High accuracy and volume facial recognition on mobile platforms

Patent metadata

Field	Value
Publication number	US-11087121-B2
Application number	US-201916376345-A
Country	US
Kind code	B2
Filing date	Apr 5, 2019
Priority date	Apr 5, 2018
Publication date	Aug 10, 2021
Grant date	Aug 10, 2021

Abstract

Disclosed are systems and methods related to facial recognition. An image of a subject can be captured via a camera on a mobile device. The image can be classified according to a device type, whether the image is captured indoors or outdoors, and a standoff distance. Facial features can be extracted from the image based on the image category. The facial features can be compared with a predefined set of facial features in a database. An identification of the subject can be made in response to the comparison.

First claim

The invention claimed is:

1. A system, comprising:
at least one computing device; and
at least one application executable on the at least one computing device, wherein, when executed, the at least one application causes the at least one computing device to at least:
obtain a first image of an individual from a client device;
determine an image category for the first image based at least in part on a hierarchical classification that classifies by applying machine or deep learning algorithms to:
determine a mobile device type of a mobile device used to capture the first image, the mobile device comprising a smartphone or a tablet, and the mobile device type comprising a smartphone type or a tablet type;
determine whether the first image is captured indoors or outdoors following determining the mobile device type; and
determine a standoff distance of the first image following determining whether the first image is captured indoors or outdoors;

generate a reduced set of facial features based at least in part, on an initial set of facial features and a plurality of parameters that are based at least in part on the image category;
compare the reduced set of facial features of the first image with individual reduced sets of facial features associated with a plurality of second images within a database;
identify the individual within the first image based at least in part on the comparison; and
transmit data encoding an identification of the individual to the client.

2. The system of claim 1 , wherein the database comprises at least one of a CASIA face database or a Labeled Faces in the Wild (LFW) face database.

3. The system of claim 1 , wherein identification of the individual occurs in less than 0.1 seconds.

4. The system of claim 1 , wherein, when executed, the at least one application further causes the at least one computing device to at least generate the database, the database being generated by:
receiving the plurality of second images from at least one other computing device;
for individual second images of the plurality of second images:
identifying a respective image category based at least in part on at least one of a respective mobile device type associated with a respective second image, whether the respective second image is captured indoors or outdoors, and a respective standoff distance;
determining a first set of second image facial features; and
determining a second set of second image facial features based at least in part on the respective image category; and

storing the plurality of second images in a data store, the plurality of second images being classified according to the first set of second image facial features and the second set of second image facial features.

5. The system of claim 1 , wherein, when executed, the at least one application further causes the at least one computing device to at least detect a face within the first image.

6. The system of claim 1 , wherein, when executed, the at least one application further causes the at least one computing device to at least classify the standoff distance as a close distance or a far distance.

7. The system of claim 1 , wherein, when executed, the at least one application further causes the at least one computing device to at least extract the initial set of facial features from the first image.

8. The system of claim 1 , wherein the first image is encoded, and when executed, the at least one application further causes the at least one computing device to at least decode the first image in response to obtaining the first image from the client.

9. A system, comprising:
a mobile device comprising a smartphone or a tablet; and
at least one application executable on the first computing device, wherein, when executed, the at least one application causes the first computing device to at least:
capture an image of an individual using the mobile device;
transmit the image to a second computing device, the second computing device being configured to;
determine an image category for the image based at least in part on a hierarchical classification that classifies by applying machine or deep learning algorithms to:
determine a mobile device type of the mobile device used to capture the image, the mobile device type comprising a smartphone type or a tablet type;
determine whether the image is captured indoors or outdoors following determining the mobile device type; and
determine a standoff distance of the image following determining whether the image is captured indoors or outdoors;

generate a reduced set of facial features based at least in part on an initial set of facial features and a plurality of parameters that are based at least in part on the image category;
identify the individual in the image based at least in part on a comparison of the reduced set of facial features of the image with individual reduced sets of facial features associated with a database of images categorized according to mobile device type, a lighting type, and a standoff distance;

receive an identification of the individual from the second computing device; and
render a user interface including the identification for display via the first computing device.

10. The system of claim 9 , wherein, when executed, the at least one application causes the first computing device to at least encode the image.

11. The system of claim 9 , wherein the identification is made in less than 0.1 seconds.

12. A method, comprising:
obtaining, via at least one computing device, a first image of a subject captured from a mobile device comprising a smartphone or a tablet;
determining, via the at least one computing device, an image category associated with the first image based at least in part on a hierarchical classification that classifies by applying machine or deep learning algorithms to:
determine a mobile device type of the mobile device used to capture the first image, the mobile device type comprising a smartphone type or a tablet type;
determine whether the first image is captured indoors or outdoors following determining the mobile device type; and
determine a standoff distance of the first image following determining whether, the first image is captured indoors or outdoors:

determining, via the at least one computing device, an initial set of facial features of a face detected within the first image;
generating, via the at least one computing device, a reduced set of facial features based at least in part on the image category and the initial set of facial features;
comparing, via the at least one computing device, the reduced set of facial features with individual reduced sets of facial features associated with a plurality of second images in a database; and
identifying, via the at least one computing device, the subject based at east in part on the comparison.

13. The method of claim 12 , wherein the first image of the subject is captured by a camera on the mobile device, and wherein obtaining the first image comprises receiving the first image from the mobile device.

14. The method of claim 12 , wherein the image category is based at least in part on a hierarchical classification according to the mobile device type, the lighting type, and the standoff distance.

15. The method of claim 14 , wherein the standoff distance is classified as a far distance or a near distance.

16. The method of claim 12 , further comprising generating the database by:
receiving the plurality of second images from at least one other computing device;
for individual second images of the plurality of second images:
identifying a resp

Assignees

• Univ West Virginia

Inventors

• Bourlai Thirimachos
• Martin Michael

Classifications

• G06V40/168

  Feature extraction; Face representation · CPC title

• G06F16/583 Primary

  using metadata automatically derived from the content · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06V40/172

  Classification, e.g. identification · CPC title