Year: 2025
Type: article
Abstract: : The review paper traces the development of face recognition from initial approaches to current deep learning techniques (Hasan et al., 2021; Sáez-Trigueros et al., 2018). The earliest methods employed distinct features using SIFT and LBP (Balaban, 2015), but they lacked adequate solutions for complex situations. Statistical subspace methods improved the representation of faces when implemented. Deep Face systems initiated a transformative change in the field by achieving human-level performance through the extensive use of diverse datasets (Taigman et al., 2014). Research continues to develop robust face recognition systems that address ethical concerns regarding bias and privacy, as well as fairness, to create more acceptable solutions. Keywords: Face recognition; Illumination; partial occlusion; pose invariance.
DOI: 10.55041/ijsrem45987
Author: Yasha Sharma
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4409837889