摘要
以小学语文课本的三千个印刷体汉字作统计材料,每个字划分为四个象限,把在左上角象限、右下角象限的笔划构成的形状分别统计,形状相同或类似的归为一类并称之为子模式。统计结果表明,左上角的子模式组字能力比右下角的强。但左上角子模式的形状较复杂,除构成子模式的部首的笔划外,总带有其它笔划。右下角子模式的形状较简单,除它本身的笔划外,无其它笔划。两者比较,在设计汉字多步自动识别的系统时,以右下角的子模式作初分类的依据,比较有利。
The structure of 3000 printed Chinese characters which were selectedfrom the Chinese textbook for elementary school was analysed. Each characteris divided into four quadrants i. e. upper-left (UL), upper-right (UR), bot-tom-left (BL) and bottom-right (BR).When the shapes composed of the strokesin UL and BR were the same or similar respectively they were attributed toone group and were called subpatterns. The result showed that the subpatterns of BR were more simple than thatof UL and the number of junctions of the former was less than that of thelater. The author argued that the subpatterns of BR are used as group maskin the early stage of the processing of multi-stage matching the computationtime may be reduced.
出处
《心理学报》
1982年第4期385-390,共6页
Acta Psychologica Sinica
