小波滤波的移动最小二乘图像变形方法  

Image Deformation Based on Wavelet Filter Using Moving Least Squares

作　　者：李洪安[1] 康宝生[1] 张雷[1] 

机构地区：[1]西北大学信息科学与技术学院,西安710127

出　　处：《小型微型计算机系统》2013年第8期1900-1903,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61272286)资助;西北大学研究生创新教育项目(YZZ12092)资助

摘　　要：为了获得具有真实感的变形图像,改变以往直接对图像进行变形的做法,首先对图像进行预处理,提出基于小波滤波的移动最小二乘法图像变形.对原始图像先进行滤波,把图像分成低频子图像和高频子图像,只对低频部分使用基于控制点集的移动最小二乘法进行变形,对不同部位的轮廓和边缘进行不同尺度的变形,较好的描述了图像中的形状和轮廓信息;对高频部分不作处理,有效的保持了图像的细节信息.再将变形后的低频部分与原图像的高频部分相加得到最后的变形结果.实验表明这种方法很好的保留了图像的高频信息,可以使图像产生平滑的变形,获得具有真实感的变形效果.In order to achieve a realistic deformed image, an image deformation method based on wavelet filter using moving least squares is proposed, preprocessing the image firstly instead of the old ways which deform the image directly. This method will filter the original image into a high and a low frequency sub image, and then only use the moving least squares method based on control points to deform the low-frequency sub image which can better describe the outline and edge of the image, but not the high-frequency sub image which can keep the detail information effectively, and take different scales to deform the different parts of the outline and edge. Then the final result will appear by adding the deformed low-frequency part to the high-frequency part of the original image. Experiments show that the high-frequency information is well preserved in the result images using this method, and so the deformation results are satisfactory and realistic.

关 键 词：图像变形 移动最小二乘法 小波滤波 高频子图像 低频子图像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]