APPLYING FPGA TECHNOLOGY FOR AI FACIAL RECOGNITION CAMERA

Le Ba Cuong; Nguyen Quang Dich

Le Ba Cuong ELOM Technology Media Joint Stock Company; Hanoi University of Science and Technology
Nguyen Quang Dich Hanoi University of Technology and Education

Keywords: FPGA, AI Camera, Facial recognition camera, Deep learning

Abstract

Recently, the fast development of artificial intelligence (AI) has led to increasingly requirements of powerful hardware for real-time processing of AI models with huge computational volumes. Among them, FPGA has emerged as a new and potential direction in building hardware for AI camra because FPGA has a number of outstanding advantages such as: Parallel computing ability for higher processing performance, ability to customize hardware depending on the application helps reduce energy consumption, and more flexible in embedded systems due to reprogrammability. This article introduces the application of FPGA technology for AI Face recognition cameras with accuracy up to 99 %, recognition range within a distance of 2 - 4 m, recognition speed reaching about 6 frames per second with a power consumption of about 7.5 W.

References

Li, L., Mu, X., Li, S. and Peng, H., A review of face recognition technology. IEEE access, 2020, 8, pp.139110-139120.

Andrejevic, M. and Selwyn, N., Facial recognition technology in schools: Critical questions and concerns. Learning. Media and Technology, 2020, 45(2), pp.115-128.

Galligan, C., Rosenfeld, H., Kleinman, M. and Parthasarathy, S., Cameras in the classroom: Facial recognition technology in schools. University of Michigan, 2020.

Ollivier, S., Li, S., Tang, Y., Cahoon, S., Caginalp, R., Chaudhuri, C., Zhou, P., Tang, X., Hu, J. and Jones, A.K., Sustainable AI Processing at the Edge. IEEE Micro, 2022, 43(1), pp.19-28.

Adnan, S., Ali, F. and Abdulmunem, A.A., 2020, November. Facial feature extraction for face recognition. In Journal of Physics: Conference Series, Vol. 1664, No. 1, p. 012050, IOP Publishing.

Ushiroyama, A., Watanabe, M., Watanabe, N. and Nagoya, A., 2022, January. Convolutional neural network implementations using Vitis AI. In 2022 IEEE 12th Annual Computing and Communication Workshop and Conference (CCWC), pp. 0365-0371.

https://docs.xilinx.com/r/en-US/ug1144-petalinux-tools-reference-guide.