Difference between revisions of "2014 Project Week:MultiAtlas MultiImage Segmentation"
| Line 3: | Line 3: | ||
<gallery> | <gallery> | ||
Image:PW-SLC2014.png|[[2014_Winter_Project_Week#Projects|Projects List]] | Image:PW-SLC2014.png|[[2014_Winter_Project_Week#Projects|Projects List]] | ||
| − | Image: | | + | Image:groundtruthMABMIS.png| Manual labels (ground truth) in IXI dataset |
| + | Image:resultMABMIS.png| Our automatic labeling result | ||
</gallery> | </gallery> | ||
Revision as of 18:47, 2 January 2014
Home < 2014 Project Week:MultiAtlas MultiImage Segmentation
Manual labels (ground truth) in IXI dataset
Our automatic labeling result
Key Investigators
- Minjeong Kim, Dinggang Shen, UNC Chapel Hill
- Xiaofeng Liu, Jim Miller, GE Research
Project Description
Objective
To deal with the limitations of existing pairwise registration methods between images with large shape difference, we develop an algorithm for multi-atlas-based multi-image segmentation of brain images and release it as a Slicer module.
Approach, Plan
Our algorithm in the module performs 1) a novel tree-based groupwise registration method for concurrent alignment of both the atlases and the target images, and 2) an iterative groupwise segmentation method for simultaneous consideration of segmentation information propagated from all available images, including the atlases and other newly segmented target images.
Progress
We have developed the Slicer module called MABMIS and fully tested it using LONI LPBA40 and IXI datasets. The result by our Slicer module shows 2% improvement compared to pairwise registration framework in terms of the averaged overlap ratio between automatic segmentation and ground truth labels. We are releasing our Slice module in NITRC (https://www.nitrc.org).
