AI帮你理解科学

AI 生成解读视频

AI抽取解析论文重点内容自动生成视频

生成解读视频

AI 溯源

AI解析本论文相关学术脉络

生成溯源树

AI 精读

AI抽取本论文的概要总结

微博一下：

This work presents a means of incorporating prior shape information into the geodesic active contour method of medical image segmentation

Statistical Shape Influence in Geodesic Active Contours

CVPR, (2002): 1316-1323

引用：1668|浏览88

EI WOS

下载 PDF 全文

原文链接

其它链接

dblp.uni-trier.de academic.microsoft.com

引用

微博一下

关键词

maximum a posteriorishapecomputed tomographylevel setsynthetic data更多(2+)

摘要

A novel method of incorporating shape information into the image segmentation process is presented. We introduce a representation for deformable shapes and define a prob- ability distribution over the variances of a set of training shapes. The segmentation process embeds an initial curve as the zero level set of a higher dimensional surfa...更多

代码：

数据：

精读

下载 PDF 全文

PPT (上传PPT)

相似论文

引用论文

被引论文

简介

The anatomical structures that appear in magnetic resonance (MR) or computed tomography (CT) scans are often explicitly extracted or segmented from the image for use in surgical planning, navigation, simulation, diagnosis, and therapy evaluation.
The authors refer to the process of labeling individual voxels in the volumetric scan by tissue type, based on properties of the observed intensities as well as anatomical knowledge about normal subjects.
With CT data, segmentation of some structures can be performed just using intensity thresholding or other low-level image processing.
The distribution of intensity values corresponding to one structure may overlap those of another structure, defeating intensity-based segmentation techniques

重点内容

The anatomical structures that appear in magnetic resonance (MR) or computed tomography (CT) scans are often explicitly extracted or segmented from the image for use in surgical planning, navigation, simulation, diagnosis, and therapy evaluation
This work presents a means of incorporating prior shape information into the geodesic active contour method of medical image segmentation
The shape representation of using Principal Component Analysis on the signed distance map was chosen with the intention of being robust to slight misalignments without requiring exact point-wise correspondences
The representation and the curve evolution technique merge well together since the evolution requires a distance map of the evolving curve, which is inherent in the shape model
A different statistical shape model could be tied into the evolution method, or a different method of model-based matching could be used with the proposed shape model

结果

The authors have tested the segmentation algorithm on synthetic and real shapes, both in 2D and in 3D.
A training set of rhombi of various sizes and aspect ratios was generated to define a shape model.
Test images were constructed by embedding the shapes of two random rhombi with the addition of Gaussian speckle noise and a low frequency diagonal bias field.
Figure 9 illustrates several steps in the segmentation of the synthetic objects.
The yellow curve illustrates the MAP shape and pose at each time step.

结论

This work presents a means of incorporating prior shape information into the geodesic active contour method of medical image segmentation.
The representation and the curve evolution technique merge well together since the evolution requires a distance map of the evolving curve, which is inherent in the shape model.
These two modules need not be coupled.
A different statistical shape model could be tied into the evolution method, or a different method of model-based matching could be used with the proposed shape model

表格

Table1: Partial Hausdorff distance between our segmentation and the manually-segmented ground truth

Download tables as Excel

基金

This work was supported by NSF Contract IIS-9610249, NSF Contract DMS-9872228, and NSF ERC (Johns Hopkins University agreement) 8810-274
Shenton’s NIMH grants, K02 M-01110 and R01 MH-50747, and Dr
McCarley’s grant, R01-40799, and the Brockton Schizophrenia Center for the Department of Veterans Affairs

引用论文

V. Caselles, R. Kimmel, and G. Sapiro. Geodesic active contours. Int’l J. Comp. Vision, 22(1):61–79, 1997.
V. Caselles, J. Morel, G. Sapiro, and A. Tannenbaum. Introduction to the special issue on PDEs and geometry-driven diffusion in image processing and analysis. IEEE Trans. on Image Processing, 7(3):269–274, 1998.
L. D. Cohen. On active contour models and balloons. CVGIP: Image Understanding, 53(2):211–218, 1991.
T. Cootes, C. Beeston, G. Edwards, and C. Taylor. Unified framework for atlas matching using active appearance models. In Int’l Conf. Inf. Proc. in Med. Imaging, pages 322–333. Springer-Verlag, 1999.
Y. Guo and B. Vemuri. Hybrid geometric active models for shape recovery in medical images. In Int’l Conf. Inf. Proc. in Med. Imaging, pages 112–12Springer-Verlag, 1999.
D. Huttenlocher, G. Klanderman, and W. Rucklidge. Comparing images using the hausdorff distance. IEEE Trans. Patt. Analysis and Mach. Intell., 15:850–863, 1993.
M. Jones and T. Poggio. Multidimensional morphable models. In IEEE Int’l Conf. Comp. Vision, pages 683–688, 1998.
M. Kass, A. Witkin, and D. Terzopoulos. Snakes: Active contour models. Int’l J. Comp. Vision, 1(4):321–331, 1988.
A. Kichenassamy, A. Kumar, P. Olver, A. Tannenbaum, and A. Yezzi. Gradient flows and geometric active contour models. In IEEE Int’l Conf. Comp. Vision, pages 810–815, 1995.
L. M. Lorigo, O. Faugeras, W. Grimson, R. Keriven, R. Kikinis, and C.-F. Westin. Co-dimension 2 geodesic active contours for mra segmentation. In Int’l Conf. Inf. Proc. in Med. Imaging, pages 126–139. Springer-Verlag, 1999.
S. Osher and J. Sethian. Fronts propagating with curvaturedependent speed: Algorithms based on hamilton-jacobi formulation. J. of Comp. Physics, 79(1):12–49, 1988.
N. Paragios and R. Deriche. Geodesic active regions for supervided texture segmentation. In IEEE Int’l Conf. Comp. Vision, pages 926–932, 1999.
J. A. Sethian. Level Set Methods. Cambridge University Press, 1996.
L. Staib and J. Duncan. Boundary finding with parametrically deformable contour models. IEEE Trans. Patt. Analysis and Mach. Intell., 14(11), 1992.
M. Turk and A. Pentland. Eigenfaces for recognition. J. of Cogn. Neuroscience, 3(1):71–86, 1991.
Y. Wang and L. Staib. Boundary finding with correspondence using statistical shape models. In Proc. IEEE Conf. Comp. Vision and Patt. Recog., pages 338–345, 1998.
A. Yezzi, A. Tsai, and A. Willsky. A statistical approach to snakes for bimodal and trimodal imagery. In IEEE Int’l Conf. Comp. Vision, pages 898–903, 1999.