In multiple sclerosis, conventional MRI is primarily used for manual assessment of brain abnormalities, with quantitative analysis limited by arbitrary scales and poor interpretability of image intensities. This thesis introduces novel approaches for conventional MRI image analysis, including a deep learning-based tool for accurate and automated lesion segmentation, and intensity scaling methods for quantifying the biologically relevant information content in T1-weighted image intensities. These advances open new possibilities for MS research with conventional MRI.
«
In multiple sclerosis, conventional MRI is primarily used for manual assessment of brain abnormalities, with quantitative analysis limited by arbitrary scales and poor interpretability of image intensities. This thesis introduces novel approaches for conventional MRI image analysis, including a deep learning-based tool for accurate and automated lesion segmentation, and intensity scaling methods for quantifying the biologically relevant information content in T1-weighted image intensities. These...
»