Face tracking and recognition in low quality video sequences with the use of particle filtering

Abstract

A system for parallel face detection, tracking and recognition in real-time video sequences is being developed. The paper describes its particle filtering based face recognition module, which operates on low quality video sequences and utilizes the results of the preceding phases of face detection and tracking. The temporal information from video sequence is utilized for the purpose of object tracking and identity recognition through knowledge cumulation. The performance of the solution is presented in closed-set and open-set identification scenarios.