View source for 2013 Summer Project Week:3D prostate segmentation of Ultrasound image

__NOTOC__
<gallery>
Image:PW-MIT2013.png|[[2013_Summer_Project_Week#Projects|Projects List]]
Image:prostatelittle.jpg|The segmentation result of localized active contour.
</gallery>


==Key Investigators==
* Nanjing University of Science and Technology: Xu Li
* BWH: Andriy Fedorov, Tina Kapur , William Wells

<div style="margin: 20px;">
<div style="width: 27%; float: left; padding-right: 3%;">

<h3>Objective</h3>
Segmentation of prostate gland from 3D Transrectal Ultrasound images.
Our goal is to work out a deformable model that can segment the prostate gland well, we are trying to find an appropriate energy function that can have good segmentation result for Ultrasound images. 







</div>

<div style="width: 27%; float: left; padding-right: 3%;">
<h3>Approach, Plan</h3>
The method we now using is localization active contour, we use a ball which centred at the coutour to make the active contour method more localized. so the localized forground and background will only considered on the local ball. Our preliminary results are based on the algorithm by Lankton and Tannenbaum. 
</div>

<div style="width: 40%; float: left;">

<h3>Progress</h3>
Since the segmentation results of prostate on 2D ultrasound image slice is not very satisfying, after discussion with Yi and Matthew, they suggest me to add more constraint in the algorithm like shape prior and texture. So in the next step I will try to add some shape and texture information in the energy function and compare the result. I also learn a lot about the 3D slicer by judging the tutorials.
[http://wiki.na-mic.org/Wiki/index.php/2013_Tutorial_Contest]
</div>
</div>

==Delivery Mechanism==

This work is in very early stages of exploration.

==References==
JieHuang,Xiaoping Yang,Yunmei Chen. A fast algorithm for global minimization of maximum liklihood based on ultrasound image segmentation. Inverse Problems and Imaging.Volume 5, No. 3, 2011, 645–657.

Shawn Lankton,Allen Tannenbaum.Localizing Region-Based Active Contours.IEEE Trans Image Process. 2008 November; 17(11): 2029–2039.