Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Parallel image processing: methodology, modeling, and analysis

Thesis/Dissertation ·
OSTI ID:5215457
The field of image processing undoubtedly requires the computational resources of high performance parallel processing. This dissertation is broadly concerned with issues in the identification and utilization of multiprocessor system organizations for parallel image processing. In this context two main problems are addressed. First is that of formulating parallel solutions to image processing problems. Provided with an image processing task specified in a conventional sequential manner, two algorithms are developed for formulating parallel versions of this task. These algorithms are derived based on a paradigm that is proposed for parallel image processing. The second problem addressed is the matching of such parallel formulations and parallel architectures. It is required to determine an assignment of parallel tasks to the processors of a parallel architecture in a manner that minimizes the resulting interprocessor communication. Bounds on the complexity of intertask communication are derived for processors interconnected by ring networks and a class of switching networks. An algebraic characterization of these networks is developed and is used to quantify various performance metrics and develop schemes for utilizing them.
Research Organization:
Texas Univ., Austin (USA)
OSTI ID:
5215457
Country of Publication:
United States
Language:
English

Similar Records

Dependability evaluation of parallel/distributed computer networks
Thesis/Dissertation · Tue Dec 31 23:00:00 EST 1985 · OSTI ID:6747866

Design methodology for multiprocessor architecture
Thesis/Dissertation · Thu Dec 31 23:00:00 EST 1987 · OSTI ID:5145961

Intelligent allocation algorithm for parallel processing
Technical Report · Thu Dec 31 23:00:00 EST 1987 · OSTI ID:5180906