Translator Disclaimer
17 May 1989 Macropipelined Multicomputer Systems For Image Analysis
Author Affiliations +
Proceedings Volume 1058, High Speed Computing II; (1989)
Event: OE/LASE '89, 1989, Los Angeles, CA, United States
We present a scheme for macropipelining in multicomputer systems to achieve high speeds in processing multiple images. Most image processing applications consist of a sequence of tasks - - e.g., preprocessing, detection, segmentation, feature extraction, and classification. This sequence lends itself to a pipelining strategy. To minimize the effects of bottlenecks in this pipeline, we introduce a performance model for data partitioning which includes both the computation and the communication aspects of parallel processing. With the help of this model, we assign the appropriate number of processors to each task so that the workloads are well-balanced. Then we generate a problem graph describing the relationships among tasks and subtasks. We use an estimator of the frame processing time of the image processing system as an objective function for choosing a mapping of the problem graph to a system graph. This estimator takes account of computation times and communication intensities among the subtasks in the problem graph, and accounts for link contentions. To find an efficient mapping, we use a heuristic optimization technique in which possible bottlenecks are given high priority in the mapping procedure. We tested our macropipelining scheme on a typical image processing application in a simulated hypercube computer system. The results support our belief that this scheme yields effective architectures for high-speed processing of long sequences of images.
© (1989) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Youngshik Moon, Nader Bagherzadeh, and Jack Sklansky "Macropipelined Multicomputer Systems For Image Analysis", Proc. SPIE 1058, High Speed Computing II, (17 May 1989);

Back to Top