In this technical demonstration, we introduce a real time face detection and recognition prototype. The proposed system can work with different image sources such as still images, videos from web cameras, and videos from ip cameras. The captured images are firstly processed by a cascaded classifier of Modified Census Transform (MCT) features to detect the faces. Then, facial features are detected inside the face region. These features are used to align and crop the face patches. Detection phase can be considerably improved by incorporating a tracking scheme to increase the hit rate while decreasing the false alarm rate. The registered faces are recognized using a novel method called Local Zernike Moments (LZM). A probabilistic decision step is employed in the final inference phase to provide a confidence margin. Introducing new identities via system's user interface is considerably simple since the system does not require retraining after each new identity.