小波分析在图像处理中的应用

本文介绍小波分析在图像处理中的应用,有多尺度边缘检测、图像识别、图像分割.边缘检测包括可去除多余边缘的一般灰度图像实例,彩色肿瘤图像的边缘提取,汉字放大.图像识别中,对于汉字或字母的识别,是将二维图像信号降维后再进行小波分解,通过计算各层次的分形维数形成特征向量从而可以进行识别;仿射变换后的汉字或其它图像可用多尺度小波的仿射不变函数进行对比识别.最后介绍了两种图像分割算法,首先利用基于能量分析的方法将纹理部分和平滑部分先分开,然后再利用基于奇异性分进行细分,得到最终分割的结果.

Some Applications in Image Procession with Wavelets

This paper introduces some applications in image processing with wavelets,such as edge detection ,pattern recognition and image segmentation .Edge detection includes contour extraction and background removal for an illustration of gray-level,extraction of tumor region in color image and enlargement of type size of Chinese character.As for recognition of Chinese character or character at first, we can reduce the dimensionality of two dimensional patterns to one dimensional sign,then do wavelets decomposition for these one dimensional patterns,at last compute fractal dimension of each sub-pattern,so feature vectors can be obtained and used to recognition.The charaters after affine transformation can also be recognized by a dyadic wavelet affine invariant function. At last we introduce two means of image segmentation. One is based on energy analysis ,another is singularity analysis.