A TRANSFER LEARNING MODEL FOR IDENTIFIER-BASED SERVICES

Abstract

Identity-based services are becoming more and more popular and bring
many benefits to users. Particularly, automatic identification helps bring highclass service experiences to beneficiaries in many fields such as education, resort
travel, health care, customer care. Many models and methods have been
proposed to solve the problem of user identification, in which face image-based
techniques are widely used due to many advantages in terms of data collection
ability, personalization. However, an identification system with high accuracy
and real-time speed is still the goal of many studies in recent times. In this paper,
we introduce a transfer learning based method that combines CNN and SVM
models for the face identification problem. A CNN architecture is proposed and
used as a feature extractor and then, the SVM model for object classification. The
obtained results show a significant improvement in the accuracy of the image
classification as well as the training time.