A human-based technique for measuring video data similarity

The increasing use of multimedia streams nowadays necessitates the development of efficient and effective methodologies for manipulating databases in storing them. Moreover, content-based access to multimedia databases requires in its retrieval stage to effectively asses the similarity of video data. This work proposes a new technique for measuring video data similarity that attempts to model some of the factors that reflect human notion in evaluating video data similarity. This model presents one step towards designing intelligent content-based video retrieval systems capable of measuring the similarity among video clips in a way similar to what humans do. The performance of the proposed model was tested where the system yielded very satisfactory values of recall and precision under various testing scenarios.