Parallel Implementation & Analysis of Low-Level Vision Algorithms.
Abstract
This in-house effort developed and implemented low level image processing routines in parallel on any array of transputers. A selected set of low-level image processing algorithms was implemented onto a single transputer, then onto multiple transputers in two (2) different network topologies, a straight pipe configuration and a ring configuration, and implemented on multiple numbers of transputers to determine the efficiency of hardware utilization. Performance analysis includes achieved speed-up, ease of implementation, efficiency, as well as advantages and disadvantages of parallel implementation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1995
- Accession Number
- ADA307309
Entities
People
- Lee A. Uvanni
Organizations
- Rome Laboratory