Face Recognition Using Local Binary Patterns Histograms (LBPH) on an FPGA-Based System on Chip (SoC)

The need for facial recognition systems that are fast and accurate is continuously increasing. In this paper, a face recognition implementation on a System on Chip (SoC), integrated with an FPGA, is presented. This implementation utilizes Local Binary Patterns Histograms to extract features from test face images and Manhattan distance to retrieve the correct match from the system's face database. The SoC utilized is a Zynq-7030. The feature extraction and the distance computations, between the database, are implemented on the FPGA. The ARM processor of the SoC is responsible for receiving the input stream and presenting the output result, using the acquired distances. Real-time face recognition, with an execution time of 8.6 ms and accuracy of 79%, is achieved through this implementation.