Face Skin Color Detection Method Based on YUV-KL Transform

Abstract

In recent years, with the development of activities such as e-commerce, face recognition has become a possible way to verify identity characteristics. The purpose of this paper is to study the method of face skin color detection based on YUV-KL transform. A method based on skin color and facial features is proposed to detect human faces. First, the skin color information is used for detection, and then the face features are used for verification. The first is face detection based on skin color. The work to be done in this process includes: the selection of color space, the analysis of YCbCr color space, the introduction of Gaussian density function estimation skin color model, and the use of simple skin color model. The second is KL transform to extract face features (features of eyes and mouth), and verify the detected faces. YUV-KL color space and KL feature extraction are used in face detection, two important contents, and compared with the traditional face skin color detection method. The experimental results show that the number of missed detections of this method is only three images.