Abstract: Image collage?is?a?technique for?fast browsing and?generating digital montage?of?a?collection of images. This paper proposes a method to measure the similarity of two images and remove the redundant images based on this measurement, then uses graph cuts optimization technique to find the best seam of smooth in color and gradient between close images. And then it uses the image blending to make the result consistency and remove the artifacts in the seam. Compared with the original technique, this collage framework can obtain better smooth result with time insensitive.

Key words: graph cuts, image collage, image blending

摘要： 图像贴图是为了对一个图像集合进行快速浏览和生成数字蒙太奇的一种技术手段，提出了一种度量图像之间相似性的方式来衡量图像之间的相似度，并除去图像中相似度较高的图像以达到去除冗余信息的效果；采用图割（graph cuts）优化技术寻找相邻图像之间颜色和梯度上最平滑的最佳绗缝，并利用图像融合技术消除直接在绗缝处拼接所出现的不自然现象。与原有技术相比，该贴图系统具有平滑效果好和速度快的优点。

关键词: 图割, 贴图, 图像融合