Face Detection and Location System Based on Software and Hardware Co-design

摘要

Face localization was an important research direction of face recognition technology. Its aim was to segment the face from the background of the detecting image. This technology was widely used in many areas of research, such as identity verification, HMI, visual communication, virtual reality, public files, etc. In this paper, we firstly constructed skin model and calculated the similarity of the image data, then calculated the face boundaries based space projection. Importantly, in order to improve the real-time, we utilized Xilinx high-level synthesis tool Vivado HLS (AutoESL) to achieve a hardware from a C program, which was based on Zynq platform. And the hardware module is used to realize the face location greatly improved the computing speed. The simulation results show that the proposed method worked well and the efficiency increased by 80%.