UPI-RJAIML-2018-2 (Abstract)

Case Study (Funded Research)

Content Based Video Judgments - A Comparative Case Study   

Rashed Mustafa*, Dingju Zhu

1Department of Computer Science and Engineering, University of Chittagong, Bangladesh.

2School of Computer Science, South China Normal University, Guangzhou, China.

Abstract

Searching desired videos is now challenging due to scattered collection of huge amount of videos all over the world. It is a tedious task to find different video applications for different video contents required by different users from huge repositories. Usually, the selection is based on user’s choice or expert recommends, however different users have different options and distinguished experts have different recommendations and only few judgments are fortunately appropriate and when the judgments are erroneous, the video content required by different users will not be fully satisfied by the selected video applications, which will diminishing the video service quality. The contribution of this paper is to prioritize videos by using a multi-criteria decision analysis tool ‘Analytical Hierarchical Process (AHP)’ for selecting, ranking and judging particular video from semantically classified videos and comparing performance of the system with other Multi Criteria Decision Analysis (MCDA) approach and Evidential Reasoning Method (ER). In this study a simple case is presented over some classification of videos where particular object/person needed to be identified. It is shown that both approach prioritized shot-stock videos for selecting particular object/person but ER approach is better than AHP (it was found that the ER approach is better than the AHP).

Key words: ER, AHP, MCDA, Eigenvector, Pair wise matrix.

Full Article PDF