An Effective Scheme for Shot Boundary Detection and Key Frame Extraction for Video Retrieval
Gs Naveen Kumar1, Vsk Reddy2
1GS NAVEEN KUMAR:Research Scholar, Faculty of Engineering, Lincoln University College, Malaysia
2VSK REDDY: Professor, Faculty of Engineering, Lincoln University College, Malaysia.
Manuscript received on November 11, 2019. | Revised Manuscript received on November 20 2019. | Manuscript published on 30 November, 2019. | PP: 11061-11065 | Volume-8 Issue-4, November 2019. | Retrieval Number: D5436118419/2019©BEIESP | DOI: 10.35940/ijrte.D5436.118419
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The rapid development of devices for image capture and information sharing has resulted in the availability of huge amounts of online video for various applications such as education, news, entertainment, etc. This leads to problems and difficulties when users query any content-related video. The reason for this scenario is that the presently available techniques of content representation and retrieval are based primarily on annotation. It therefore provides insufficient information for understanding and retrieving the content to match the query of the user. Content Based Video Retrieval (CBVR) is one of the promising new ways for finding content in a large video archive, rather than simply searching terms. The primary steps for indexing, summarizing and retrieving video are shot transition recognition and representative frame extraction. We have proposed a key point matching algorithm for a superior and robust Scale Invariant Feature Transform (SIFT) followed by the collection of representative frames from each segmented shot using the Image Information Entropy method. By using the Rough Set Theory, we can get better the concert of this scheme through removing unnecessary representative frames. All the methods suggested to prove the efficacy were tested on TRECVID datasets and contrasted with state-of – the-art approaches.
Keywords: Shot Transition, Representative frames, Image Entropy, SIFT, Rough Set Theory, Retrieval.
Scope of the Article: Knowledge Representation and Retrievals.