摘要
该文提出了一种基于小波矩的新型图形识别算法。小波矩除了具有矩特征的平移、缩放和旋转不变及抗噪性强的特性以外，还增加了小波对图象结构精细特征的把握能力强的优点。文章结合FFT变换的优点提出了改进的离散傅立叶－小波矩，并由所提取的小波矩特征与最小距离分类器结合形成一个图象识别系统，成功实现了对精细复杂汉字图符的识别。通过与采用传统的几何矩识别结果比较，在识别准确率、识别速率和抗噪性上性能都有较大的提高。
： In this paper we present a new method based on wavelet moment for recognizing complex patterns．Besides having the invariant to the translation，scaling and rotation，the wavelet moment has the multiresolution properties so it is suitable for classing the very similar objects．Combining FFT transform，we proposed the reformative discrete FFT－Wavelet moment．Using a minimum－distance classifier together with our FFT－Wavelet invariant moment，we have constructed a recognition system which succeed in classing the very liken Chinese character．Comparing with the geometrical moment，the classification rate，recognizing efficiency and antinoise capability have been improved．
出处
《计算机工程与应用》
CSCD
北大核心
2001年第7期47-49,52,共4页
Computer Engineering and Applications
基金
自然科学基金资助!(编号:990384-11)
关键词
小波矩
FFT变换
图形识别
算法
模式识别
计算机
： Wavelet moment，translation scaling and rotation invariant，FFT transform，minimum－distance classifier
