Organ shape plays an important role in various clinical practices, e.g., diagnosis, surgical planning and treatment evaluation. It is usually derived from low level appearance cues in medical images. However, due to diseases and imaging artifacts, low level appearance cues might be weak or misleading. In this situation, shape priors become critical to infer and refine the shape derived by image appearances. Effective modeling of shape priors is challenging because: 1) shape variation is complex and cannot always be modeled by a parametric probability distribution; 2) a shape instance derived from image appearance cues (input shape) may have gross errors; and 3) local details of the input shape are difficult to preserve if they are not statistically significant in the training data. In this paper we propose a novel Sparse Shape Composition model (SSC) to deal with these three challenges in a unified framework. In our method, a sparse set of shapes in the shape repository is selected and composed together to infer/refine an input shape. The a-priori information is thus implicitly incorporated on-the-fly. Our model leverages two sparsity observations of the input shape instance: 1) the input shape can be approximately represented by a sparse linear combination of shapes in the shape repository; 2) parts of the input shape may contain gross errors but such errors are sparse. Our model is formulated as a sparse learning problem. Using L1 norm relaxation, it can be solved by an efficient expectation-maximization (EM) type of framework. Our method is extensively validated on several medical applications, 2D lung localization in X-ray images, 3D liver segmentation in low-dose CT scans, and 3D rat brain structure segmentation in MR microscopy. Compared to state-of-the-art methods, our model exhibits better performance in these studies.


title = "Towards robust and effective shape modeling: Sparse shape composition",
journal = "Medical Image Analysis",
volume = "16",
number = "1",
pages = "265 - 277",
year = "2012",
author = "Shaoting Zhang and Yiqiang Zhan and Maneesh Dewan and Junzhou Huang and Dimitris N. Metaxas and Xiang Sean Zhou",

title = "Deformable segmentation via sparse representation and dictionary learning",
journal = "Medical Image Analysis",
volume = "16",
number = "7",
pages = "1385 - 1396",
year = "2012",
author = "Shaoting Zhang and Yiqiang Zhan and Dimitris N. Metaxas",

Selected publications on segmentation

Shaoting Zhang, Yiqiang Zhan, Maneesh Dewan, Junzhou Huang, Dimitris N. Metaxas, Xiang Sean Zhou: Towards robust and effective shape modeling: Sparse shape composition. Medical Image Analysis 16(1): 265-277 (2012) (Top 25 hottest articles in Medical Image Analysis in 2012 full year [Link])

Shaoting Zhang, Yiqiang Zhan, Dimitris N. Metaxas: Deformable segmentation via sparse representation and dictionary learning. Medical Image Analysis 16(7): 1385-1396 (2012) (its conference version has been selected as MICCAI'11 Young Scientist Award Finalist)

Guotai Wang, Shaoting Zhang, Feng Li, Lixu Gu: A New Segmentation Framework Based on Sparse Shape Composition in Liver Surgery Planning System. Medical Physics, Volume 40, Issue 5, 051913, (2013).
Chosen as the cover of Medical Physics, May 2013

Tian Shen, Shaoting Zhang, Junzhou Huang, Xiaolei Huang and Dimitris Metaxas: Integrating Shape and Texture in 3D Deformable Models: From Metamorphs to Active Volume Models. Book chapter in Multi Modality State-of-the-Art Medical Image Segmentation and Registration Methodologies. Volume I, Chapter 1, A.S. El-Baz, R. Acharya U, and M. Mirmehdi (Editors), Springer.

Zhennan Yan, Shaoting Zhang, Xiaofeng Liu, Dimitris Metaxas, Albert Montillo: Accurate segmentation of brain images into 34 structures combining a non-stationary adaptive statistical atlas and a multi-atlas with applications to Alzheimer's disease. ISBI, 2013. (a collaboration with GE Global Research, through an I/UCRC program)