基于抠像技术的图像无缝融合算法

Seamless Cloning Algorithm Based on Image Matting Technique

为了更快更好的进行图像融合,在用自由融合算法对泊松图像编辑与抠像技术进行改进的基础上,实现了一套图像自由无缝融合算法。该融合算法首先利用抠像技术来提取图像中的物体边缘,然后迭代求解泊松方程,从而取得了自然的融合效果。该算法与最新的无缝融合算法相比,其优点在于:①对于前景与背景的颜色变化复杂的图像,仍可得到准确的边缘;②可以使用户能引导前景映射图的走向;③无论图像前景层中存在多少个洞,融合效果将不受任何影响。这使得该算法的应用具有相当的灵活性。对于目前的图像融合技术而言,该算法可以获得更高的融合质量、更快的处理速度以及多样的融合效果。

In this paper, we present a method called arbitrary cloning in which image matting technique and Poisson image editing are applied to accomplish the image seamless cloning. First, the object boundary is extracted through image matting technique. And Poisson equation is solved to achieve natural result for seamless cloning. Compared with state-of-art techniques, the algorithm shows advantages that great performance can be reached in bad conditions when there is complicated color variation in the foreground and background image. The α matte can be guided by the user input and the algorithm is not affected even when there are many holes in the foreground image. The algorithm is very flexible in application. Within current cloning technique, the algorithm guarantees better output, higher speed and diverse results.