Authors: Amit Phadikar; Nishant Kumar; Baisakhi Sur Phadikar; Goutam Kumar Maity
Addresses: Department of Information Technology, MCKV Institute of Engineering, Liluah, Howrah-711204, West Bengal, India ' Department of Computer Science, Techxell School of Management and Technology, Jaina More, Bokaro-829301, Jharkhand, India ' Cognizant Technology Solutions, Salt Lake, Kolkata, India ' Department of Electronics and Communication Engineering, MCKV Institute of Engineering, Liluha, Howrah, WB, India
Abstract: Most of the methods for video summarisation rely on complicated clustering algorithms that make them too computationally complex for real time applications. This paper presents an efficient approach for video summary generation that does not rely on complex clustering algorithms and does not require frame length as a parameter. The present scheme combines colour histogram and edge histogram features with optimum global thresholding to detect key frames. The optimum threshold is selected based on genetic algorithm (GA) to increase the performance of the proposed system. For each shot, key frames are extracted and similar key frames are eliminated. Experimental results duly support those claims.
Keywords: colour histograms; edge histograms; genetic algorithms; video summarisation; optimum global threshold; video summary generation; key frames.
International Journal of Innovative Computing and Applications, 2016 Vol.7 No.1, pp.1 - 12
Received: 24 Feb 2015
Accepted: 09 Jul 2015
Published online: 23 Mar 2016 *