将图像傅里叶频谱分析应用于汉字视标的选择  

Selection of Chinese visual targets by Fourier Frequency spectrum analysis

作　　者：王晨晓[1] 陈湘君[1] 贺极苍[1] 金成鹏[1] 瞿佳[1] 吕帆[1] 

机构地区：[1]温州医学院附属眼视光学院,325027

出　　处：《中华眼科杂志》2008年第7期587-590,共4页Chinese Journal of Ophthalmology

基　　金：浙江省科技计划资助项目（2005C14001）;温州市科技局资助项目（Y20060227）

摘　　要：目的以汉字图像各空间频率的频谱能量相似为原则，筛选汉字作为汉字视力表的备选视标。方法从2500个常用汉字中选取两类笔画数的汉字：少笔画数汉字（2、3、4画，共172个）和中笔画数汉字（7、8、9画，共890个）。将其中的每个汉字制作成BMP图片格式。然后，利用傅里叶频谱分析，获取每个汉字图像各空间频率成分的频谱分布。以此分别对所有的少笔画数和中笔画数汉字的各空间频率的频谱能量求均值，将其作为参照值，采用聚类分析将汉字图像进行聚类来选取汉字。结果与少笔画数和中笔画数汉字各空间频率的频谱能量均值相似的汉字分别为50和40个。结论通过汉字图像的傅里叶频谱分析，以汉字图像的各空间频率的频谱能量相似作为筛选视标的主要原则，符合了视觉系统具有的傅里叶频谱分析特性。此方法应用于汉字视标的筛选具有重要的意义。Objective To select Chinese characters for Chinese visual acuity chart by using the method of Fourier frequency spectrum analysis. Methods BMP images were made for 1062 Chinese characters with small or medium number of strokes. The images were processed to derive frequency spectrum at the frequency domain by using a MatLab program. Euclidean distance from Cluster analysis was then used to classify and select the Chinese characters for optotypes of the visual acuity chart, based on their mean frequency spectrum from all the characters analyzed. Results Fifty and forty Chinese characters were selected from the groups of the small and the middle number of strokes respectively. Condusions It is important and interesting to select Chinese characters by using Fourier analysis method for designing Chinese visual acuity chart that is consisting with the property of the visual system of the eye.

关 键 词：傅里叶分析 视力测定法 汉字视标 研究设计 

分 类 号：R686[医药卫生—骨科学]