Implementation of Face Recognition Processing Using an Embedded Processor

摘要

Embedded processors are conventionally difficult to use in face recognition in the security and robotic fields because of the tremendous amount of processing required. We implemented face recognition processing with a multicore based embedded processor having low power consumption and high performance. The single-chip multiprocessor is manufactured using a 0.15μm process with two M32R cores, 512KB of SRAM, and peripheral circuits integrated on a single-chip. It has a power supply voltage of 1.5V, a frequency of 600MHz, and power consumption of 800mW.