基于小波的自动图像变形

图像变形算法中最费时的是在源图像和目标图像中确定变形特征基元；而到目前为止，大部分图像变形算法都是通过人工或半人工的方式来实现的．如果两幅源图像间的特征基元的确定能自动实现，则不但可以减轻用户的负担，而且可以极大地节省变形算法所需的时间，同时提高了算法的效率．提出了一种基于小波变换的自动图像变形方法．该方法完全依赖于给定的两幅源图像，自动生成期间的变形图像，不需要预先在两幅源图像中指定特征基元．实验结果证明，用本文的算法不仅能自动完成变形操作，而且也取得了满意的结果．

Automatic Image Morphing Based Wavelet Transform

The most tedious part of image morphing is to establish the correspondence of features between images by an animator. But, up to now, the features on an image are usually specified manually by an animator .If an algorithm to automatically select the morphing markers would reduce the burden of feature specification and save a great deal of time and money, at the same time, it improve the efficiency of the algorithm .A automatic image morphing algorithm based on wavelet transform was presented. It depends on completely the given two source image and produces the good morphing image, without specifying the feature marker beforehand .The result shows that the algorithm can create morphing completely automatically and gain the good morphing.