Invention Grant
- Patent Title: System and method for automatic detection and recognition of people wearing personal protective equipment using deep learning
-
Application No.: US16996668Application Date: 2020-08-18
-
Publication No.: US11482030B2Publication Date: 2022-10-25
- Inventor: Radu Tudor Ionescu , Mariana Iuliana Georgescu , Adrian Şandru , Georgian-Emilian Duţă
- Applicant: SecurifAI LLC
- Applicant Address: RO Azuga
- Assignee: SecurifAI LLC
- Current Assignee: SecurifAI LLC
- Current Assignee Address: RO Azuga
- Agency: Leason Ellis LLP
- Main IPC: G06V40/10
- IPC: G06V40/10 ; G06V30/194 ; G08B21/02

Abstract:
A system and method for determining whether persons depicted in images are wearing personal protective equipment (PPE) is disclosed. The system includes a PPE detection platform that, during a testing phase, comprises a person detector configured to detect persons in images and a deep neural network PPE classifier. The PPE classifier comprises a spatial attention module and is trained to predict whether or not a PPE item is present in the image sub-region corresponding to the detected person. During a training phase, the PPE detection platform comprises a body pose estimator that, based on the output of the person detector, determines a pose of the detected person's body and provides a spatial attention supervision signal used to train the spatial attention module of the PPE classifier and tune the PPE classifier. The PPE platform is configured to bypass the body pose estimator during the inference phase.
Public/Granted literature
Information query