Abstract—The automatic and computerized recognition of Regions of Interest (ROI) is a crucial step in the process and analysis of medical images. The reasons are many and include the increase of available medical image data, the wide variety of devices and methods for image acquisition and the need to provide mechanisms making the analysis more accurate and the clinicians’ job faster. Within the study on multiple sclerosis, the goal is the recognition of the damaged brain areas by processing images captured through magnetic resonance imaging. In this context, the proposed work is a study on the relationship between brain images obtained by magnetic resonance imaging, using different types of acquisitions. The goal is to understand whether it is somehow possible to identify the different regions of the brain, through a process of segmentation, using a method which allows the user’s independence. The employed volumes are acquired in three different modalities T1-weighted, T2-weighted, and PD for synthetic database; T1-weighted, T2-weighted and FLAIR for real database. The purpose of this paper is to provide the doctor with a tool helping with diagnosis and detecting the possible areas of doubt. Two databases were taken into account, a synthetic one and a real one, and for the synthetic database the parameters of the confusion matrix have been calculated. 
 

Index Terms—segmentation, fuzzy processing, connectedness, T1-weighted, T2-weighted, PD, FLAIR, MS lesion

Cite: Sonia Nardotto, Silvana G. Dellepiane, and Luca Patrone, "Semi-Automatic Segmentation of Multiple Sclerosis Lesion in 4D Modality," International Journal of Signal Processing Systems, Vol. 5, No. 1, pp. 28-33, March 2017. doi: 10.18178/ijsps.5.1.28-33