Local feature descriptor extracting apparatus, local feature descriptor extracting method, and program

• US 9,697,435 B2
• Filed: 11/15/2012
• Issued: 07/04/2017
• Est. Priority Date: 11/18/2011
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Iwamoto, Kota, Mase, Ryota
• Primary Examiner(s): Harandi, Siamak
• Assistant Examiner(s): Duffy, Kate R
• Application Number:
  US14/359,063
• Publication Number:
  US 20140328543A1
• Time in Patent Office: 1,692 Days
• Field of Search: None
• CPC Class Codes: G06F 18/2135 based on approximation criteria, e.g. principal component analysis
• CPC Class Codes: G06T 2207/20021 Dividing image into blocks, subimages or windows
• CPC Class Codes: G06T 2207/20112 Image segmentation details
• CPC Class Codes: G06V 10/462 Salient features, e.g. scale invariant feature transforms [SIFT]
• CPC Class Codes: G06V 10/7715 Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods

Abstract

An object of the present invention is to reduce a size of a feature descriptor while maintaining accuracy of object identification. A local feature descriptor extracting apparatus includes: a feature point detecting unit which detects a plurality of feature points in an image and which outputs feature point information that is information regarding each feature point; a feature point selecting unit which selects a prescribed number of feature points in an order of importance from the plurality of detected feature points, based on the feature point information; a local region acquiring unit which acquires a local region corresponding to each selected feature point; a subregion dividing unit which divides each local region into a plurality of subregions; a subregion feature vector generating unit which generates a feature vector of a plurality of dimensions for each subregion in each local region; and a dimension selecting unit which selects a dimension from the feature vector for each subregion so that a correlation between neighboring subregions is lowered, based on a positional relationship between subregions in each local region and which outputs an element of the selected dimension as a feature descriptor of the local region.

Claims

1. A local feature descriptor extracting apparatus comprising:

a memory that stores a program; and

a processor configured to run the program stored in the memory that is configured to cause the processor to

detect a plurality of feature points in an image and output feature point information that is information regarding each feature point,

select a prescribed number of feature points in a descending order of scale from the plurality of detected feature points, based on the feature point information,

acquire a local region corresponding to each selected feature point,

divide each local region into a plurality of subregions,

generate a feature vector of a plurality of dimensions for each subregion in each local region, and

select a dimension from the feature vector for each subregion so that at least one gradient direction differs between neighboring subregions, based on a positional relationship between subregions in each local region and output an element of the selected dimension as a feature descriptor of the local region.

2. The local feature descriptor extracting apparatus according to claim 1, wherein the program is further configured to cause the processor to determine the number of feature points to be selected and the number of dimensions to be selected.

3. The local feature descriptor extracting apparatus according to claim 2, wherein the program is further configured to cause the processor to receive specified feature descriptor size information that is information for specifying a total size of a feature descriptor of the selected feature points and determine the number of feature points and the number of dimensions based on the specified feature descriptor size information.

4. The local feature descriptor extracting apparatus according to claim 3, wherein the program is further configured to cause the processor to determine the number of feature points and the number of dimensions based on information indicating a correspondence relationship between the total size and the number of dimensions and on the specified feature descriptor size information.

5. The local feature descriptor extracting apparatus according to claim 3, wherein the program is further configured to cause the processor to determine the number of feature points and the number of dimensions based on information indicating a correspondence relationship between the total size and the number of feature points and on the specified feature descriptor size information.

6. The local feature descriptor extracting apparatus according to claim 3, wherein the program is further configured to cause the processor to determine the number of feature points and the number of dimensions based on information indicating a correspondence relationship among the total size, the number of feature points, and the number of dimensions and on the specified feature descriptor size information.

7. The local feature descriptor extracting apparatus according to claim 1, wherein the program is further configured to cause the processor to receive selection result information indicating a selection result of feature points and determine the number of dimensions based on the selection result information.

8. The local feature descriptor extracting apparatus according to claim 7, wherein

the selection result information includes importance information indicating an importance of each selected feature point, and

the program is further configured to cause the processor to determine the number of dimensions for each selected feature point based on the importance information.

9. A local feature descriptor extracting method for causing a computer to:

detect a plurality of feature points in an image and output feature point information that is information regarding each feature point;

select a prescribed number of feature points in a descending order of scale from the plurality of detected feature points, based on the feature point information;

acquire a local region corresponding to each selected feature point;

divide each local region into a plurality of subregions;

generate a feature vector of a plurality of dimensions for each subregion in each local region; and

select a dimension from the feature vector for each subregion so that at least one gradient direction differs between neighboring subregions is lowered, based on a positional relationship between subregions in each local region, and output an element of the selected dimension as a feature descriptor of the local region.

10. A non-transitory computer-readable storage medium storing a program for causing a computer to realize the functions of:

detecting a plurality of feature points in an image and outputting feature point information that is information regarding each feature point;

selecting a prescribed number of feature points in a descending order of scale from the plurality of detected feature points, based on the feature point information;

acquiring a local region corresponding to each selected feature point;

dividing each local region into a plurality of subregions;

generating a feature vector of a plurality of dimensions for each subregion in each local region; and

selecting a dimension from the feature vector for each subregion so that at least one gradient direction differs between neighboring subregions is lowered, based on a positional relationship between subregions in each local region, and outputting an element of the selected dimension as a feature descriptor of the local region.