Analysis of video segmentation for spatial resolution reduction video transcoding

In this paper different methods of video segmentation are analyzed to perform spatial resolution reduction video transcoding with multiple processing units. A distributed video transcoder is built in which different processing units perform the transcoding operation. To fully utilize the computational power of different processing units the distribution of computational load should be equal. In video transcoding different frames require different computational power hence inefficient video segmentation will lead towards lower performance. We have analyzed three possible methods of video segmentation: (1) each segment has equal size, (2) each segment has equal number of frames, and (3) each segment has equal number of group of pictures. The performance of the system, the relationship between the processing units used and speed in computation is measured in terms of standard deviation of transcoding time of different processing units.