الفهرس Only 14 pages are availabe for public view
 |  | from | 106 |  |  |
| | | from | 106 | | |
Abstract
A video sequence, often visualized as a volume of data, can be conceived as a sequence of still images or frames. Video processing techniques tend to predict frames, either forward or backward, with different motion estimation techniques then subtract the predicted frame from the original one to get the error which is processed and sent to the receiver where same prediction technique is applied and the erroF is corrected with the received error signal. In the mean time insignificant information is removed while processing video sequence in order.to compress the size as maximum as possible. It is clear that the insignificance here is measured against the desired quality after signal reconstruction.
In some cases, like master films or medical imaging, losses in reproduced video are not acceptable at all and here it comes the need for having a lossless video compression technique that can achieve high compression ratio while maintaining the original video quality.
This work proposes a lossless video coding scheme. It is a hybrid technique with several processing stages, specially designed for video streams and based on information theory. Not like other lossless compression algorithms designed for mainly text compression and consider sequential single dimension data streams, it considers video stream special characteristics. Experimental results showed that the proposed algorithm succeeded to be less than best performer of the test lossless compression algorithms with more than 40%.