Image pyramid segmentation using parallel machines
This thesis describes two algorithms: parallel fuzzy pyramid linking (FPL) image segmentation for shared memory systems and parallel FPL image segmentation for distributed memory systems. The algorithms are designed for the purpose of increasing the speed of FPL which is a specific approach to pyramid image segmentation. Both algorithms divide the original image among the available processors, and each of the processors performs FPL on each sub-image. The main disadvantage in processing time for shared memory systems is scheduling the memory usage. The main disadvantage in speed for distributed memory systems is inability of processors to share information. Both algorithms were evaluated on the process speedup. The results show that the algorithm for the shared memory systems proved to be more efficient.
Thesis92.S272.pdf
8.47 MB
Adobe PDF
41e8f91535d9a1a250637d210f2144e9