People Tracking in Low Rank Representation and Face Recognition Using Neural Network

Automated video analysis plays an important role in video surveillance, traffic monitoring, vehicle navigation and in many applications. Face recognition is also a challenging problem in the field of image analysis especially when dealt with video sequences. This work is on people tracking and their face recognition in video analytics. An algorithm named Low Rank Representation is used to efficiently seperate the foreground person and the background. Low rank does not need any training sequence and it can deal with the complex background. The face of the tracked person is captured from the video frame and then it is compared with the suspected images stored in the database. Face recognition is done by the morphological neural network which uses backpropagation algorithm.