A System for Disguised Face Recognition with Convolution Neural Networks

Abstract

Face recognition technology has been quite advanced in recent years and has been applied to various daily necessities and applications. However, people may make a false positive feature of the masked camouflage face because of makeup or wearing different equipment. In this paper, a two-stage disguise face recognition method based on CNN is proposed for the disguised face wearing equipment. In the first stage, we train a network that identifies the type of equipment and extracts the remaining faces that are not disguised. In the second stage of identification, the extracted remaining faces use the identified network for identity identification. The experimental results show that the proposed method has reached an average of 97.6% accuracy in the first stage of equipment type recognition. In the second stage of disguise face identification, 72.4% identification rate was obtained. The proposed method in this paper has reached the identification rate of the disguise identification research in recent years. The results of the above two stages show that the proposed method can effectively identify the type of disguise worn when people wear disguise. Then, the facial information of the disguise is removed to achieve a certain identity recognition effect.