全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...

融合KernelPCA形状先验信息的变分图像分割模型

DOI: 10.11834/jig.20150806

Keywords: 图像分割,变分方法,形状先验,核主成分分析(KernelPCA),姿态不变性

Full-Text   Cite this paper   Add to My Lib

Abstract:

目的基于能量最小化的变分图像分割方法已经受到研究人员的广泛重视,取得了丰硕成果。但是,针对图像中存在的噪音污染、目标被遮挡等情况,则难以正确分割。引入先验形状信息是解决该问题的一个重要方向,但是随之而带来的姿态变化问题是一个难点。传统的做法是在每步迭代过程中单独计算姿态变换参数,导致计算量大。方法在基于KernelPCA(KPCA)的形状先验模型基础上,提出一种具有内在的姿态不变性的KPCA形状先验模型,并将之融合到C-V变分图像分割模型中。结果提出模型无须在每步迭代中显式地单独计算姿态变换参数,相对于C-V模型分割正确率能够提高7.47%。同时,针对KPCA模型中计算高斯核函数的参数σ取值问题,也给出一种自适应的计算方法。结论理论分析及实验表明该模型能较好地解决先验形状与目标间存在的仿射变化问题,以及噪音、目标被遮挡等问题。

References

[1]  Kass M, Witkin A, Terzopoulos D. Snakes: active contour models[J].Int. J. Comput. Vis., 1988, 1(4): 321-331.[DOI: 10.1007/BF00133570]
[2]  Zheng Q, Dong E, Cao Z, et al. Active contour model driven by linear speed function for local segmentation with robust initialization and applications in MR brain images[J]. Signal Processing, 2014, 97: 117-133.[DOI: 10.1016/j.sigpro.2013.10.008]
[3]  Liu W, Shang Y, Yang X. Active contour model driven by local histogram fitting energy[J]. Pattern Recognit. Lett., 2013, 34: 655-662.[DOI: 10.1016/j.patrec.2013.01.005]
[4]  Caselles V, Kimmel R, Sapiro G. Geodesic active contours[J]. Int. J. Comput. Vis., 1997, 22(1): 61-79.
[5]  Wang Y, Xiang S, Pan C, et al. Level set evolution with locally linear classification for image segmentation[J]. Pattern Recognit., 2013, 46: 1734-1746.[DOI: 10.1016/j. patcog. 2012. 12. 006]
[6]  Liu L, Zeng L, Shen K, et al. Exploiting local intensity information in C-V model for noisy image segmentation[J]. Signal Processing, 2013, 93: 2709-2721.[DOI: 10.1016/j. sigpro. 2013. 03. 035]
[7]  Chan T, Zhu W. Level set based shape prior segmentation[C]//Proceedings of IEEE Computer Society Conference on CVPR’05. Washington DC; IEEE, 2005,2: 1164-1170.[DOI: 10. 1109/CVPR. 2005. 212]
[8]  Chan T, Vese L. Active contours without edges[J]. IEEE Trans. image Process., 2001, 10(2): 266-277.[DOI: 10. 1109/83. 902291]
[9]  Leventon M, Grimson W, Faugeras O. Statistical shape influence in geodesic active contours[C]//Proceedings of IEEE Conference on CVPR’2000. Washington DC: IEEE, 2000, 1: 316-323. [DOI: 10.1109/CVPR.2000.855835]
[10]  Yan P, Zhang W, Turkbey B, et al. Global structure constrained local shape prior estimation for medical image segmentation[J]. Comput. Vis. Image Underst., 2013, 117(9): 1017-1026.[DOI: 10.1016/j.cviu.2013.03.006]
[11]  Dambreville S, Rathi Y, Tannenbaum A. A framework for image segmentation using shape models and kernel space shape priors[J]. Pattern Analysis and Machine Learning, 2008, 30(8): 1385-1399.[DOI: 10.1109/TPAMI.2007.70774]
[12]  Chen Y, Tagare H. Using prior shapes in geometric active contours in a variational framework[J]. Int. J. Comput. Vis., 2002, 50(3): 315-328.[DOI: 10.1023/A:1020878408985]
[13]  Liu W, Shang Y, Yang X. A shape prior constraint for implicit active contours[J]. Pattern Recognit. Lett., 2011, 32: 1937-1947.[DOI: 10.1016/j.patrec.2011.09.012]
[14]  Cremers D, Osher S. Kernel density estimation and intrinsic alignment for shape priors in level set[J]. Int. J. Comput. Vis., 2006, 69(3): 335-351.[DOI: 10.1007/s11263-006-7533-5]
[15]  Charpiat G, Faugeras O, Keriven R. Approximations of shape metrics and application to shape warping and empirical shape statistics[J]. Found. Comput. Math., 2005, 5(1): 1-58.[DOI: 10. 1007/s10208-003-0094-x]
[16]  Shen J, Li Y X, Zhou Z M, et al. Shape prior constrained KPCA object segmentation with parameter adaption[J]. Journal of Image and Graphics, 2013, 18(7): 783-789. [沈霁, 李元祥, 周则明, 等.参数自适应的KPCA先验形状约束目标分割[J]. 中国图象图形学报, 2013, 18(7): 783-789.] [DOI: 10. 11834/jig.20130714]
[17]  Wang Q. Kernel principal component analysis and its applications in face recognition and active shape models[J]. arXiv Prepr. arXiv1207.3538, 2012.

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133