Documentation/4.0/Modules/EMSegment Easy

Introduction and Acknowledgements

This work was funded by the ARRA Supplement to the Neuroimage Analysis Center (NAC), funded by the National Institutes of Health. Information on NAC can be obtained from the NAC website. Author: Kilian Pohl, UPenn Contributor1: Daniel Haehn, UPENN Contact: Kilian Pohl, <email>pohl.kilian@gmail.com</email>

University of Pennsylvania   Surgical Planning Laboratory

Module Description

This module is designed for users who want to do a quick intensity based image segmentation. The remainder of this sections describes the work flow of the two modes in further detail:

Use Cases

MR image segmentation

Tutorials

N/A

Panels and their use

Step 1: Define Input Datasets

Define the volumes to be segmented. Each volume has to be from the subject and has to represent a different mode, such as T1 and FLAIR, than the other input volumes.

Step 2: Define Structure

The user specifies the number of structures to be segmented and the label of each structure. You can further add class to an existing structure by right-clicking on the structure and selecting "Add sub-class". The label and corresponding color of a structure are modified by selecting the structure in the tree.

Step 3: Sampling and Class Weights

Define the intensity distribution for each structure of interest by taking samples in the image of interest. Users can also specify the relative to weight of a node in the tree with respect to other structures. If a class should be overestimated in the final label map then this can be adjusted by lowering the weight of that structure and repeating the segmentation.

Step 4: Edit Node-based Parameters

Users specify the relative to weight of a node in the tree with respect to other structures which are children of the same parent node. The first tab also specifies the weight of the input channels as well as the atlases. The value 'Alpha' specifies the smoothing applied to the structure (via MRFs). The second tab (Stopping Condition) lists the number of iterations associated with the segmentation task. By default, the Bias iteration is set to -1 which means that it is performed each iteration. If the value is greater -1 then the inhomogeneity computation is stopped after n iterations. The third tab specifies printing out intermediate results, which are saved in the working directory specified in the next step

Step 5: Miscellaneous

This panels allows the user to specify the region of interest to be segmented. Pressing the 'Segment' button starts the segmentation.

Once the parameters are specified, the target images are segmented using the EM Segmentation algorithm (Pohl et al. TMI 2007). The label map with corresponding statistics is returned after successful completion of the algorithm.

Similar Modules

Please visit the segmentation section for similar modules

References

• Pohl K, Bouix S, Nakamura M, Rohlfing T, McCarley R, Kikinis R, Grimson W, Shenton M, Wells W. A Hierarchical Algorithm for MR Brain Image Parcellation. IEEE Transactions on Medical Imaging. 2007 Sept;26(9):1201-1212. [bib]