Automatic target recognition method for millimeter-wave body scanner

Abstract

Automatic target recognition is a significant function of millimeter-wave body scanner. Based on the principle of saliency and sparse coding, this paper proposes a new method for millimeter-wave body image processing and recognition. According to the characteristics such as shape, texture and gray level of the millimeter wave image, the method realizes automatic recognition for concealed objects. A large number of experiments verify the accuracy of the method on detecting objects concealed in different position of human body.