# US 8,280,173 B2 · Feature point location determination method and apparatus

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 8,280,173 B2
- Document ID: US8280173B2
- Status: Granted
- Kind: B2
- Application number: 12/393,508
- Filed: 2009-02-26
- Published: 2009-09-03
- Granted: 2012-10-02
- Earliest claimed: 2008-02-29
- Assignees: Canon Kabushiki Kaisha
- Inventors: Masami Kato; Takashi Suzuki
- Web page: https://patentgrep.com/patent/US8280173B2

## CPC

- G06V 40/171: {Local features and components; Facial parts (eye characteristics G06V40/18); Occluding parts, e.g. glasses; Geometrical relationships}
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A method for determining feature point locations in an image performs a first search in a predetermined first search area to search for locations of plural feature points in the image, corrects the locations of the plural feature points based on a geometric layout relationship among the plural feature points searched for, sets a second search area based on the corrected location of each of the feature points, and performs a second search in the second search area to search for the location of each of the feature points. Then, the method determines reliability of the location of each feature point searched for by the second search and selects one of the corrected location and the location searched for by the second search, as a location of the feature point.

## Claims

### Claim 1

A feature point location determination method comprising: the following steps and using a processor to perform one or more of said steps:
a first search step of searching for locations of a plurality of feature points in a predetermined first search area in an image;
a location correction step of correcting the locations of the plurality of feature points based on a geometric layout relationship among the plurality of feature points searched for by in said first search step;
a second search step of setting a second search area based on the locations of the plurality of feature points corrected by in said location correction step and searching for locations of the feature points in the second search area;
a reliability calculation step of calculating reliability of each of the feature points searched for from the second search area based on reliability distribution in the second search area;
a distance calculation step of calculating distance between the locations of each of the feature points before and after correction in said location correction step;
a determination step of determining a threshold based on the distance calculated in said distance calculation step; and
a selection step of selecting one of the location corrected in said location correction step and the location searched for in said second search step, as the location of each of the feature points to be localized, based on a comparison result between the reliability calculated in said reliability calculation step and the threshold determined in said determination step.

### Claim 2

A feature point location determination apparatus comprising:
a first search unit configured to search for locations of a plurality of feature points in a predetermined first search area in an image;
a location correction unit configured to correct the locations of the plurality of feature points based on a geometric layout relationship among the plurality of feature points searched for by said first search unit;
a second search unit configured to set a second search area based on the locations of the feature points corrected by said location correction unit and configured to search for locations of the feature points in the second search area;
a reliability calculation unit configured to calculate reliability of each of the feature points searched for from the second search area based on reliability distribution in the second search area;
a distance calculation unit configured to calculate distance between the loactions of each of the feature points before and after by said location unit:
a determination unit configured to determine a threshold based on the distance calculated by said distance calculation unit; and
a selection unit configured to select one of the location corrected by said location correction unit and the location searched for by said second search unit, as the location of each of the feature points to be localized, based on a comparison result between the reliability calculated by said reliability calculation unit and the threshold determined by said determination unit.

### Claim 3 (depends on 1)

A non-transitory computer-readable medium storing a program which makes a computer perform the feature point location determination method according to claim 1.

### Claim 4 (depends on 1)

The method according to claim 1, wherein the second search area is smaller than the first search area.

### Claim 5 (depends on 1)

The method according to claim 1, wherein said location correction step generates a projection vector by projecting an input vector whose elements are made up of coordinate values of the feature points onto an eigenspace using a pre-generated eigenvector, inverse-projects the projection vector using the eigenvector, and thereby acquires corrected feature point locations.

### Claim 6 (depends on 5)

The method according to claim 5, wherein the eigenvector is generated by principal component analysis.

### Claim 7 (depends on 1)

The method according to claim 1, further comprising a generation step of generating a reliability map which represents a distribution of reliability of existence of each of the feature points in the image, wherein
said first search step and said second search step search for the location of each of the feature points based on the reliability map.

### Claim 8 (depends on 1)

The method according to claim 1, further comprising a generation step of generating a reliability map which represents a distribution of reliability of existence of each of the feature points in the image,
wherein said first search step searches for the location of each of the feature points based on the reliability map; and
said second search step searches for the location of each of the feature points using the image.

### Claim 9 (depends on 1)

The method according to claim 1, further comprising a change step of changing size of the second search area according to the distance calculated in said distance calculation step.

### Claim 10 (depends on 1)

The method according to claim 1, further comprising a change step of changing the second search area for each type of feature point to be localized.

### Claim 11 (depends on 1)

The method according to claim 1, wherein said determination step determines the threshold based on conditions set for each type of feature point to be localized.

### Claim 12 (depends on 1)

The method according to claim 1, wherein the feature points to be localized are facial organs of an animal or characteristic parts of an organ.
