Skull sex identification has a wide range of applications in forensic science, anthropology, and facial restoration. In this paper, the automatic gender identification of the skull was achieved by extracting the skull contour and applying CNN(Convolutional Neural Network). First, the three-dimensional model of the skull is reconstructed by preprocessing the skull data. Second, the Canny edge detection algorithm is improved to extract the skull contour. Finally, a skull gender recognition model is designed based on CNN. The method was validated on 115 complete skulls and 20 broken skull data sets, respectively. The accuracy rate of gender identification for the complete skull was 95.5%, and for the four types of incomplete skulls were 95.0%, 92.5%, 90.0%, and 77.5% respectively. The results show that the application of CNN for skull gender recognition does not require extraction of skull features, and the skull contour has good discrimination performance.
展开▼