计算机应用与软件Issue(4)：210-214,251,6.DOI:10.3969/j.issn.1000-386x.2015.04.050
一种基于HS V空间的自然图像抠图算法
A NATURAL IMAGE MATTING ALGORITHM BASED ON HSV SPACE
摘要
Abstract
Natural image matting is a technology to extract the foreground object from a natural image.We present in this paper an HSV space-based matting algorithm for natural image.The algorithm takes GrabCut as the basis,employs main colour waveform graph of HSV space of the image to dynamically set the parameters of Gauss mixture model,and speeds up the convergence rate by introducing an iterative difference of energy function.The proposed algorithm improves the segmentation efficiency and accuracy of GrabCut algorithm in image matting.Applying to colour images with less background colours,the algorithm achieves better image matting effect.关键词
HSV空间/能量函数/高斯混合模型/迭代差值Key words
HSV space/Energy function/Gauss mixture model/Iterative difference分类
信息技术与安全科学引用本文复制引用
兰红,柳显涛..一种基于HS V空间的自然图像抠图算法[J].计算机应用与软件,2015,(4):210-214,251,6.基金项目
江西省教育厅重点项目（赣教技字［12770］号）；江西省教育厅科技项目（GJJ14430）。 （）
