FACE RECOGNITION USING INFRARED IMAGING

Abstract

Face recognition is an area with enormous practical potential, including an extensive span of business and regulation implementation practices. It is amongst the utmost dynamic analysis of fields of machine vision. It continues to advance even after over three decades of intense exploration, profiting from advances in different research fields such as computer graphics, image processing, pattern recognition, and physiology. Visible spectrum pictures are the maximum researched face recognition modality, and structures based on this modality have reached a substantial level of maturity with some real-world success. However, visible spectrum images face tests in the existence of radiance, posture and facial changes, as well as facial disguises; all of these issues can decline recognition accuracy.

Numerous methods which have been projected to overcome these restrictions; however, the use of infrared (IR) imaging has developed as a particularly capable research direction. This report presents a complete and timely review of the works on face recognition using infrared imaging. The crucial contributions present are (i) a summary of the intrinsic properties of infrared imaging which makes this modality hopeful in the framework of face recognition; (ii) systematic analysis of the most important methods, along with a focus on evolving common trends as well as critical differences between other practices. The statistical description of the face varies drastically with changes in pose, illumination and expression. These variations make face recognition (FR) even more challenging.

vi In this report, two novel techniques are proposed, viz., Dual Objective Feature Selection to learn and select only discriminant features and Scaled Euclidean Classification to exploit within-class information for smarter matching. The 1-D discrete cosine transform (DCT) is used for efficient feature extraction. A complete FR system for enhanced recognition performance is presented. Experimental results on the face database Surveillance cameras face database (SCface) illustrate the promising performance of the proposed techniques for face recognition.