Feature Extraction from Video for Cricket Highlight Generation
Amruta D. Aphale1, P. M. Kamde2
1Amruta D. Aphale, Department of Computer Engineering, SPP University, SCOE, Pune, Maharashtra, India.
2Prof. P. M. Kamde, Department of Computer Engineering, SPP University, SCOE, Pune, Maharashtra, India.
Manuscript Received on July 01, 2015. | Revised Manuscript Received on July 02, 2015. | Manuscript published on July 20, 2015. | PP: 42-46 | Volume-3 Issue-8, July 2015. | Retrieval Number: H0652073815/2015©BEIESP
Open Access | Ethics and Policies | Cite
© 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 most rapidly increasing component in various sectors is Internet Technology. Where the information is being searched based on images, texts, videos. There exists various methods to extract the required information from the raw data which is in the form of text and images. There are multiple information engines where a detailed information could be searched, one of popularly being used is Goggle. However those uses text based retrieval techniques. Being a critical aspect of Information technology Video has become most synergistic channel of communication in day to day life. The steep volume of video makes it enormously hard to browse through and get the interned information. Its difficult to search a video without knowing the content. Performing manual analysis on the contents and then indexing the same is pretty time consuming task. The apparent alternative is to detect such events in the video automatically. The initial step in automating the system is event detection which breaks the massive volume of video into smaller chunks called shots. Our work aims in identifying such events. Although attempts have been made to detect shot boundaries having smooth transitions in between the results are not as successful as for detecting shots separated by hard cuts. Performing a detailed analysis on Video database is most complex task as it involves number of variables and having the analysis done on larger number of such requires larger amount of memory with huge computation power. A video database describes what actually happens in a video and its perception by a human which is termed as Semantic Information. These days we have number of national and international broadcasting news, sports channels, which continuously broadcasts the sport events happening around the globe. There are many of them who have got a special devoted segment for sports. Even having these facilities one cannot remain stuck to watch the complete event due to certain time constraints. With this as an encouragement to find a technique that could provide desired results, this report discusses various algorithms and sketches out the main features that have been so far used for event detection. Here an systematic approach has been attempted to extract prominent features and events in Cricket sport videos. This system also classifies every event sequence into a concept by sequential association mining.
Keywords: Browsing, event detection, multimedia, retrieval, semantic gap, video database.