RELM for MRI Images
Description
This study introduces a brain tumor classification method that combines a hybrid feature extraction technique with a regularized extreme learning machine (RELM). The method extracts features from brain images, applies principal component analysis (PCA) to enhance them, and then classifies the tumor type using RELM. Experimental results on a public brain image dataset demonstrated that this approach improves classification accuracy from 91.51% to 94.23%, outperforming existing methods.
Creator
King Saud University
Information
Pediatrics or Adult
Adult
Speciality
Neurology
Modality
MRI
Training
3,064 brain tumor MRI images from 233 patients at different planes
Github
Publication
FDA