A Fast Detection Method Based on Deep Learning of Millimeter Wave Human Image

Abstract

Millimeter wave imaging technology has been a leading field for its perfect performance at airports and other secure locations. This paper tried to detect items concealed on human body with the deep learning method. The deep learning needs a lot of images to achieve an excellent result. As a result of that, this paper firstly collected lots of millimeter wave images and established corresponding human dataset for detection, then proposed a detection method based on threshold segmentation and Faster-RCNN. The experimental results on testing datasets valid the effectiveness of the proposed detection method.