Intelligent classification techniques for face recognition using deep learning algorithms from hyper spectral images

Abstract

Face recognition is an important task in the recent years due to the introduction of automation in many real-life applications. In the examination system, face recognition is necessary to identify the genuineness of candidates. In airport application, the persons with fake passport can be identified by the use of FR systems. It has many other applications such as military, home security, FR based hospital information systems, banking and reservation systems. With regard to gain a larger size of the facial datasets, it is necessary to explore the use of Hyper Spectral Imaging (HSI) techniques on facial datasets. The hyper spectral imaging techniques provide improved face recognition accuracy by acquiring added biometric features like spectral features. However, as the number of faces to be tested rises, the efficacy of 2-Dimensional image-based methods decreases because of the shrinkage in the inter-object space in the facial recognition domain. In this condition, the hyper spectral imaging technique must be used to improve the efficacy owing to the vast size of features. Hence, it is necessary to apply Artificial Intelligence (AI) algorithms such as Machine Learning (ML) with HSI to perform the task of face recognition more accurately. newlineRecently, the Convolutional Neural Network (CNN), Recursive Neural Networks (RNN) and vision transforms are identified as the appropriate techniques for performing the computer vision including the face recognition, image and video analytics. The different types of tasks like image classification, objects detection process, and face recognition tasks that are aided from the CNN. newline