This project proposes to develop a real-time face recognition algorithm which will be part of our people tracking from multiple cameras. Using the face recognition, we could re-identify pedestrians that the system already knows.
The project will be composed of the following steps; the first step includes independent detection of faces at each camera. Then, using the people detection algorithm, those faces will be tracked across cameras, and a model will be build for each one of them. Finally, the faces will be compared with "already identified" faces in order to identify the tracked people.
Students should have experience in The student should have experience in C++ programming. Experience in vision and 3D geometry is an advantage.
30% Theory, 40% Implementation (C++), 30% Practical experiments
For further information, send an e-mail regarding the project.
Contact: Horesh BenShitrit (office BC 307, tel 3 81 91).