Combining several images can of course be done consecutively or two at a time. To save time, we propose to do it in one step. We present a mosaic technique for multiple images of irregular shapes. First, we present a method of multi-masks to treat the problem of blending selected regions in several images almost seamlessly. And then, with a multiresolution structure of multi-mask signals as the weighting functions, wavelet transform is used to combine images in different resolution levels to reduce the effect of the seam. The results of experiments using portrait images show that the combined image is smooth and almost seamless.%对在多个图象间进行非规则形状图象拼接的多分辨整合技术进行了研究。运用 Multi-mask s方法将多个图象的选定区域同时融合在一起，减小了计算量，并利用小波变换对图象进行 多分辨分解，采用与频段宽度成正比的加权拼接宽度来消除拼缝的影响。实例表明，拼接图 象平滑自然，效果较好。
展开▼
