摘要
提出了一种基于小波变换的图像变形技术。首先,利用小波变换对源图像和目标图像进行多分辨率分解,得到对应的低频近似信号分量和高频细节信号分量;然后,通过基于时变权函数的多分辨率插值,得到各变形结果图像的小波系数;最后,通过小波重构得到各变形图像。实验结果表明,利用本算法可以方便地实现自然的图像变形。
We propose a wavelet transformation based image-morphing algorithm. First, we decompose source and target images into multi-resolution, and get the corresponding detail and approximate coefficients. Then we calculate the resulting wavelet coefficients through multi-resolution interpolation using a time-varied weight function Finally, we reconstruct the morphing result images through inverse wavelet transformation. The experiment demonstrates the effectiveness of this algorithm in achieving smooth and vivid image morphing.
出处
《现代计算机》
2006年第6期28-30,共3页
Modern Computer
关键词
小波变换
图像变形
图像融合
Wavelet Transform
Image Morphing
Image Fusion
