Method, apparatus, computer program, computer system, and computer-readable storage medium for representing and searching for an object in an image

• US 7,492,972 B2
• Filed: 10/30/2007
• Issued: 02/17/2009
• Est. Priority Date: 07/05/1999
• Status: Expired due to Fees

• Current Assignee: Rakuten Group, Inc.
• Original Assignee: Mitsubishi Denki Kabushiki Kaisha (Mitsubishi Electric Corporation)
• Inventors: Bober, Miroslaw Z.
• Primary Examiner(s): Mariam; Daniel G
• Application Number:
  US11/929,434
• Publication Number:
  US 20080050049A1
• Time in Patent Office: 476 Days
• Field of Search: 382/103, 382/181, 382/190, 382/199, 382/201, 382/203, 382/305, 358/403, 707 1- 10
• US Class Current: 382/305
• CPC Class Codes: G06F 16/5854 using shape and object relationship
• CPC Class Codes: G06V 10/46 Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features colour feature extraction G06V10/56
• CPC Class Codes: G06V 10/752 Contour matching

Abstract

A method of representing an object appearing in a still or video image, by processing signals corresponding to the image, comprises deriving a plurality of numerical values associated with features appearing on the outline of an object starting from an arbitrary point on the outline and applying a predetermined ordering to said values to arrive at a representative of the outline.

Claims

1. A method of deriving a factor indicating the reliability for a descriptor representing an object appearing in an image or sequence of images, by processing signals corresponding to the image or images, the method comprising:

deriving a plurality of numerical values associated with features appearing on the outline of an object to form a descriptor of said outline;

deriving a factor indicating the reliability of said descriptor using a relationship between at least two of said values; and

storing the derived factor.

2. A method as claimed in claim 1 wherein the factor is based on the ratio between two of said values.

3. A method as claimed in claim 2 wherein the ratio is of the two greatest values.

4. A method as claimed in claim 1 wherein said values are values of a CSS curvature scale space representation.

5. A method as claimed in claim 4 wherein said values are peak height co-ordinate values.

6. A method as claimed in claim 1 wherein a curvature scale space representation of the outline is obtained by smoothing the outline in a plurality of stages using a smoothing parameter sigma, resulting in a plurality of outline curves, using values for the maxima and minima of the curvature of each outline curve to derive curves characteristic of the original outline, and selecting the co-ordinates of peaks of said characteristic curves as said numerical values.

7. A method as claimed in claim 1 further comprising applying a predetermined ordering to said numerical values to arrive at a representation of the outline.

8. A method as claimed in claim 7 wherein the predetermined ordering is such that the resulting representation is independent of the starting point on the outline.

9. A method as claimed in claim 7 wherein the values are ordered starting from the greatest value.

10. A method as claimed in claim 9 wherein the values are ordered in decreasing size.

11. A method as claimed in claim 7 wherein the values are ordered starting from the smallest value.

12. A method of searching for an object in an image or sequence of images, by processing signals corresponding to images, the method comprising

inputting a query in the form of a two-dimensional outline;

deriving a query descriptor of said outline;

deriving a factor for said query descriptor using a method as claimed in claim 1;

obtaining a descriptor of objects in stored images;

comparing said query descriptor with each descriptor for a stored object; and

selecting and displaying at least one result corresponding to an image containing an object for which the comparison indicates a degree of similarity between the query and said object.

13. A method as claimed in claim 12 wherein a factor is derived for each descriptor for a stored object by deriving a plurality of numerical values associated with features appearing on the outline of an object to form a descriptor outline;

deriving a factor indicating the reliability of said descriptor using a relationship between at least two of said values;

storing the derived factor ; and

wherein the comparison is made using a predetermined ordering only or the predetermined ordering and some other ordering depending on said factors.

14. An apparatus comprising:

an image database storing still and video images;

a descriptor database storing descriptors for objects or parts of objects appearing in the images stored in the image database; and

a processor configured to perform the method as claimed in claim 1 or claim 12.

15. A computer-readable storage medium storing computer-executable process steps for implementing a method as claimed in claim 1 or claim 12.

16. A computer system comprising:

an image database storing still and video images;

a descriptor database storing descriptors for objects or parts of objects appearing in the images stored in the image database; and

a control unit programmed to perform the method as claimed in claim 1 or claim 12.