# US 2024/0404320 A1 · System and method for facial recognition accuracy

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

- Number: US 2024/0404320 A1
- Document ID: US20240404320A1
- Status: Published application
- Kind: A1
- Application number: 18/731,304
- Filed: 2024-06-02
- Published: 2024-12-05
- Earliest claimed: 2018-05-10
- Assignees: Wicket, LLC
- Applicants: Wicket, LLC
- Inventors: Gennady Livitz; Patrick Quinlan; Samson Timoner
- Web page: https://patentgrep.com/patent/US20240404320A1

## CPC

- G06V 40/172: {Classification, e.g. identification}
- G06F 16/784: {the detected or recognised objects being people}
- G06F 18/22: Matching criteria, e.g. proximity measures
- G06V 10/761: {Proximity, similarity or dissimilarity measures}
- G06V 20/52: Surveillance or monitoring of activities, e.g. for recognising suspicious objects (recognising microscopic objects G06V20/69)
- G06V 40/167: {using comparisons between temporally consecutive images}
- G06V 10/32: Normalisation of the pattern dimensions
- G06V 40/23: {Recognition of whole body movements, e.g. for sport training}
- G07C 9/00: Individual registration on entry or exit
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G07C: TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE

## Abstract

The invention provides, in some respects, a digital data processing method of identifying a person of interest in a video stream. The method, which executes on a digital data processing system including one or more digital data processing devices in communications coupling with one another, includes detecting using a detector and without using a facial recognizer one or more faces in a video stream. It further includes using a tracker to associate, without using a facial recognizer, one or more of the detected faces with a same apparent individual, and scoring, again, without using a facial recognizer, each of the detected faces that are associated with the same apparent individual. The method further includes applying selected ones of the scored detected faces of the same apparent individual to a facial recognizer to identify a person of interest, where selection is based on scores of the detected faces.

## Claims

### Claim 1

A digital data processing method of identifying a person of interest in a video stream comprising executing, on a digital data processing system that includes one or more digital data processing devices that are in communications coupling with one another, the steps of
A. detecting, using a detector and without using a facial recognizer, one or more faces in a video stream,
B. using a tracker, and without using a facial recognizer, to associate one or more of the detected faces with a same apparent individual,
C. scoring, without using a facial recognizer, each of the detected faces that are associated with the same apparent individual,
D. applying selected ones of the scored detected faces of the same apparent individual to a facial recognizer to identify a person of interest, where selection is based on scores of the detected faces.

### Claim 2 (depends on 1)

The digital data processing method of claim 1, comprising executing on the digital data processing system a step of modifying any of the video stream and a presentation thereof to highlight the person of interest.

### Claim 3 (depends on 1)

The digital data processing method of claim 1, comprising executing on the digital data processing system a step of generating an alert regarding the person of interest.

### Claim 4 (depends on 1)

The digital data processing method of claim 1, comprising executing on the digital data processing system a step of adjusting a threshold associated with the facial recognizer as a function of the number of selected images applied thereto.

### Claim 5 (depends on 1)

The digital data processing method of claim 1, wherein step (D) comprises foregoing application to the facial recognizer of scored detected faces whose scores fall below a threshold.

### Claim 6 (depends on 1)

The digital data processing method of claim 1, comprising executing on the digital data processing system a step of monitoring, over a continuous interval of time, scores of scored detected faces associated with the same apparent individual.

### Claim 7 (depends on 6)

The digital data processing method of claim 6, wherein step (D) comprises applying to the facial recognizer one or more of the scored detected faces being monitored, if their monitored scores fail to increase within a specified period of time.

### Claim 8 (depends on 7)

The digital data processing method of claim 7, wherein step (D) comprises applying to the facial recognizer one or more of the scored detected faces being monitored, if after the specified period of time their scores increase.

### Claim 9 (depends on 1)

The digital data processing method of claim 1, wherein
step (A) comprises detecting, using the detector and without using a facial recognizer, a said face of a first individual in a first video frame of the video stream, and
step (B) comprises using the tracker to track in multiple frames of the video stream a said face apparently associated with the first individual.

### Claim 10 (depends on 9)

The digital data processing method of claim 9, comprising executing on the digital data processing system the step of selecting with a selector function faces in the multiple frames apparently associated with the first individual, and applying those selected faces to a facial classifier of the facial recognizer.

### Claim 11 (depends on 10)

The digital data processing method of claim 10, comprising executing on the digital data processing system the steps of
using an embedder to map each face selected in the selecting step to a vector of values and using the vectors of values in the facial classifier to identify a face as being that of a person of interest.

### Claim 12 (depends on 1)

The digital data processing method of claim 1, wherein step (B) includes disambiguating detected faces to identify those associated with the same apparent individual.

### Claim 13 (depends on 12)

The digital data processing method of claim 12, wherein step (B) includes disambiguating detected faces based on any of motion and/or position in fields of view captured by the video stream.

### Claim 14 (depends on 13)

The digital data processing method of claim 13, wherein the step of disambiguating includes comparing features of the detected faces without using a facial recognizer.

### Claim 15 (depends on 1)

The digital data processing method of claim 1, wherein step (D) includes disambiguating among scored detected faces that are apparently associated with the same apparent individual by clustering the scored detected faces.

### Claim 16 (depends on 15)

The digital data processing method of claim 15, wherein the clustering step includes using an embedder to map each of the scored detected faces to a respective vector of values.

### Claim 17 (depends on 15)

The digital data processing method of claim 15, wherein step (D) includes utilizing time as a factor in clustering the selected detected faces to identify those associated with different individuals.

### Claim 18

A digital data processing method of identifying a person of interest in a video stream, comprising executing on a digital data processing system that includes one or more digital data processing devices that are in communications coupling with one another the steps of
A. detecting, using a detector and without using a facial recognizer, a face of a first individual in a first video frame of a video stream,
B. using a tracker to track, without using a facial recognizer, faces associated with the first individual detected in multiple frames of the video stream,
C. selecting, with a selector function, faces in the multiple frames associated with the first individual, and applying those selected faces to a facial classifier,
D. using an embedder to map each face selected in the selecting step to a vector of numbers,
E. using the vectors of numbers in the facial classifier to identify the first individual,
F. scoring, without using a facial recognizer, each of the detected faces of the first individual,
G. applying selected ones of the scored detected faces of the first individual to a facial recognizer to identify whether he or she is a person of interest, where selection is based on scores of the detected faces.

### Claim 19 (depends on 18)

The digital data processing method of claim 18, comprising executing on the digital data processing system a step of adjusting a threshold associated with the facial recognizer as a function of the number of selected images applied thereto.

### Claim 20 (depends on 18)

The digital data processing method of claim 18, wherein step (G) comprises foregoing application to the facial recognizer of scored detected faces whose scores fall below a threshold.

### Claim at position 21

21-26. (canceled)
