Automatic Sex Identification Based on Convolution Neural Network and Least Square Method

Abstract

Sex identification has great application value in the field of forensic science and facial reconstruction. In view of the problem that traditional methods are mainly depend on plenty of artificial intervention, we present a novel automatic sex identification method based on Convolution Neural Network and Least Square Method. Firstly, multiple images of each sample are captured on the three-dimensional digitized skulls. Secondly, the probability values of sample images can be assessed by the Convolution Neural Network. Finally, we achieve sex identification using the Least Square Method to weight the probability values of sample images. This method abandons tedious manual measurement, and is easy to be applied by researchers without professional qualification. We implement our algorithm on 90 skulls and the experiments show that the method performs better than the state-of-the-art sex identification methods.