Parallel image analysis on recursive Torus architecture

The authors describe a new parallel image understanding machine RTA/1 design based on the recursive Torus architecture, and proposed a data level parallel processing scheme using parallel data structures. Various types of parallel image analysis algorithms were developed to demonstrate the effectiveness of both RTA/1 and the parallel data structures. The performance of all algorithms were evaluated by simulation and it was shown that about two order of magnitude speed-up can be attained by using 32 × 32 PE tours. The five types of parallel data structures described here are only fundamental ones and one needs to explore new distributed data representations and coordinated communication procedures to increase the flexibility, utility, and efficiency of the parallel processing on RTA/1. The Pascal like notation used to describe the parallel data structures and algorithms is to be improved to a complete parallel programming language. However the development of the parallel programming language should follow the investigation of the control level parallel processing on RTA/1, since parallel image understanding systems require both data and control level parallel processing capabilities