Method for efficient coding of shape descriptor parameters

• US 7,447,373 B2
• Filed: 10/30/2007
• Issued: 11/04/2008
• Est. Priority Date: 02/29/2000
• Status: Expired due to Term

• Current Assignee: Rakuten Group, Inc.
• Original Assignee: Mitsubishi Electric Corporation
• Inventors: Bober, Miroslaw
• Primary Examiner(s): Do; Anh Hong
• Application Number:
  US11/928,895
• Publication Number:
  US 20080056596A1
• Time in Patent Office: 371 Days
• Field of Search: 382/251, 382/293, 382/295, 382/218, 382/219, 382/232, 345/420, 345/202, 704/500, 704/503
• US Class Current: 382/251
• CPC Class Codes: G06T 9/20 Contour coding, e.g. using detection of edges
• CPC Class Codes: G06V 10/40 Extraction of image or video features
• CPC Class Codes: G06V 10/44 Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
• 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

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 sets of co-ordinate values representing the shape of the object and quantising the co-ordinate values to derive a coded representation of the shape, and further comprises quantising a first co-ordinate value over a first quantisation range and quantising a smaller co-ordinate value over a smaller range.

Claims

1. A method of 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 sets of co-ordinate values representing the shape of the object; and

quantising the co-ordinate values to derive a coded representation of the shape, wherein, for a first set of coordinate values having a value for a given co-ordinate less than the corresponding co-ordinate value in a second set of co-ordinate values, the number of bits allocated to the quantised representation of the given coordinate value for the first set of values is less than the number of bits allocated to the quantised representation of the corresponding co-ordinate value for the second set of values.

2. A method as claimed in claim 1 wherein the quantisation range for the given co-ordinate value of the first set of co-ordinate values is based at least on the corresponding co-ordinate value of the second set of co-ordinate values.

3. A method as claimed in claim 2 wherein, for a sequence of decreasing co-ordinate values, the quantisation range for each co-ordinate value is based on one or more of the preceding, higher, co-ordinate values, where they exist.

4. A method as claimed in claim 1 wherein the range over which the given co-ordinate value is quantised is less for the first set of co-ordinate values than for the second set of co-ordinate values.

5. A method as claimed in claim 1 wherein the number of bits allocated to the quantised representation of the given co-ordinate value for the first set of values is same as the number of bits allocated to the quantised representation of the corresponding co-ordinate value for the second set of values.

6. A method as claimed in claim 1 wherein the sets of co-ordinate values are co-ordinate pairs, and the quantisation range varies for at least one coordinate value of each pair of co-ordinate values.

7. A method as claimed in claim 6 wherein the co-ordinate pairs correspond to the positions of the peaks in a CSS representation of a shape.

8. A method as claimed in claim 7 wherein the varying quantisation range is used for the co-ordinate value corresponding to the peak height.

9. A method as claimed in claim 1 wherein the quantisation range is the same for a plurality of co-ordinate values.

10. A method as claimed in claim 1 comprising the step of ordering co-ordinate values in decreasing or increasing size.

11. 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 object, deriving a representation of the query object, comparing the representation with the representations derived using a method as claimed in claim 1, and selecting and displaying those objects for which the representations indicate a degree of similarity to the query.

12. A computer readable medium having stored thereon computer executable program for searching for an object in an image or sequence of images by processing signals corresponding to images, the program when executed causes a computer to perform the steps of:

inputting a query object;

deriving a representation of the query object;

comparing the representation with the representations derived using a method as claimed in claim 1; and

selecting and displaying those objects for which the representations indicate a degree of similarity to the query.

13. An apparatus for representing an object appearing in an image or sequence of images, comprising:

means for deriving a plurality of sets of co-ordinate values representing the shape of the object; and

means for quantising the co-ordinate values to derive a coded representation of the shape;

wherein, for a first set of coordinate values having a value for a given co-ordinate less than the corresponding co-ordinate value in a second set of co-ordinate values, the number of bits allocated to the quantised representation of the given coordinate value for the first set of values is less than the number of bits allocated to the quantised representation of the corresponding co-ordinate value for the second set of values.

14. An apparatus searching for an object in an image or sequence of images, comprising:

means for inputting a query object;

means for deriving a representation of the query object;

means for deriving a plurality of sets of co-ordinate values representing the shape of the object;

means for quantising the co-ordinate values to derive a coded representation of the shape, wherein, for a first set of coordinate values having a value for a given co-ordinate less than the corresponding co-ordinate value in a second set of co-ordinate values, the number of bits allocated to the quantised representation of the given coordinate value for the first set of values is less than the number of bits allocated to the quantised representation of the corresponding co-ordinate value for the second set of values;

means for comparing the representation of the query object with the coded representation of the shape;

means for selecting and displaying those objects for which the representations indicate a degree of similarity to the query.

15. A computer readable medium having stored thereon computer executable program for representing an object appearing in an image or sequence of images, by processing signals corresponding to the image or images, the program when executed causes a computer to perform the steps of:

deriving a plurality of sets of co-ordinate values representing the shape of the object; and

quantising the co-ordinate values to derive a coded representation of the shape, wherein, for a first set of co-ordinate values having a value for a given co-ordinate less than the corresponding co-ordinate value in a second set of co-ordinate values, the number of bits allocated to the quantised representation of the given co-ordinate value for the first set of values is less than the number of bits allocated to the quantised representation of the corresponding co-ordinate value for the second set of values.