Parallel algorithms for all nearest neighbors of binary images on the BSP model

Proceedings Fourth International Symposium on Parallel Architectures, Algorithms, and Networks (I-SPAN'99) Pub Date : 1999-06-23 DOI:10.1109/ISPAN.1999.778970

Takashi Ishimizu, A. Fujiwara, M. Inoue, T. Masuzawa, H. Fujiwara

引用次数: 0

Abstract

We present two parallel algorithms for computing the nearest neighbors of an n/spl times/n binary image on the Bulk-Synchronous Parallel (BSP) model. The first algorithm is for weighted distance, and the second algorithm is for L/sub p/ distance. Both algorithms run in O(n/sup 2//p+L) computation time and O(g/sup n///spl radic/p+L) communication time using p (1/spl les/p/spl les/n) processors and in O(n/sup 2//p+(d+L)log p/n/log(d+1)) computation time and in O(gn//spl radic/p+(gd+L)log p/n/log(d+1)) communication time using p (n

查看原文本刊更多论文

基于BSP模型的二值图像的所有近邻并行算法

提出了两种基于批量同步并行(BSP)模型的n/spl次/n二值图像的最近邻计算算法。第一种算法是加权距离，第二种算法是L/下标p/距离。对于任意整数d (1/spl les/d/spl les/p/n)，这两种算法都在O(n/sup 2//p+L)计算时间和O(g/sup n///spl径向/p+L)通信时间内运行，使用p (1/spl les/p/spl les/n)处理器，在O(n/sup 2//p+(d+L)log p/n/log(d+1))计算时间和O(gn//spl径向/p+(gd+L)log p/n/log(d+1))通信时间使用p (n

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊