Deep Learning-based Face Identification System

JIITA, vol.3 no.4, pp.315-317, 2019


Deep Learning-based Face Identification System

Juangtak Ryu 1), Jong Kwan Lee 2,*)
1) School of Electronic and Electrical Engineering, Daegu University, Gyeongsan, Republic of Korea
2) Department of Computer Science, Bowling Green State University, Ohio, U.S.A.

Abstract: In this extended abstract paper, a practical, deep learning-based system for automated face identification is presented. The system is based on the deep hidden identity features which exploits GPU-based deep learning processing. In particular, after employing a deep-learning based image aligned registration and feature extraction, the feature vectors are compared using the cosine similarity measure to identify different faces. The experimental results show that the system is very efficient and accurate.

Keyword: face identification; deep learning; feature extraction; image registration