Face recognition and registration for home surveillance system using iterative closest point and HAAR cascade lgorithm

Creating a new way to recognize and register human face, using 3D sensor at a consumer level computer or laptop. We develop method to recognize and register human face using three dimensional pixel values and two dimensional pixel values respectively. This uses IR depth sensors and RGB camera of Kinect for windows to collect depth input and frames from video respectively. We use threshold value, left and right value for the recognized face. This data is the features of the face. These features are stored in xml format which is then used to compare the recognized face against the existing face in the database.;We register group of members in the system. The features of the face are different even for same family members. We input an alert sound after registering all the users. When a new unregistered person enters in to the field of camera we generate the alert sound.;We can register only one person at a time, then the system can recognize 7 persons in the frame of the video. An alert is generated unless one registered person is in the frame. We use Iterative Closet Point algorithm for 3D face recognizing and haar cascade algorithm to extract features of the face.