LIVING BODY DETECTION METHOD, COMPUTER APPARATUS AND COMPUTER-READABLE STORAGE MEDIUM

摘要

A living body detection method, a computer apparatus and a computer-readable storage medium. The method comprises: training a multi-layer perceptron using a pre-set training set, so as to determine a multi-layer perceptron model (S101); acquiring N consecutive frames of a human face image to be detected (S102); transforming an intermediate frame of the human face image of the N consecutive frames of the human face image from a first color space to a second color space (S103); extracting textural features of the transformed intermediate frame of the human face image and dynamic mode features of the N consecutive frames of the human face image (S104, S105); fusing the textural features and the dynamic mode features to acquire fusion features (S106); performing feature mapping on the fusion features by using the multi-layer perceptron model, and outputting mapping features and performing normalization processing so as to obtain a predicted probability value of a living body tag and a predicted probability value of a non-living body tag (S107); and then determining the N consecutive frames of the human face image to be a living body or a non-living body human face image (S108, S109). The fusion features in the living body detection method, the computer apparatus and the computer-readable storage medium include textural features and dynamic mode features, and therefore, the recognition accuracy and safety of living body detection can be improved.