Image Acquisition for VLSI Face Detection System

Abstract

Face detection is the process to identify the face biometrics of people while image acquisition is the fundamental of face detection. Image acquisition system is a series process in capturing and loading image to Field-Programmable Gate Array (FPGA) board. This study implemented Viola-Jones face detection algorithm into DE2-115 board using TRDB-D5M as camera to acquire image. The captured image is segmented and stored to allocated memory address for image processing and external monitor is used to display the output. The first part of this study is using OV7670 camera and DE2-115 board to collect data for face detection in Very High Speed Integrated Circuit Hardware Description (VHDL) language while the second part is using TRDB-D5M and DE2-115 to test the function of TRDB-D5M camera in Verilog language. Then the last part of this study is the implementation of the TRDB-5M camera into the face detection system in VHDL language. The result obtained had shown that with OV7670 camera as the input, the acquisition of image is successfully done, stored and displayed on the monitor. However, the camera resolution is very low and the displayed images are very small in size. In enhancing the face detection system, TRDB-D5M camera is used to replace the OV7670 camera. The new setup is successful in displaying stream images showing that the implemented segmentation of the acquired image to the allocated memory is success, but for the still image only the default image is displayed.