Ayelet Akselrod-Ballin 

We introduce a multiscale approach that combines segmentation with classification to detect abnormal brain structures in medical imagery, and demonstrate its utility in automatically detecting Multiple Sclerosis (MS) lesions in 3D multi-channel magnetic resonance (MR) images. Our method uses segmentation to obtain a hierarchical decomposition of a multi-channel, anisotropic MR scans. It then produces a rich set of features describing the segments in terms of intensity, shape, location, neighborhood relations, and anatomical context. These features are then fed into a decision forest classifier, trained with data labeled by experts, enabling the detection of lesions at all scales. Unlike common approaches that use voxel-by-voxel analysis, our system can utilize regional properties that are often important for characterizing abnormal brain structures. We provide experiments on two types of real MR images: a multi-channel Proton density-, T2- and T1-weighted data set of 25 MS patients and a single channel fluid attenuated inversion recovery (FLAIR) data set of 16 MS patients. Comparing our results with lesion delineation by a human expert and with previously extensively validated results shows the promise of the approach.
 

DownLoads

Multiple Sclerosis database:

Requires permission of Prof. M. Fillipi head of the Neuroimaging Research Unit, Hospital San Raffaele, Milan, Italy