A device and method for automatically recreating a content preserving and compression efficient lecture video
A device and method for automatically recreating a content preserving and compression efficient lecture video is provided. The device comprises of a computer based video recreating means (3) connected to a video receiving means (2) to an input side thereof and to a video reproducing means (4) at the output side thereof, wherein the video recreating means (3) is designed to split the video into visual and audio data, split the visual data into a plurality of scenes, classify each scene into a number of activity scenes, select activity scenes pre-determined to be critical for preserving the semantics of the lecture video and determine a key frame thereof, recreate the visual data by effecting a time based merger of the key frames of the selected activity scenes, recreate the audio data by removing voiceless data and noise therein and recreate the lecture video by effecting a synchronized time based merger of the recreated visual and audio data.
The video is shot detected and recognized for content and non-content shots, the content frames followed full a clear full-page writing, tracked for current pen-position while writing, good quality key frame detected, skew eliminated, frame is super-resolved (if clarity demands), played in synchronization with audio, key-hole animated at client site. The method is fully automated! Data size is massively compressed due to non-sending of video (at least a factor of 10-15), streaming is possible. 3 patents already filed, and product is copy-protected. Most of the media size is now due to audio only (about 90-95%) due to choice of mp3 player. Can be further reduced by selecting more efficient audio-codec.
To know more visit the link below:www.ee.iitb.ac.in/~lec-to-mobile
- Subashis Chaudhuri, Ranjith Ram A, Trupti N Thete