Postprint version. Published in 2009 IEEE 7th Symposium on Application Specific Processors Proceedings: San Francisco, CA, July 27, 2009, pages 66-69.
Copyright © 2009 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The definitive version is available at http://dx.doi.org/10.1109/SASP.2009.5226339
NOTE: At the time of publication, the author Bridget Benson was not yet affiliated with Cal Poly.
This paper presents a parallelized architecture for hardware acceleration of multi-view face detection. In our architecture, the multi-view face detection system generates rotated image windows and their integral image windows for each classifier which perform parallel classification operations to detect non-upright (rotated) and non-frontal (profile) faces in the images. We use the training data from OpenCV to detect the frontal and profile faces based on the Viola and Jones algorithm. The proposed architecture for multi-view face detection has been designed using Verilog HDL and implemented in a Xilinx Virtex-5 FPGA. Its performance has been measured and compared with a Jones' and Viola's software implementation of multi-view face detection.
Electrical and Computer Engineering