Robust Face Recognition over Rotation

Abstract

An inspiring announcement made by unique identification authority of India (UIDAI) on 6th June 2025, there are two hundred and eleven crore biometric authenticated transactions are made in May 2025. Biometrics are ubiquitous and face is vital for smart applications because of the nonintrusive virtue and works better in non-cooperative environment. In this paper, we propose a robust biometric algorithm to recognize human faces towards the variation of rotation angles. The input images are refined to extract face through Robert’s edge detection and cropping. It removes the background information, which is redundant. Unique statistical features are extracted through matrix norm, standard deviation, peak to peak filtering and normalized gray-level co-occurrence matrix (NGLCM). The obtained results are simply outperforms over other conventional approaches.