Authors : R. Krishnamoorthy and M. Braveen
Abstract: In this study, a new optimal feature selection scheme with orthogonal polynomials and Ant Colony Optimization (ACO) for Content-Based Video Retrieval System (CBVRS) is proposed. Initially, the video file is divided in to smaller number of chunks as shots in orthogonal polynomials transform domain. In order to identify the key frames to represent a shot, each video image inside a shot is then applied with same orthogonal polynomials to yield Direct Coefficients (DC) images. In this research, the DC image which has the maximum DC value is modeled to be a key frame. From the identified key frames, low level feature such as color, edge and texture information are extracted in the same orthogonal polynomials domain. Since, the extracted features are larger in size, ACO scheme is adopted to select optimal features that represent a key frame for content-based video retrieval system.
R. Krishnamoorthy and M. Braveen, 2017. A New Optimal Feature Selection Scheme with Orthogonal Polynomials and Ant Colony Optimization for Content Based Video Retrieval System. Asian Journal of Information Technology, 16: 383-391.