摘要
本文在介绍汉字识别PCLI特征的基础上,结合并行处理技术,设计了PCLI特征提取的VLSI并行算法.算法用了N个处理单元(PEs),其时间复杂度由相应串行算法的O(N^2)降到O(N).这种处理方法将有利于汉字识别速度的提高和其它性能的改进.
In this paper, based on an introduction to classification feactures——P.C.L.I. features for Chinese character Recognition, we design a parallel VLSI algorithm and its architecture for P.C.L.I features extraction. In this algorithm, we use N array processors. The time complexity of the parallel algorithm is O(N) while its serial algorithm's time complexity is O(N^2). The idea of parallel processing will improve recognition speed and other recognition performance.
出处
《计算机研究与发展》
EI
CSCD
北大核心
1992年第11期35-41,53,共8页
Journal of Computer Research and Development
关键词
汉字识别
VLSI
并行算法
特征
Chinese character recognition
VLSI parallel algorithm
systolic array
