Video Retrieval Using Automatically Extracted Audio

Videos are a powerful and expressive media that can capture and present information. As these videos cover many subjects / genres, it is critical task to segregate videos as per the needs or interest of the users which requires classifications. In this paper we present a model for video classification using embedded audio. Video classification is usually accompanied with video annotation which helps in retrieving the archives. Efficient and effective video classification and annotation demands automated unsupervised classification and annotation of videos based on its embedded video content as manual indexing is unfeasible. As a first step the audio content of video is extracted and cleaned for further processing the next step converts audio into textual format. The text is processed upon to get the prime keywords in the video using text mining. The videos are classified and annotated on the keywords thus found. The annotated videos are stored in an object oriented database for future retrieval.