Personal Identification Methods by Using Facial and Clothing Features for Entry and Exit Management and Its Identification Using Multiple Frames

Abstract

In this research, we develop user identification methods using facial and clothing features as methods without user operation. We propose two types of integration methods of facial and clothing features: the one based on thresholds and weights and the other using machine learning. Further, we propose an identiflcation method from multiple frames. The evaluation results demonstrate the effectiveness of the proposed methods.