Content Based Video Retrieval System

Abstract

Content based video retrieval is an approach for facilitating the searching and browsing of video collections based on content i.e. text, image. It includes methods such as Indexing, Searching and Retrieval, and Ranking of video results. Traditional video retrieval methods fail to meet technical challenges due to large and rapid growth of multimedia data, demanding effective retrieval systems. Therefore, a more efficient method for video retrieval within large video archives is needed. The existing systems relay mainly on textual search. Among them, content-based video retrieval is a challenging and important problem of practical value. It can help users to retrieve desired video from video dataset efficiently based on the video contents. Content Based Video Retrieval System is a desktop application that is developed to retrieve videos using multimodal query i.e. user can search the video dataset using textual or visual query. Extracted Features from the uploaded query are matched with stored content in video index (divided in textual, visual) and then relevant videos are retrieved. Retrieved Videos for monomodal search are ranked whereas for multimodal search they are merged and then ranked. Ranked Results are displayed to the user.