Title: Video segmentation using Metropolis Hastings Algorithm for the VCR operations

Authors: R. Ashok Kumar, K. Ganesan

Addresses: School of Information Technology and Engineering, VIT University, Vellore 632 014, Tamil Nadu, India. ' School of Information Technology and Engineering, VIT University, Vellore 632 014, Tamil Nadu, India

Abstract: In this paper, we have investigated the video pre-processing and the VCR operations for the Video on Demand (VoD) system. We have proposed a novel technique for the segmentation of the movies using Markov Chain Monte Carlo (MCMC). Our proposed segmentation technique will segment the videos by identifying the scene boundary locations. The objective of this paper is to reduce the accessing time for the new requests of the movies using M-chaining technique and also to reduce the seeking time for the VCR operations using our segmentation technique. We have numerically analysed the existing segmentation techniques with our proposed segmentation technique. We found that our segmentation technique shows the less computational time compared with other segmentation techniques. The performance of the segmentation techniques is measured in terms of Precision as well as Recall and our technique gives more accurate results compared with other segmentation techniques. The VCR functionalities are simulated by applying our segmentation technique and the results of the simulation show less accessing and seeking time. We have also addressed the practical issues of efficient utilisation of overall bandwidth and buffer of the VoD system using our segmentation technique.

Keywords: chaining; frames; MCMC; Markov chain Monte Carlo; segmentation; scenes; shots; VoD; video on demand; video pre-processing; VCR operations; scene boundary locations; simulation.

DOI: 10.1504/IJAMC.2010.034661

International Journal of Advanced Media and Communication, 2010 Vol.4 No.3, pp.274 - 297

Received: 11 Feb 2010
Accepted: 06 Jul 2010

Published online: 14 Aug 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article