (Nokia Corporation)
• Original Assignee: Lucent Technologies, Inc. (Nokia Corporation)
• Inventors: Quadranti, Laurent Lorenzo Balmelli, Mojsilovic, Aleksandra
• Primary Examiner(s): Dastouri, Mehrdad
• Application Number:
  US09/390,388
• Time in Patent Office: 1,411 Days
• Field of Search: 382/190, 382/195, 382/203, 382/224, 382/225, 382/232, 382/260, 382/263, 382/264, 382/279, 382/296, 382/240, 382/108, 375/240.18, 375/240.19, 375/240.08
• US Class Current: 382/203
• CPC Class Codes: G06T 7/42 using transform domain methods

Abstract

An image or other information signal having a texture is analyzed along perceptual dimensions such as directionality, symmetry, regularity and type of regularity. A wavelet decomposition or other type of filtering operation is first applied to the information signal, in order to generate oriented subbands. Texture-related features, e.g., coarse features, vertical edges, horizontal edges, corners, etc., are then extracted from the subbands. The extracted features are then used to characterize the above-noted perceptual dimensions of the texture. A classification process may subsequently be used to classify the texture as directional, symmetric, regular and/or having a particular regularity type, e.g., a small, medium or large regularity, based on comparison with established thresholds. Advantageously, the classification process can also be used to determine the replicability of the texture.

Claims

1. A processor-implemented method of analyzing an information signal having a texture, the method comprising the steps of:

extracting texture-related features from the information signal; and

generating an output signal representative of at least one of a directionality, a symmetry and a, regularity of the texture, based on the extracted features, wherein the step of extracting texture-related features further includes the step of computing a feature vector as

2. The method of claim 1 further comprising the step of:

determining replicability of the texture utilizing one or more of the directionality, the symmetry, the regularity and regularity type of the texture.

3. The method of claim 1 wherein the information signal comprises at least a portion of at least one image.

4. The method of claim 3 further including the step of separating the image into a plurality of regions, rotating at least a subset of the regions such that each region has a common orientation, and applying a wavelet decomposition to each of the regions.

5. The method of claim 1 wherein the step of extracting texture-related features further includes implementing a wavelet decomposition of the information signal.

6. The method of claim 5 wherein the step of extracting texture-related features further includes the steps of:

extracting vertical edges of the information signal utilizing a first subband of the wavelet decomposition; and

extracting horizontal edges of the information signal utilizing a second subband of the wavelet decomposition.

7. The method of claim 6 wherein the step of extracting texture-related features further includes the step of extracting coarse features of the information signal utilizing a third subband of the wavelet decomposition.

8. The method of claim 6 wherein the step of extracting texture-related features further includes the step of extracting corner edges of the information signal utilizing a fourth subband of the wavelet decomposition.

9. The method of claim 1 wherein the step of determining replicability of the texture further includes the steps of:

determining the directionality of the texture; and

identifying the texture as replicable if the texture has a designated directionality.

10. A processor-implemented method of analyzing an information signal having a texture, the method comprising the steps of:

extracting texture-related features from-the information signal;

generating an output signal representative of at least one of a directionality, a symmetry and a regularity of the texture, based on the extracted features; and

determining replicability of the texture utilizing one or more of the directionality, the symmetry, the regularity and a regularity type of the texture wherein the step of determining replicability of the texture further includes the steps of;

determining the directionality of the texture; and

identifying the texture as replicable if the texture has a designated directionality, wherein if the texture is not identified as replicable based on the directionality, the step of determining replicability of the texture further includes the steps of;

determining the symmetry of the texture; and

identifying the texture as non-replicable if the texture does not have a designated symmetry.

11. The method of claim 10 wherein if the texture is not identified as replicable based on the directionality or as non-replicable based on the symmetry, the step of determining replicability of the texture further includes the steps of:

determining the regularity of the texture; and

identifying the texture as non-replicable if the texture does not have a designated regularity.

12. The method of claim 11 wherein if the texture has a designated regularity, the step of determining replicability of the texture further includes the steps of:

determining a type of the regularity of the texture;

identifying the texture as replicable if the regularity type is small; and

identifying the texture as non-replicable if the regularity type is medium or large.

13. An apparatus for processing an information signal having a texture, the apparatus comprising:

a filter bank for extracting texture-related features from the information signal, wherein the extracted features are used to determine at least one of a directionality, a symmetry and a regularity of the texture, and to determine replicability of the texture utilizing one or more of the directional, the symmetry, the regularity and a regularity type of the texture,

wherein a feature vector is computed as

14. The apparatus of claim 13 wherein the information signal comprises at least a portion of at least one image.

15. The apparatus of claim 13 wherein the filter bank implements a wavelet decomposition of the information signal.

17. The apparatus of claim 15 wherein filter bank further includes a third filter for extracting coarse features of the information signal utilizing a third subband of the wavelet decomposition.

18. The apparatus of claim 15 wherein the filter bank further includes a fourth filter for extracting corner edges of the information signal utilizing a fourth subband of the wavelet decomposition.

16. The apparatus of claim 16 wherein the filter bank further includes:

a first filter for extracting vertical edges of the information signal utilizing a first subband of the wavelet decomposition; and

a second filter for extracting horizontal edges of the information signal utilizing a second subband of the wavelet decomposition.

19. An apparatus for processing an information signal having a texture, the apparatus comprising:

a memory for storing at least a portion of the information signal; and

a processor coupled to the memory and operative to implement a filter bank for extracting texture-related features from the information signal and to determine at least one of a directionality, a symmetry and a regularity of the texture based on the extracted features, wherein the processor is further operative to extract the texture-related features by computing a feature vector of the form

20. The apparatus of claim 19 wherein the filter bank implements a wavelet decomposition of the information signal.

21. The apparatus of claim 19 wherein the processor is operative to determine replicability of the texture utilizing one or more of the directionality, the symmetry, the regularity and a regularity type of the texture.

22. A machine-readable medium storing one or more software programs for processing an information signal having a texture, wherein the one or more software programs when executed by a processor include:

a step for extracting texture-related features from the information signal;

a step for generating an output signal representative of at least one of a directionality, a symmetry and a regularity of the texture, based on the extracted features; and

a step for determining replicability of the texture utilizing one or more of the directionality, the symmetry, the regularity and a regularity type of the texture,

wherein the step of extracting texture-related features further includes the step of computing a feature vector as