汉字印刷字符视觉心理尺度评价研究-计算机应用技术专业论文.docxVIP

摘要在设计领域中，许多问题都涉及到人类的视觉心理感受。目前，由于这些问 摘要 在设计领域中，许多问题都涉及到人类的视觉心理感受。目前，由于这些问 题必须要人的介入才能解决，就大大限制了设计的自动化进程。因此如何将这些 感性认识转化为计算机能够处理的定量信息，是一个值得探索的课题。 ≮。’本研究汉字印刷字符视觉心理尺度评价，其目标就是要寻找一种工程的手段 来解决评价汉字直观大小，这一通常被认为纯粹属于感性认识的问题。通过对已 有算法的分析，本文提出了一种新的汉字视觉心理尺度评价方法——矩心特征 法。本方法将汉字的视觉心理尺度这一连续的、无法用物理量直接分析评价的问 题进行离散化，利用模式识别的手法加以研究。根据形状知觉尤其是汉字知觉的 心理法则，抽取汉字字符的黑点数、零次心距、一次心距、二次心距作为评价汉 字字符视觉心理尺度的依据，利用人工神经网络来构造边界，从而将汉字字符分 为“偏大”、“适中”和“偏小”三类。本方法以人类视觉心理本质为出发点，因 此从根本上克服了原有算法的不足。同时，实验也证实了本方法对于汉字印刷字 符视觉心理尺度评价的有效性。本研究的成果对于汉字字符设计具有重要的意 义，此外，也将对类似心理量的工程性处理产生一定的启示作用。 关键字 心理尺度有效特征心距模式识别神经网络 一——一 ～。——～—————一 AB AB STRACT In the field of designing，there are several problems relating to the visual psychology of human beings．The dependence on the judgement of experts will slow down the process of designing．How to solve such problems by computer automatically?That will be a 900d question． This research focuses on the evaluation of printed Chinese characters’visual psychological scale，which means to find a way to evaluate the scale of Chinese characters by computer．After discussing some algorithms，we come up with our new method———Cemml Distances as Valid Features．The visual psychological scale of Chinese characters is continual，which can not be analysed with physical quantity directly．For this reason，the new method is to disperse the problem and study it by means ofpattern recognition．During the course we extracted some features which can reflect the visual psychological scale of Chinese characters，such as sum of black points，central distance degree 0 and SO on．And then,boundaries were formed by a neural network．So that we can divide Chinese characters into three classes——bigger proper and smaller．Our research based on the essence of visual psychology,which entirely got over the lack of the existing algorithms Further more，the result of experiments has also proved the validity of the new method．The fruit of this research will be significant in designing ofChinese characters．Moreoveg it may be a reference to all