An approach for automated video indexing and video search in large lecture video archives

E-Learning is the use of educational technology, communication and information technologies and electronic media in education. E learning contains various types of media including images, video, audio, streaming videos, animation, web based learning, video based learning, audio based learning, E books etc. Distance learning can be done without school or collages, anyone can learn from their home or office. E-Learning industry is economically remarkable and it was work out in 2000 to be above 50$ billion corresponding to traditionalist estimates. Lecture Audio, video data on internet is growing rapidly. Hence there is immediate need for method by which we can retrieve audio, videos on internet. In this paper we have presented a technology for video search in lecture video archive. Initially, we can introduce segmentation of videos and key frame detection for offering rules for navigation of video contents. By applying ASR (Automatic Speech Recognition) on lecture audio and OCR (Optical Character Recognition) on video content we can extract metadata. OCR can be used in Data entry for business document, Automatic Number plate Recognition, Extracting business card information into a contact list and so on. ICR (Intelligent character recognition) focuses on handwritten documents as well as cursive character one at a time usually it involves in Machine Learning. Speech recognition system can classify into continuous or discrete system which can be speaker independent, speaker dependent or adaptive. Discrete system focuses on a separate acoustic model for each single word, sentence, phrase etc. are said to be isolated word speech recognition (ISR). CSR (Continuous Speech Recognition) System focuses on user who speaks sentences continually.