Anomaly detection and action recognition for mobile cameras

In this project we want to develop a few AI algorithms for the security and public health monitoring applications that can be implemented on a mobile camera. This camera can be either mounted on the autonomous mobile robot or on the wearable devices that security guards are equipped with. This project is aiming to solve the following challenges: understanding the location of the camera based on the footage, anomaly detection, and action recognition. These goals can be achieved through a combination of deep learning, traditional computer vision, and machine learning methods. This understanding can help the security guard or mobile robot in performing the patrolling missions by increasing the response time and improving the consistency and quality of the reports. Currently, most of the research, and a large portion of the datasets are focused on the action recognition and anomaly detection of the stationary camera footages. In this research we want to specifically solve these problems for mobile camera footage.

Faculty Supervisor:

Mo Chen

Student:

Mohammad Hadi Salari

Partner:

Tellext

Discipline:

Computer science

Sector:

Manufacturing

University:

Simon Fraser University