Performance Optimization of Face Detection Algorithm

Abstract

Human face detection is known as one of the most concerned topics in numerous fields such as monitoring and surveillance system. In today's era, the existed face detection system is needed to be modified and optimized to adapt the changing demands. In this study, the face detection system with Viola-Jones algorithm had been used, modified, and optimized. The compilation of the optimized face detection system is conducted in Very High Speed Integrated Circuit Hardware Description (VHDL) language by using Quartus Prime 15.1 Lite Edition software. The new optimized face detection system is applied on the Altera DE2-115 Development and Education board. Several performance measurements such as power consumption, computing frequency and accuracy are carried out to examine the performance of the new optimized and implemented structure. To evaluate the accuracy of the new face detection system, the results of the measurements were compared to the results of the original face detection system. The obtained results had shown some improvement on the performance of the face detection system