物体轮廓形状超像素图割快速提取方法

被引:13
作者
张荣国 [1 ]
刘小君 [2 ]
董磊 [3 ]
李富萍 [1 ]
刘焜 [2 ]
机构
[1] 太原科技大学计算机科学与技术学院
[2] 合肥工业大学机械与汽车工程学院
[3] 中北大学机械与动力工程学院
关键词
图像分割; 超像素; 图割; 水平集; 轮廓形状;
D O I
10.16451/j.cnki.issn1003-6059.201504007
中图分类号
TP391.41 [];
学科分类号
080203 ;
摘要
提出一种水平集框架下物体轮廓形状超像素图割快速提取方法.该方法首先均匀化放置种子点,通过对超像素化演化力的设定,生成具有区域相似特征的超像素,这些超像素对原图像的划分既能保持目标轮廓形状的几何特性,又可避免超像素间的互相重叠.然后构建超像素标号和Heaviside函数的关联关系,应用图割建立M-S能量函数的优化模型.最终利用超像素图割提取目标轮廓的几何形状.实验表明,超像素化的图像像素数目大幅度减少,转化后的优化模型符合图割对能量函数进行优化的要求,图割中最小割/最大流方法避开微分方程的求解,这些措施在保证轮廓形状提取效果的基础上提高提取效率.
引用
收藏
页码:344 / 353
页数:10
相关论文
共 11 条
  • [1] Smooth Chan–Vese segmentation via graph cuts[J] . Pattern Recognition Letters . 2012 (10)
  • [2] Active contours driven by local and global intensity fitting energy with application to brain MR image segmentation
    Wang, Li
    Li, Chunming
    Sun, Quansen
    Xia, Deshen
    Kao, Chiu-Yen
    [J]. COMPUTERIZED MEDICAL IMAGING AND GRAPHICS, 2009, 33 (07) : 520 - 531
  • [3] Fast image segmentation based on multilevel banded closed-form method[J] . Shoudong Han,Wenbing Tao,Xianglin Wu,Xue-cheng Tai,Tianjiang Wang.Pattern Recognition Letters . 2009 (3)
  • [4] Image segmentation evaluation: A survey of unsupervised methods[J] . Hui Zhang,Jason E. Fritts,Sally A. Goldman.Computer Vision and Image Understanding . 2007 (2)
  • [5] Graph cuts and efficient N-D image segmentation
    Boykov, Yuri
    Funka-Lea, Gareth
    [J]. INTERNATIONAL JOURNAL OF COMPUTER VISION, 2006, 70 (02) : 109 - 131
  • [6] Object segmentation using graph cuts based active contours[J] . Ning Xu,Narendra Ahuja,Ravi Bansal.Computer Vision and Image Understanding . 2006 (3)
  • [7] Lazy snapping
    Li, Y
    Sun, J
    Tang, CK
    Shum, IY
    [J]. ACM TRANSACTIONS ON GRAPHICS, 2004, 23 (03): : 303 - 308
  • [8] A multiphase level set framework for image segmentation using the Mumford and Shah model
    Vese, LA
    Chan, TF
    [J]. INTERNATIONAL JOURNAL OF COMPUTER VISION, 2002, 50 (03) : 271 - 293
  • [9] Evolution, Implementation, and Application of Level Set and Fast Marching Methods for Advancing Fronts[J] . J.A. Sethian.Journal of Computational Physics . 2001 (2)
  • [10] A study on continuous max-flow and min-cut approaches. Jing Y,Egil B,Xue-Cheng T. IEEE Conference on Computer Vision and Pattern Recognition (CVPR) . 2010