Research on Living Face Detection Based on Dual Camera Contour Matching

Combining the advantages of visible-light camera and thermal-imaging camera, a research method of living face detection based on contour feature matching is proposed. In view of the different focal lengths of the two cameras, firstly, this paper establishes a camera positioning model to make the thermal-imaging image and the visible-light image spatial display range consistent; Secondly, the thermal-imaging image is preprocessed and the face contour is extracted by the improved multi-line convex hull detection algorithm; Thirdly, based on the YCbCr ellipse clustering model, the CbCr space of the visible-light skin color range is segmented, then the multi-motion convex hull detection is used to extract the visible-light image face contour; Finally, the contour matching algorithm is used to match the contours of the face extracted by the thermographic image and the visible-light image to determine whether it is a living face. The test results show that the method of this paper solves the application requirements of the national security department such as airports for living face identification.