Optimal image algorithms on an orthogonally-connected memory-basedarchitecture

摘要

] while providingoptimal speedup for several problems in image analysis and vision. Suchproblems include labeling connected regions, computing minimum convexcontainers of regions, and computing nearest neighbors of pixels andregions. Optimal algorithms are presented for histogramming andcomputing the Hough transform. Such problems arise in medium-levelvision and require global operations or dense data movement. It is shownthat for these types of problems, the proposed organization is superiorto the mesh and pyramid organizations