Implementing OpenVX on a polymorphous array processor

This paper presents an implementation of OpenVX on the PAAG array processor (Polymorphic Array Architecture for Graphics and image processing), which is a novel polymorphous architecture designed for graphics rendering and image processing. OpenVX is a new standard for implementing and optimizing low-level image processing and computer vision released by Khronos. It uses a graph-based execution model to combine function enabling both task and data-independent execution. We propose two OpenVX graph mapping modes for PAAG: Whole Graph Mapping (WGM) and Single Node Mapping (SNM). Experimental results indicate that the two modes are both effective, but WGM is better that SNM in terms of efficiency, the experiments also illustrates PAAG has a great potential on image processing and computer vision on array processor.