Towards user transparent data and task parallel image and video processing: An overview of the Parallel-Horus project

In the research area of image and video processing, the Horus library [5] has become a valuable system for software driven rapid prototyping, and an essential vehicle for knowledge transfer at the level of universities and expertise centers. Due to its strictly sequential implementation, however, Horus can not always satisfy all demands for high performance. As computational requirements for the processing of large image sets and video streams continue to increase, it is essential to provide a Horus implementation that applies to high performance computers. As researchers in image and video processing can not be expected to also become experts in parallel computing, the Parallel-Horus project aims to shield Horus users from all intrinsic complexities of parallelization. This paper presents an overview of the Parallel-Horus project; it discusses the project's goals, as well as current and future research directions. Also, the efficiency of the current Parallel-Horus implementation is demonstrated by evaluating a strictly sequential state-of-the-art imaging application.