Medical ultrasound strain imaging for malignant breast tumors Segmentation 

Adam Slater
Matt McCormick
ECE 533 Project Proposal
11.7.2005

	The purpose of this project is to perform image segmentation to
benefit a researcher at UW-Madison.  The researcher is involved in
medical ultrasound strain imaging1.  A strain image is a spatial map of
the local deformation that a tissue experiences after the application of
mechanical load.  Often, stiffer tissue regions will deform less than
regions that are softer.  Stiff regions will show up as dark areas in
the image.  It has been found that malignant breast tumors are stiffer
than the surrounding tissue2. 

	The researcher's aim is to quantify the contrast that exists
between a tumor and its background.  Therefore, the boundary between the
tumor and the surrounding tissue must be defined.  Our goal is to define
the region that contains the tumor.  There is one tumor in each image. 

	Boundaries drawn by radiologists are available to assist in the
development and validation of our segmentation algorithm. 

	Our approach is to use segmentation methods developed in class
such as dilation, erosion, boundary extraction, and level set methods as
suggested by department faculty3-5.  MATLAB toolboxes have been
developed for the level set method6. 


1.  Ophir J, Kallel F, Varghese T, Konofagou E, Alam SK, Krouskop T,
Garra B, Righetti R.  Elastography.  Comptes Rendus De L Academie Des
Sciences Serie Iv Physique Astrophysique.  2001;2:1193-1212. 

2.  Hall TJ, Zhu YN, Spalding CS.  In vivo real-time freehand palpation
imaging.  Ultrasound in Medicine and Biology.  2003;29:427-435. 

3.  Osher S, Sethian JA.  Fronts propagating with curvature dependent
speed: Algorithms based on Hamilton-Jacobi Formulations.  Journal of
Computational Physics.  1988;79:12-49. 

4.  Tsai R, Osher S.  Level Set Methods and Their Applications in Image
Science.  Communications in Mathematical Sciences.  2003;1:623-656. 

5.  Sethian JA.  Level set methods : evolving interfaces in geometry,
fluid mechanics, computer vision, and materials science: Cambridge :
Cambridge University Press; 1996. 

6.  Sumengen B.  In: Vision Research Lab at UC Santa Barbara. 
http://barissumengen.com/level_set_methods/