Color Content Based Video Retrieval Using Discrete Cosine Transform Applied On Rows and Columns of Video Frames with RGB Color Space
Sudeep D. Thepade, Ajay A. Narvekar, Ameya V. Nawale · 2013
The volume of video data is increasing because of advanced technology in imaging communication. This tends to increase the popularity of the videos. The conventional textual description based video retrieval which is based on text entered by user, so, suffers from user perception. Hence, content based video retrieval is more effective than text based retrieval, as it gives result closer to semantic view. Here, a novel content based video retrieval based on discrete cosine transform (DCT) applied on rows and columns of video frames to extract features as selected high energy is proposed with the help of RGB colour space. Video database of 500 video samples spread across 10 categories is considered as a test bed for experimentation. Performance comparison is done with the help of cross-over point of average precision and average recall values of 500 query videos fired on the whole database. The experimental results have shown that the video retrieval with DCT applied on columns gives better performance.