Shape Registration and Low-Rank for Multiple Image Segmentation

Shape similarity is a useful cue for multiple image segmentations. In this paper, a shape registration and low rank based active contour model is developed to segment similar shapes. Under the assumption that the object shapes have similar contours, shape registration is quite helpful to object segmentation and the matrix formed by the shapes of object has a low-rank property. Given a group of multiple test images, normalization and circulant shift are used to register the shapes of objects, and a low-rank constraint is employed during the evolution of object contours. The proposed method could handle complex transformations of image group well, such as large-angle rotation. Experiments on the synthesized and real multiple images show that the proposed approach consistently improves the performance of active contour model and yields more accurate contours than previous methods.

[1]  A. Yezzi,et al.  A variational framework for joint segmentation and registration , 2001, Proceedings IEEE Workshop on Mathematical Methods in Biomedical Image Analysis (MMBIA 2001).

[2]  Guillermo Sapiro,et al.  Geodesic Active Contours , 1995, International Journal of Computer Vision.

[3]  Allan Hanbury,et al.  Metrics for evaluating 3D medical image segmentation: analysis, selection, and tool , 2015, BMC Medical Imaging.

[4]  V. Caselles,et al.  A geometric model for active contours in image processing , 1993 .

[5]  Huimin Yu,et al.  Deep Learning Shape Priors for Object Segmentation , 2013, 2013 IEEE Conference on Computer Vision and Pattern Recognition.

[6]  Rémi Ronfard,et al.  Region-based strategies for active contour models , 1994, International Journal of Computer Vision.

[7]  Y. Nesterov.  Gradient methods for minimizing composite objective function , 2007 .

[8]  Tony F. Chan,et al.  Level set based shape prior segmentation , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[9]  Rui Caseiro,et al.  Exploiting the Circulant Structure of Tracking-by-Detection with Kernels , 2012, ECCV.

[10]  Robert M. Gray,et al.  Toeplitz and Circulant Matrices: A Review , 2005, Found. Trends Commun. Inf. Theory.

[11]  Anthony J. Yezzi,et al.  A variational framework for integrating segmentation and registration through active contours , 2003, Medical Image Anal..

[12]  Tony F. Chan,et al.  Active contours without edges , 2001, IEEE Trans. Image Process..

[13]  Anthony J. Yezzi,et al.  A geometric snake model for segmentation of medical imagery , 1997, IEEE Transactions on Medical Imaging.

[14]  Nanning Zheng,et al.  Automatic salient object segmentation based on context and shape prior , 2011, BMVC.

[15]  B. S. Manjunath,et al.  Shape prior segmentation of multiple objects with graph cuts , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[16]  Roman Goldenberg,et al.  Fast Geodesic Active Contours , 1999, Scale-Space.

[17]  Xavier Bresson,et al.  A Variational Model for Object Segmentation Using Boundary Information and Shape Prior Driven by the Mumford-Shah Functional , 2006, International Journal of Computer Vision.

[18]  Daniel Cremers,et al.  Towards Recognition-Based Variational Segmentation Using Shape Priors and Dynamic Labeling , 2003, Scale-Space.

[19]  Lixu Gu,et al.  A homotopy-based sparse representation for fast and accurate shape prior modeling in liver surgical planning , 2015, Medical Image Anal..

[20]  Rachid Deriche,et al.  Geodesic active regions for supervised texture segmentation , 1999, Proceedings of the Seventh IEEE International Conference on Computer Vision.

[21]  James S. Duncan,et al.  Active Contours with Group Similarity , 2013, 2013 IEEE Conference on Computer Vision and Pattern Recognition.

[22]  Demetri Terzopoulos,et al.  Snakes: Active contour models , 2004, International Journal of Computer Vision.

[23]  Emmanuel J. Candès,et al.  A Singular Value Thresholding Algorithm for Matrix Completion , 2008, SIAM J. Optim..