A Parallel Implementation to Schedule a Video Sequence by a Parallel Genetic Evolution Algorithm Approach

The problem of scheduling a set of dependent or independent tasks to be processed in a parallel fashion is one of the most challenging problems in parallel computing. The goal of a scheduler is to assign tasks to available processors such that precedence requirements between tasks are satisfied and the overall length of time required to execute the entire program, the schedule length or make span is minimized. A Parallel Genetic Algorithm Approach has been developed to the problem of task scheduling. GA is competitive in terms of solution quality if it has sufficient resources to perform its search. The Job taken for the Scheduling is the Detection of a Moving Object in a Video Sequence. The Moving Object Segmentation is suitable for real time content-based multimedia communication systems. First a background registration technique is used to construct as reliable background image from the accumulated frame difference information. The moving object region is then separated from the background region by comparing the current frame with the constructed background image. The implementation is optimized using parallel processing and achieved on a personal computer with a 3.0 GHZ Pentium IV Processor. Good segmentation performance is demonstrated by the simulation results.