Smart ROI

Tracking Aorta

The SmartROI tool is used to track the aorta in MR PCA sequence (15 dynamics).


Boundary

The SmartROI tool is first used to find the brain boundary in a slice in this T2 perfusion sequence. Then a predefined relationship between boundary control points is used to extract watershed territories as new ROI's.
Overview
An important problem in automated analysis of images is the requirement to locate variable image structures (ie: anatomical regions). We have implemented the simple but powerful technique of deformable template location within Tina. We have already used this technique to locate the aorta for analysis of shape and area change through the cardiac cycle (see image above left) and this will be demonstrated on data sets from subjects with Marfans syndrome. A comparison with human segmentation has shown not only a dramatic increase in data throughput but also a factor of two improvement in repeatability [1]. In addition we have extended the technique to support the identification of regions away from the main structure. This process we call an affine transfer region of interest (ATROI) as it is based on linear combinations of control points from the deformable model. The technique is fully integrated into the Tina system so that an arbitrary region can be used to identify any region of the data for processing with any of the other tools. The technique is stable and accurate and allows the researcher to simply subdivide anatomical regions in data which otherwise have no discernible image structure. This process will be demonstrated using an automatic perfusion analysis system which computes mean parametric statistics from predefined geometric regions determined from the edge of the brain (see image above right).

Downloads
source
Library toolkit
binary 1
Sun Sparc - Solaris 2
binary 2
Intel x86 - Linux 2
Publications
[1] A J Lacey, N. A. Watson, N A Thacker, and A Jackson. Automatic tracking of the aortic boundary in blood flow analysis using cardiac MR images.' In Proc. ECR 2000, volume 9. European Congress of Radiology, Springer, March 2000
the end
[Back to software]