Authors : A. Packialatha and A. Chandrasekar
Abstract: In recent past, videos form a large part of data in the internet. Due to its ability to attract audience of all age groups, a large group of websites dedicated to browsing and watching videos have mushroomed. But regrettably, search engines do not give sufficient importance to retrieve videos. In all the video-oriented websites, the search is based on keywords typed in which produces unrelated output in terms of relevancy. To achieve content-based video retrieval, we suggest PSPF (Pencil Screened Pattern Filter). In the proposed system, the background key frame and object key frame of each and every shot of the video is obtained. Then the morphological shape of the background and objects are acquired and represented as a matrix of 0 and 1 based on grid encoding. The eigen values of these grid encoded matrices are used as metadata for indexing and retrieving videos. An index table is maintained based on the eigen values of the background key frame matrices. The index table is organized in such a way that the background key frame matrices of videos with related eigen values divide the adjacent field in the table. The eigen values of objects in the video are associated with the background key frame in the index table by means of linear chaining. Thus, we have effectively utilized the squat level characteristics of video to achieve ontological recovery of videos.
A. Packialatha and A. Chandrasekar, 2017. Effective Video Retrieval Using Pencil Screen Pattern Filter and Eigen based Indexing. Asian Journal of Information Technology, 16: 486-492.