Automatic Ischemic Stroke Lesion Segmentation Using Single MR Modality and Gravitational Histogram Optimization Based Brain Segmentation

摘要

In this paper the automatic and customized brain segmentation followed by a stroke lesion detection technique is presented applying single modality Magnetic Resonance Images (MRIs). A novel intensity-based segmentation technique called gravitational histogram optimization is developed for this purpose. By applying histogram gravitational optimization algorithm the brain can be segmented into discriminative area including stroke lesion. The mathematical descriptions as well as the convergence criteria of the developed algorithm are presented in detail. The application of the proposed algorithm in the segmentation of single Diffusion-Weighted Images (DWI) modality of healthy and lesion MR image slices for different number of segments is presented and the results are discussed. The segmented areas are then employed in automatic lesion slice detection and lesion extraction technique. The stroke lesion is extracted from the recognized lesion slice with acceptable accuracy.