摘要
小波变换可以有效的用于图像的特征提取 ,但二元小波变换的计算量随其阶数的增加呈指数级增长。本文提出小波变换并行算法 ,并在银河系列并行计算机上进行了图像特征提取的实验 ,结果表明计算时间明显减少。
Wavelet transform can be used to image feature extracting. But the amount of computing will increase with exponent level when using two-variable wavelet transform . In this paper we describe a parallel wavelet transform algorithm. We use this algorithm to process a 512× 512 image feature extracting on parallel-computer YH series. In the end, the results are given, which show that the way that we describe can greatly improve the speed-up ratio.
出处
《微机发展》
2000年第4期38-39,共2页
Microcomputer Development
基金
国家 8 63高技术资助项目!(863 - 391 -ZT0 3 - 0 6 - 0 3)
