Continuity Labeling Technique of Multiple Face in Multiple Frame

Research on recognizing and tracking objects have recently been carried out actively. There are especially high numbers of application fields using face recognition and tracking. Existing methods for recognizing and tracking objects have many difficulties when the target is multiple and of the same type. This study is on the continuous labeling method of the same face between frames in videos in which the faces of multiple people are included. The core of the algorithm is divided into detecting the face regions in one frame and recognizing faces using the previous frame, while applying the suitable methods. The usefulness of the proposed method was proven through experimentation and somewhat achievements were acquired through the test results.