Grab Cut数字抠图方法的研究与实现

在图像处理中,关于图像分割方法的高效性研究有着重要意义.传统的图像分割方法或者是应用图片色彩的信息(全局),例如常用图像处理工具PhotoShop中的魔术棒工具;或者是利用图片边界信息(局部),如PhotoShop中的智能套索工具.Graph Cuts分割算法有效的结合在图像处理中所用到的两种图像信息.在Graph Cuts算法的基础上发展而来的Grab Cut算法,在交互工作相当简单的基础上,可以得到良好的分割结果.结合作者所研究的课题,对Grab Cut方法在图像分割部分的实现进行了比较详细的介绍.

Studying and implementing Grab Cut digital matting algorithm

In the digital image processing,the research of efficiency of interactive image segmentation is of great practical importance.The traditional method of image segmentation uses either texture(color) information,e.g. magic wand in PhotoShop,or edges(contrasts) information,e.g.intelligent scissors in PhotoShop.An approach called Graph Cuts has been developed successfully combing both types of information.The Grab Cut algorithm whichis based on Graph Cuts is accurate and it needs less interactive work.The authors combine their research,and introduce the implementing the segmentation phase of the Grab Cut algorithm.