Real-time face recognition system with dynamic training and enhanced multi-algorithm face recognition

Abstract

In today's world, the scope and applications of face recognition systems either as standalone or as an integrated module into a larger system cannot be belittled. Remarkable face recognition algorithms have been proposed in last decade offering towering levels of accuracy and precision. However, the flexibility and robustness of the existing systems remain highly questionable. To further intricate these problems, most of the face recognitions systems are designed under strict background constraints, conveniently ignoring variations in illumination and at the apex of the predicament, lies the intricacy involved in the creation and maintenance of training sets in the databases. Even the most sophisticated systems designed posed grave difficulties in deployment and required maintenance by skilled personnel, leading to poor response from the global market. With an objective to realize an intelligent, maintenance-free system with impeccable design, our research team developed an immaculate real-time face recognition system with dynamic training and enhanced multi-algorithm face recognition. The system would facilitate user-friendly, dynamic creation of training sets by means of an innovative approach which would make the need of skilled maintenance personnel obsolete. The proposed system has been comprehensively tested to achieve remarkable precision and accuracy of 99.9%. Proposed system archetype design would result in manifold applications in developing simplified face recognition systems (for enrolling masses) for institutions, business organizations, research labs, military applications, etc. for authentication and authorization purposes.