Method and system for searching for images based on color and shape of a selected image

• US 6,606,417 B1
• Filed: 04/19/2002
• Issued: 08/12/2003
• Est. Priority Date: 04/20/1999
• Status: Expired due to Term

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Brechner, Eric L.
• Primary Examiner(s): Ahmed, Samir
• Assistant Examiner(s): BHATNAGAR, ANAND P
• Application Number:
  US10/127,075
• Time in Patent Office: 480 Days
• Field of Search: 382/253, 382/305, 382/306, 382/240, 375/240.19, 707/104.1
• US Class Current: 382/240
• CPC Class Codes: G06F 16/5838 using colour
• CPC Class Codes: G06F 16/5854 using shape and object relationship
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

A method and system that allows a user to find images using a query by example is disclosed. Images similar to the selected image are retrieved based on color or shape. Each clip in a clip catalog is assigned a color metric and a wavelet, i.e., shape, metric. The selected clip is also assigned a color metric and a wavelet metric. Those clips that are most similar based on either the wavelet metric or the color metric are returned to the user.

Claims

1. A method for computing a wavelet signature for an image comprising:

(a) centering the image on a field;

(b) converting the image to a standard color space;

(c) normalizing colors;

(d) performing a wavelet transform;

(e) finding a plurality of the largest coefficients by magnitude for each color plane;

(f) encoding a row/column location for each of the largest coefficients by increasing resolution;

(g) encoding the sign of the largest coefficients; and

(h) storing weight values for the row/column location for the largest coefficients.

2. The method of claim 1, wherein the field is white.

3. The method of claim 1, wherein the standard color space is YIQ.

4. The method of claim 1, wherein normalizing colors comprises scaling colors to vary from 0-1.

5. The method of claim 1, wherein wavelet transform is a standard Haar transform.

6. The method of claim 1, wherein the plurality of the largest coefficients is no more than 40 of the largest coefficients.

7. The method of claim 1, further comprising ignoring coefficients less than a threshold value.

8. A computer-readable medium having computer executable instructions for performing the method recited in any one of claims 1-7.

9. A system for computing a wavelet signature for an image comprising:

(a) a processing unit; and

(b) a storage medium coupled to the processing unit, the storage medium storing program code implemented by the processing unit for;

(i) centering the image on a field;

(ii) converting the image to a standard color space;

(iii) normalizing colors;

(iv) performing a wavelet transform;

(v) finding a plurality of the largest coefficients by magnitude for each color plane;

(vi) encoding a row/column location for each of the largest coefficients by increasing resolution;

(vii) encoding the sign of the largest coefficients; and

(viii) storing weight values for the row/column location for the largest coefficients.

10. The system of claim 9, wherein the field is white.

11. The system of claim 9, wherein the standard color space is YIQ.

12. The system of claim 9, wherein normalizing colors comprises scaling colors to vary from 0-1.

13. The system of claim 9, wherein wavelet transform is a Haar transform.

14. The system of claim 9, wherein the plurality of the largest coefficients is no more than 40 of the largest coefficients.

15. The system of claim 9, further comprising ignoring coefficients less than a threshold value.