Regularized Multi-Label Fast Marching And Application To Whole-Body Image Segmentation

In this paper, we propose a computationally efficient regularization strategy for the Fast Marching (FM) segmentation of multiple organs. Segmentation is based on interactive seeds placements, where the seeds define either organs of interest or the background. Regularization efficiently compensates for the sensitivity of the FM to narrow bridges between different organs with similar intensities. It also leads to segmentations that are far less sensitive to seeds location than by using the standard FM cost. The driving motivation of our work is the quantitative analysis of Chronic Lymphocytic Leukemia (CLL), which is the most common B-cell malignancy and mostly affects elderly people. This requires the segmentation of more than ten organs in whole-body Magnetic Resonance images. The disease path and progression being highly heterogeneous with important inter-patient variability, the segmentation is highly based on clinician experience, which is difficult to automatically reproduce. In this context, the proposed segmentation algorithm compares very favorably to the tools usually available for clinicians.