A hardware/software co-design approach for face recognition

Abstract

Face recognition is a technique employed in large-scale citizen identification applications, surveillance applications, law enforcement applications such as booking stations, and kiosks. Artificial neural networks (ANNs) have been proved to be an effective way to solve this problem, but due to the long-time training process, this approach cannot be implemented efficiently by software. Although, hardware implementations can speedup the training process, this may lead to an inflexible solution. To balance flexibility (i.e., software implementations) and performance (i.e., hardware implementations), an embedded computing system consisting of both a processor and dedicated hardware on a field programmable gate array (FPGA) chip is proposed to solve face recognition based on an ANN approach. Results obtained indicate that this system achieves almost twice the speedup over a pure software implementation.