A Classifier Based on Superpixels and Markov Random Fields for Multiple Sclerosis Lesions on Magnetic Resonance Images

Contenido principal del artículo

A. Reyes
A. Alba
E. R. Arce-Santana
I. Rodríguez-Leyva

Resumen

In this work we present a methodology for lesion detection in Magnetic Resonance Imaging (MRI). Many physicians rely on brain images for the diagnosis of neuropathologies such as Multiple Sclerosis (MS). Unfortunately, in Mexico, not all public health institutions have access to commercial imaging software. For this reason, physicians are interested in the development of tools that could partially replace commercial software, for instance, for the detection of brain lesions. The proposed method uses the Simple Linear Iterative Clustering method (SLIC) in order to reduce the number of variables, followed by a Gauss Markov Measure Field (GMMF) model to perform the classification. In literature, these methods have demonstrated many advantages such as: computational efficiency, border preservation and accuracy. Results obtained with the proposed method are promising and confirm these benefits.

Detalles del artículo

Cómo citar
Reyes, A., Alba, A., Arce-Santana, E. R., & Rodríguez-Leyva, I. (2017). A Classifier Based on Superpixels and Markov Random Fields for Multiple Sclerosis Lesions on Magnetic Resonance Images. Memorias Del Congreso Nacional De Ingeniería Biomédica, 3(1), 88–91. Recuperado a partir de http://memoriascnib.mx/index.php/memorias/article/view/11
Sección
Trabajos Libres 2014-2017