Code for Independent Work submission for the Princeton Computer Science Department, Spring 2023, by Maria Khartchenko, class of 2024
This is a replication of two models: Zhang et al. (1) and Wang et al. (2), to see if they perform well on a more complex dataset consisting of not only MS scans and healthy control scans, but also scans from patients with Alzheimer's, TBI, and Parkinsons. I also created two of my own models: Simple_CNN (6-layers: 4 conv, 2 FCL) and Medium_CNN (8-layers: 6 conv, 2 FCL) that performed better on a more limited dataset.
Data_sorting files allow for conversion between .dcm, .gif, .bmp, and .tif into .png, and then sorts them into folders needed for the models. Change the datapaths to fit your computer.
Please feel free to email me with any questions at [email protected].
(1): Zhang YD, Pan C, Sun J and Tang C: Multiple Sclerosis identification by convolutional neural networks with dropout and parametric ReLU. J Comput Sci 28: 818 (2018)
(2): Wang SH, Tang C, Sun J, Yang J, Huang C, Phillips P and Zhang YD: Multiple Sclerosis Identification by 14-Layer Convolutional Neural Network with Batch Normalization, Dropout, and Stochastic Pooling. Front Neurosci 12: 818 (2018)