Computer-Based Segmentation, Change Detection and Quantification for Lesions in Multiple Sclerosis

GÖÇERİ E., Songul C.

2017 International Conference on Computer Science and Engineering (UBMK), Antalya, Turkey, 5 - 08 October 2017, pp.177-182, (Full Text) identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/ubmk.2017.8093371
  • City: Antalya
  • Country: Turkey
  • Page Numbers: pp.177-182
  • Keywords: MS, lesion segmentation, lesion burden, MR images, QUANTITATIVE FOLLOW-UP, WHITE-MATTER LESIONS, AUTOMATED SEGMENTATION, MR-IMAGES, ACTIVE LESIONS, BRAIN MRI, SUBTRACTION, REPRODUCIBILITY, LOAD, MS
  • Akdeniz University Affiliated: Yes

Abstract

Multiple Sclerosis (MS) is a neurological, progressive widespread disease whose diagnosis, treatment and monitoring have vital importance. However, manual method based on visual inspection for diagnosis and time-series assessments of changes in MS lesions is not re-producible and quantitative. Also, it is subjective and yields in inter-/intra-observer variabilities. Furthermore, the conventional method is time-consuming and mostly urgent results are required in practice. Therefore, in the literature, automated techniques have been proposed to detect and segment MS lesions and also to evaluate changes in these lesions quantitatively. In this paper, these automated approaches are presented for the radiologists, neurologists and researchers who are interested in this subject or want to improve former works or want to develop novel automated methods to overcome the problems in this active research area.