2018/09/10 by Francesco La Rosa, Mário João Fartaria, La Rosa, Francesco +11
Biochemistry, Genetics and Molecular Biology · Medicine · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Herpesvirus Infections and Treatments #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Multiple Sclerosis Research Studies #RNA regulation and disease
paper · pdf · doi:10.48550/arxiv.1809.03185
openalex publication_date 2018/09/10 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
In this work, we present a comparison of a shallow and a deep learning\narchitecture for the automated segmentation of white matter lesions in MR\nimages of multiple sclerosis patients. In particular, we train and test both\nmethods on early stage disease patients, to verify their performance in\nchallenging conditions, more similar to a clinical setting than what is\ntypically provided in multiple sclerosis segmentation challenges. Furthermore,\nwe evaluate a prototype naive combination of the two methods, which refines the\nfinal segmentation. All methods were trained on 32 patients, and the evaluation\nwas performed on a pure test set of 73 cases. Results show low lesion-wise\nfalse positives (30%) for the deep learning architecture, whereas the shallow\narchitecture yields the best Dice coefficient (63%) and volume difference\n(19%). Combining both shallow and deep architectures further improves the\nlesion-wise metrics (69% and 26% lesion-wise true and false positive rate,\nrespectively).\n