Static scheduling with interruption costs for computer vision applications

Abstract

It is very difficult to find pre-emptive scheduling algorithms that consider all the main characteristics of computer vision systems. Moreover, there is no generic algorithm that considers interruption costs for such systems. Taking the interruption of tasks into account scheduling results can be improved. But it is also very important to take the costs that arise from interruptions into account because they not only increase the total execution time, but also because the scheduler can evaluate whether it is adequate to interrupt certain tasks or not. Thus, the result can be more realistic. Therefore, we present an extension to the static algorithm SASEPA for computer vision which considers interruption costs.