Review of hybrid face-based recognition system

• Main Authors: Hassan, Heba Jabbar, Wali, Mousa K., Shujaa, Mohamed Ibrahim
• Format: Conference Proceeding
• Language: English
• Subjects: Algorithms; Barriers; Effectiveness; Explainable artificial intelligence; Face recognition; Facial recognition technology; Machine learning; Parameter identification; State-of-the-art reviews
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0236494

In recent years, face recognition algorithms have made great strides thanks to the rapid development of deep learning techniques and the availability of large-scale face datasets. However, despite these developments, there are still obstacles to reaching high accuracy and durability in facial recognition systems in real-world applications. Face recognition systems that combine many modalities or algorithms are gaining popularity as a means of overcoming these challenges. The goal of this research was to compile an overview of hybrid face-based recognition systems, detailing state-of-the-art methods and their effectiveness in improving face recognition performance. First, the reason of double-mode techniques are preferable is explained, and what benefits they have over single-mode ones. This extensive analysis emphasizes the important parameters that affect the efficacy of several hybrid face-based identification systems and indicate their strengths and limits. The obstacles and potential future research topics, such as standardized evaluation methodologies and the incorporation of explainable artificial intelligence (AI) methods are discussed. Insights and suggestions for the future development of more accurate and trustworthy solutions are provided in this review, making it a great resource for researchers and practitioners working on face recognition systems. The focus of future research should be on further improving hybrid approaches and exploring new directions.