Computerized system and method for adaptive stranger detection

Patent metadata

Field	Value
Publication number	US-11126830-B2
Application number	US-201916573245-A
Country	US
Kind code	B2
Filing date	Sep 17, 2019
Priority date	Sep 17, 2019
Publication date	Sep 21, 2021
Grant date	Sep 21, 2021

Abstract

Disclosed are systems and methods for improving interactions with and between computers in computerized security and content monitoring, hosting and providing devices, systems and/or platforms. The disclosed systems and methods provide a novel framework that adaptively distinguishes between known people versus unknown people based on a dynamically applied, anonymous facial recognition methodology. The disclosed framework provides such functionality by recognizing faces within captured images without storing any information or annotations regarding or revealing the captured person's identity. The framework is configured to adaptively learn to distinguish between faces seen for the first time and faces it has previously seen by locally processing a captured image and only sending face embeddings to a network location for future comparisons of subsequently, anonymously captured images.

First claim

What is claimed is:

1. A method comprising the steps of:
identifying, via a computing device, an image comprising content depicting a person at a location;
analyzing, via the computing device, said image, and based on said analysis, determining information associated with a face of the person depicted by said content, said face information comprising data indicating characteristics of traits of said face;
comparing, via the computing device, the face information to each face embedding stored in a gallery hosted by storage, each stored face embedding comprising face information for previous person depictions captured at said location, the stored face embeddings being ordered in said gallery according to how recent a respective person depiction was observed at said location, said comparison comprising determining a similarity value for each stored face embedding indicating how similar each stored face embedding is to said face information;
identifying, via the computing device, a stored face embedding having a highest similarity value;
comparing, via the computing device, said highest similarity value to a similarity threshold;
determining, via the computing device, whether said person is a known person or a stranger based on said comparison, wherein said person is known when said similarity threshold is satisfied, wherein said person is a stranger when said similarity threshold is not satisfied.

2. The method of claim 1 , further comprising:
updating said gallery when said person is determined to be a known person, said updating comprising moving said identified stored face embedding to a first position within said gallery, said updating further comprising updating a recency value for said identified stored face embedding to indicate said determination.

3. The method of claim 1 , further comprising:
communicating an alert indicating that an unknown person is at said location when said person is identified as said stranger.

4. The method of claim 1 , wherein said face information within each face embedding is represented by a feature vector.

5. The method of claim 1 , wherein said gallery is configured as a double linked list of tuples, wherein each face embedding is represented by a tuple and each tuple is connected to its preceding and following tuple in said list.

6. The method of claim 1 , further comprising:
automatically capturing, by a security system associated with said computing device, said image, said capturing occurring automatically based on detection of said person at the location, wherein said identification of the image is based on said capturing.

7. The method of claim 1 , further comprising:
analyzing said image, and based on said analysis, identifying said face of the person within the content of the image; and
cropping said image based on said identified face, wherein said analysis of said image is based on said cropped image.

8. The method of claim 1 , wherein said image is a video frame within a captured video.

9. The method of claim 1 , further comprising:
updating said gallery when said person is determined to be a stranger, said updating comprising storing said face information as new face embedding data within said gallery, said storage comprising inserting said new face embedding data into a first position within said gallery and moving each previously stored face embedding down a position.

10. The method of claim 9 , further comprising:
determining that said insertion of the new face embedding data causes said database to exceed a predetermined size;
identifying face embedding data at the last position in the gallery; and
deleting said identified face embedding data.

11. The method of claim 10 , further comprising:
determining a usage of said computing device based at least in part on said location; and
setting said predetermined size of said gallery based on said usage, said predetermined size indicating a maximum number of face embeddings capable of being stored.

12. A non-transitory computer-readable storage medium tangibly encoded with computer-executable instructions, that when executed by a processor associated with a computing device, performs a method comprising the steps of:
identifying, via the computing device, an image comprising content depicting a person at a location;
analyzing, via the computing device, said image, and based on said analysis, determining information associated with a face of the person depicted by said content, said face information comprising data indicating characteristics of traits of said face;
comparing, via the computing device, the face information to each face embedding stored in a gallery hosted by storage, each stored face embedding comprising face information for previous person depictions captured at said location, the stored face embeddings being ordered in said gallery according to how recent a respective person depiction was observed at said location, said comparison comprising determining a similarity value for each stored face embedding indicating how similar each stored face embedding is to said face information;
identifying, via the computing device, a stored face embedding having a highest similarity value;
comparing, via the computing device, said highest similarity value to a similarity threshold;
determining, via the computing device, whether said person is a known person or a stranger based on said comparison, wherein said person is known when said similarity threshold is satisfied, wherein said person is a stranger when said similarity threshold is not satisfied.

13. The non-transitory computer-readable storage medium of claim 12 , further comprising:
updating said gallery when said person is determined to be a known person, said updating comprising moving said identified stored face embedding to a first position within said gallery, said updating further comprising updating a recency value for said identified stored face embedding to indicate said determination.

14. The non-transitory computer-readable storage medium of claim 12 , further comprising:
communicating an alert indicating that an unknown person is at said location when said person is identified as said stranger.

15. The non-transitory computer-readable storage medium of claim 12 , further comprising:
updating said gallery when said person is determined to be a stranger, said updating comprising storing said face information as new face embedding data within said gallery, said storage comprising inserting said new face embedding data into a first position within said gallery and moving each previously stored face embedding down a position.

16. The non-transitory computer-readable storage medium of claim 15 , further comprising:
determining that said insertion of the new face embedding data causes said database to exceed a predetermined size;
identifying face embedding data at the last position in the gallery; and
deleting said identified face embedding data.

17. The non-transitory computer-readable storage medium of claim 16 , further comprising:
determining a usage of said computing device based at least in part on said location; and
setting said predetermined size of said gallery based on said usage, said predetermined size indicating a maximum number of face embeddings capable of being stored.

18. A computing device comprising:
a processor; and
a non-transitory computer-readable storage medium for tangibly storing thereon program logic for execution by the processor, the program logic comprising:
logic executed by the processor for identifying, via the computi

Assignees

• Verizon Media Inc

Inventors

• Zhang Yang
• Dassa Guy
• Lahtiranta Atte

Classifications

• G06V40/168

  Feature extraction; Face representation · CPC title

• G06V10/449

  Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters · CPC title

• G06N20/00 Primary

  Machine learning · CPC title

• G06V10/7715

  Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods · CPC title

• G06V10/761

  Proximity, similarity or dissimilarity measures · CPC title