Face Features Extraction Techniques for Face Recognition of Low Quality Face Images: A Review

Abstract

The impact of smart technologies in real life applications spread over, not limited to, in areas like security and surveillance, biometric systems authenticating user transactions, a person authentication for various licensing, banking, Aadhar card activities etc.The Close Circuit Television (CCTV) cameras are one of the very useful component in these applications. All these applications need to identify a person rightly. Hence it is required to have a robust system for such identification. There are different traits for person identification but the most impactful is the face of a person. Numerous research works are done design and development of Face recognition system based on traditional methods to recent machine learning techniques. In all of these methods the extraction of observational and numerical parameters that help in identifying a person is the most impacting factor on recognition performance. In this paper we have discussed various techniques to extract such facial features, specifically related to Low Resolution (LR), Low Quality (LQ) images. We have also discussed the performance, limitation and the scope of the work carried out by different researcher. Our discussion is focused on use of Gabor based techniques for feature extraction.