Recognition, reidentification and security enhancements using autonomous machines

Patent metadata

Field	Value
Publication number	US-10496697-B2
Application number	US-201816123842-A
Country	US
Kind code	B2
Filing date	Sep 6, 2018
Priority date	Apr 24, 2017
Publication date	Dec 3, 2019
Grant date	Dec 3, 2019

Abstract

A mechanism is described for facilitating recognition, reidentification, and security in machine learning at autonomous machines. A method of embodiments, as described herein, includes facilitating a camera to detect one or more objects within a physical vicinity, the one or more objects including a person, and the physical vicinity including a house, where detecting includes capturing one or more images of one or more portions of a body of the person. The method may further include extracting body features based on the one or more portions of the body, comparing the extracted body features with feature vectors stored at a database, and building a classification model based on the extracted body features over a period of time to facilitate recognition or reidentification of the person independent of facial recognition of the person.

First claim

What is claimed is:

1. An apparatus comprising:
one or more processors including a graphics processor; and
a memory to store data, including image data generated by a camera;
wherein the one or more processors are to:
detect one or more objects within a physical vicinity based on one or more images captured by the camera, the one or more objects including a person, wherein detecting includes detecting utilizing one or more images of one or more portions of a body of the person,
extract body features based on the one or more portions of the body, wherein the one or more processors are to build a classification model based on extracted body features based on a plurality of objects over a period of time for recognition or reidentification of one or more persons independent of facial recognition,
determine whether the person can be recognized based on the extracted body features and the classification model, and, if so, perform identification of the person independent of facial recognition, and
upon determining that the person cannot be recognized based on the extracted body features and the classification model, perform a facial detection of the person.

2. The apparatus of claim 1 , wherein the one or more processors are further to train a classifier associated with a neural network to ensure the recognition or reidentification of the person independent of the facial recognition.

3. The apparatus of claim 2 , wherein the one or more processors are further to insert one or more verification checks at one or more layers of the neural network to check integrity of the neural network at each of the one or more layers to prevent malicious attacks.

4. The apparatus of claim 2 , wherein the one or more processors are further to facilitate separate parallel executions of the neural network and a software application associated with the one or more processors if the software application is subject to cyber-attacks, wherein the neural network is protected in a first execution unit, while the software application is quarantined in a second execution unit.

5. The apparatus of claim 2 , wherein the one or more processors are further to compare an output of the neural network with a pending decision of a decision-making entity, wherein based on the output of the neural network, the pending decision is at least one of altered, suspended, or maintained.

6. The apparatus of claim 1 , wherein the graphics processor is co-located with an application processor on a common semiconductor package.

7. The apparatus of claim 1 , wherein the one or more processors are further to, upon determining that a face of the person is detected in the facial detection of the person, perform facial recognition using the detected face of the person.

8. The apparatus of claim 7 , wherein the facial recognition includes storing a body image and person label from the from the facial recognition in a data store and providing data for a body feature extractor model to generate samples for a gallery of images.

9. The apparatus of claim 8 , wherein the one or more processors are further to, upon determining that a face of the person is not detected in the facial detection, perform a person re-identification for the person, including comparing the person with samples in the gallery of images.

10. A method comprising:
detecting one or more objects within a physical vicinity based on one or more images captured by a camera, the one or more objects including a person, wherein detecting includes detecting utilizing one or more images of one or more portions of a body of the person;
extracting body features based on the one or more portions of the body;
building a classification model based on extracted body features over a period of time to facilitate recognition or reidentification of persons independent of facial recognition;
determining whether the person can be recognized based on the extracted body features and the classification model, and, if so, performing identification of the person independent of facial recognition, and
upon determining that the person cannot be recognized based on the extracted body features and the classification model, performing a facial detection of the person.

11. The method of claim 10 , further comprising training a classifier associated with a neural network to ensure the recognition or reidentification of the person independent of the facial recognition.

12. The method of claim 11 , further comprising inserting one or more verification checks at one or more layers of the neural network to check integrity of the neural network at each of the one or more layers to prevent malicious attacks.

13. The method of claim 11 , further comprising facilitating separate parallel executions of the neural network and a software application associated with a graphics processor if the software application is subject to cyber-attacks, wherein the neural network is protected in a first execution unit, while the software application is quarantined in a second execution unit.

14. The method of claim 13 , wherein the graphics processor is co-located with an application processor on a common semiconductor package.

15. The method of claim 11 comprising comparing an output of the neural network with a pending decision of a decision-making entity, wherein based on the output of the neural network, the pending decision is at least one of altered, suspended, or maintained.

16. At least one non-transitory machine-readable medium comprising instructions that when executed by a computing device, cause the computing device to perform operations comprising:
detecting one or more objects within a physical vicinity based on one or more images captured by a camera, the one or more objects including a person, wherein detecting includes detecting utilizing one or more images of one or more portions of a body of the person;
extracting body features based on the one or more portions of the body;
building a classification model based on extracted body features over a period of time to facilitate recognition or reidentification of one or more persons independent of facial recognition;
determining whether the person can be recognized based on the extracted body features and the classification model, and, if so, performing identification of the person independent of facial recognition, and
upon determining that the person cannot be recognized based on the extracted body features and the classification model, performing a facial detection of the person.

17. The machine-readable medium of claim 16 , wherein the operations further comprise training a classifier associated with a neural network to ensure the recognition or reidentification of the person independent of the facial recognition.

18. The machine-readable medium of claim 17 , wherein the operations further comprise inserting one or more verification checks at one or more layers of the neural network to check integrity of the neural network at each of the one or more layers to prevent malicious attacks.

19. The machine-readable medium of claim 17 , wherein the operations further comprise facilitating separate parallel executions of the neural network and a software application associated with a graphics processor if the software application is subject to cyber-attacks, wherein the neural network is protected in a first execution unit, while the software application is quarantined in a second execution unit, wherein the graphics processor is co-located with an application processor on a common semiconductor package.

Assignees

• Intel Corp

Inventors

• Das Barnan
• Varerkar Mayuresh M
• Biswal Narayan
• Baran Stanley J
• Cilingir Gokcen
• Shah Nilesh V
• Sharma Archie
• Abdelhak Sherine
• Kotha Praneetha
• Pandit Neelay
• Weast John C
• Macpherson Mike B
• Kim Dukhwan
• Hurd Linda L
• Appu Abhishek R
• Koker Altug
• Ray Joydeep

Classifications

• G06V10/82 Primary

  using neural networks · CPC title

• G06V10/764

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

• G06F18/24143

  Distances to neighbourhood prototypes, e.g. restricted Coulomb energy networks [RCEN] · CPC title

• G06F16/784

  the detected or recognised objects being people · CPC title

• G06K9/00986

  Physics · mapped topic