# US 9,436,892 B2 · Method and apparatus for facial detection using regional similarity distribution analysis

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

- Number: US 9,436,892 B2
- Document ID: US9436892B2
- Status: Granted
- Kind: B2
- Application number: 14/268,149
- Filed: 2014-05-02
- Published: 2015-11-05
- Granted: 2016-09-06
- Earliest claimed: 2014-05-02
- Assignees: Hong Kong Applied Science and Technology Research Institute Company Limited
- Applicants: Hong Kong Applied Science and Technology Research Institute Co. Ltd.
- Inventors: Jinhui Hu; Yi Tian; Min Chen
- Web page: https://patentgrep.com/patent/US9436892B2

## CPC

- G06V 40/172: {Classification, e.g. identification}
- G06F 18/22: Matching criteria, e.g. proximity measures
- G06V 10/426: Graphical representations
- G06V 10/50: by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A method and apparatus for facial recognition by a computing device using an adaptive search window and regional similarity distribution analysis are disclosed herein.

## Claims

### Claim 1

A method of facial recognition by a computing device, the method comprising:
identifying, via a processor of the computing device, one or more features of a target image;
matching, via the processor, at least one of the one or more features of the target image with one or more features of a reference image;
determining, via the processor, a weighting contribution for the at least one matched feature;
determining, via the processor, a credibility of a facial recognition match based at least in part on the weighting contribution of the at least one matched feature, wherein the determining a weighting contribution comprises providing a higher weighting for the at least one matched feature based at least in part on a number of other matched features relative to the at least one matched feature; and
determining a utilization ratio based at least in part on a ratio of a matching area to a surface area of a detection frame.

### Claim 2 (depends on 1)

The method of claim 1, wherein the determining a weighting contribution comprises providing a higher weighting for the at least one matched feature based at least in part on a distance to one or more other matched features.

### Claim 3 (depends on 1)

The method of claim 1, wherein the matching the at least one of the one or more features comprises using an adaptive window.

### Claim 4 (depends on 1)

The method of claim 1, wherein the determining a weighting contribution comprises providing a lower weighting for the at least one matched feature based at least in part on a determination of geometric isolation.

### Claim 5 (depends on 3)

The method of claim 3, further comprising centering the adaptive window on the at least one matched feature.

### Claim 6 (depends on 1)

The method of claim 1 further comprising determining a final similarity factor of the facial recognition match based at least in part on the utilization ratio.

### Claim 7 (depends on 3)

The method of claim 3, wherein the adaptive window is a circle with a radius between
[
a
×
b
2
20
∼
a
×
b
2
5
]
,
where a refers to a detection frame height, and b refers to a detection frame width.

### Claim 8 (depends on 1)

The method of claim 1, wherein the at least one matched feature is included in a matched feature pool.

### Claim 9

An article including a non-transitory computer readable medium having instructions thereon, the instructions executable to:
identify one or more features of a target image;
match, using an adaptive window, at least one of the one or more features of the target image with one or more features of a reference image;
determine a weighting contribution for the at least one matched feature;
determine a credibility of a facial recognition match based at least in part on the weighting contribution of the at least one matched feature; and
determine a utilization ratio based at least in part on a ratio of a matching area to a surface area of a detection frame.

### Claim 10 (depends on 9)

The article of claim 9, wherein the instructions executable to determine a weighting contribution are further to provide a higher weighting for the at least one matched feature based at least in part on a distance to one or more other matched features.

### Claim 11 (depends on 9)

The article of claim 9, wherein the instructions executable to determine a weighting contribution are further to provide a higher weighting for the at least one matched feature based at least in part on a number of other matched features also within the adaptive window.

### Claim 12 (depends on 9)

The article of claim 9 wherein the instructions executable to determine a weighting contribution are further to provide a lower weighting for the at least one matched feature based at least in part on a determination of geometric isolation.

### Claim 13 (depends on 9)

The article of claim 9, wherein the instructions executable to match the at least one of the one or more local features using an adaptive window are further to center the adaptive window on the at least one matched feature.

### Claim 14 (depends on 9)

The article of claim 9, wherein the instructions are further to: determine a final similarity factor of the facial recognition match based at least in part on the utilization ratio.

### Claim 15

A method comprising:
matching, via a processor of a computing device, at least one feature of a target image with at least one feature of a reference image using an adaptive window;
determining, via the processor, a weighting contribution for the matched at least one feature;
determining, via the processor, a credibility of a facial recognition match based at least in part on the weighting contribution of the at least one matched feature; and
determining a utilization ratio based at least in part on a ratio of a matching area to a surface area of a detection frame.

### Claim 16 (depends on 15)

The method of claim 15, wherein the determining a weighting contribution comprises providing a higher weighting for the matched at least one feature based at least in part on a distance to one or more other matched features.

### Claim 17 (depends on 15)

The method of claim 15, wherein the determining a weighting contribution comprises providing a higher weighting for the matched at least one feature based at least in part on a number of other matched features within the adaptive window.

### Claim 18 (depends on 15)

The method of claim 15, wherein the determining a weighting contribution comprises providing a lower weighting for the matched at least one feature based at least in part on a geometric isolation determination.

### Claim 19 (depends on 15)

The method of claim 15, wherein the adaptive window is a circle with a radius between
[
a
×
b
2
20
∼
a
×
b
2
5
]
,
where a refers to a detection frame height, and b refers to a detection frame width.

### Claim 20 (depends on 15)

The method of claim 15 further comprising determining a final similarity factor of the facial recognition match based at least in part on the utilization ratio.