The Design And Implementation Of Face Detection And Tracking System Based On FPGA

Nowadays,with the emergence of advanced science and technology,people's demands are increasing,Facial recognition has become the focus of scholars' research on its uniqueness,such as friendliness,immediacy,and quickness.The face detection and tracking technology has become a key step in face recognition,in the in-depth study,its security sector,intelligent human-computer interaction,digital surveillance,authentication and other fields of rapid development,and good application prospects.With the continuous development of chip manufacturing process technology,making the FPGA lower prices,integrated to improve,so that the reliability of FPGAs and applications increased,it shows their own advantages in the image processing and other areas.Our customers can build their own functional modules on the FPGA according to their requirements,and they could be free to modification and updating their modules.Because of his flexibility to greatly reduce the design cost,so that FPGA play an prominent part in the development and design of electronic.First of all,in research to a large number of face detection tracking and FPGA applications on the premise of technical information,we mainly study the commonly used face detection and tracking algorithms,and comparative of their advantages and disadvantages of their implementation with FPGA,in the end,.this paper use the detection method based on skin color segmentation and the tracking method based on face feature.On the basis of elliptical skin color segmentation,the algorithm is improved,the key detection and adaptive threshold module is added,and a complete detection and tracking system of face image is built.Secondly,this paper designed a hardware development system based on CMOS camera,VGA display,FPGA development board,storage and display.We complete the Verilog code of each module,and Modelsim simulation,verify the correctness of each module,and finally we can successfully collect images with the camera in real time,the acquisition effect is good.In this paper,the color space is YCbCr,in the detection and tracking method,the format conversion,elliptical skin color segmentation,threshold key adaptive,morphological filtering,face positioning tracking,face rectangle marking module are designed and implemented by Modelsim and SignalTab II Simulation to verify,verify its feasibility.Finally,the design of the program is verified in the development board.The results show that the detection rate of the face detection and tracking system is about 90%,and the detection speed is 30 fps to meet the purpose of real-time detection in the face of the collection of color a single face,small angle rotation of the face image.