Identifying Appropriate Features for Recognizing Multi-View Video Duplication with the aid of an Adaptive Crow Search Algorithm

Abstract

Multi-view video orientation is a stimulating problem that has been handled for diverse circumstances. The significant purpose of incorporating optimization techniques is to identify appropriate feature extraction among 113-features extracted (DWT-12, shape-1, color-1, and texture-99). The computational process of identifying appropriate feature extraction techniques through manual or trial-and-error process consumes more time. To resolve, the research includes optimization techniques namely Adaptive Crow Search Algorithm (ACSA), Crow Search Algorithm (CSA), Cat Swarm Optimization (CSO) and Opposition based Whale Optimization Algorithm (OWOA).  The results show that the proposed ACSA unveils accuracy of 93.75% in both databases, which is proficient over comparative techniques. It is obvious that the adaptive nature of CSA updating strategy in associate with seeking and tracking strategy enhances the performance effectively.