What Energy Functions Can Be Minimizedvia Graph Cuts?
IEEE Transactions on Pattern Analysis and Machine Intelligence
Optimal Surface Segmentation in Volumetric Images-A Graph-Theoretic Approach
IEEE Transactions on Pattern Analysis and Machine Intelligence
3D Graph cut with new edge weights for cerebral white matter segmentation
Pattern Recognition Letters
The matrix orthogonal decomposition problem in intensity-modulated radiation therapy
COCOON'06 Proceedings of the 12th annual international conference on Computing and Combinatorics
Hi-index | 0.00 |
Medical imaging often involves the injection of contrast agents andthe subsequent analysis of tissue enhancement patterns. Manyimportant types of tissue have characteristic enhancement patterns;for example, in magnetic resonance (MR) mammography, malignanciesexhibit a characteristic "wash out" temporal pattern, while in MRangiography, arteries, veins and parenchyma each have their owndistinctive temporal signature. In such image sequences, there aresubstantial changes in intensities; however, this change is dueprimarily to the contrast agent rather than the motion of sceneelements. As a result, the task of segmenting contrast-enhancedimages poses interesting new challenges for computer vision. Inthis paper, we propose a new image segmentation algorithm for imagesequences with contrast enhancement, using a model-based timeseries analysis of individual pixels. We use energy minimizationvia graph cuts to efficiently ensure spatial coherence. The energyis minimized in an expectation-maximization fashion that alternatesbetween segmenting the image into a number of non-overlappingregions and finding the temporal profile parameters which bestdescribe the behavior of each region. Preliminary experiments on MRmammography and MR angiography studies show the algorithm's abilityto find an accurate segmentation.