Optimizing Convolutional Neural Networks Using Differential Evolution Algorithm for Multiple Sclerosis Detection

Multiple Sclerosis (MS) is a chronic disease of the central nervous system that causes serious problems for patients and also challenges for doctors in diagnosis and treatment. One of the main difficulties is that MS lesions in MRI scans look very similar to other brain problems, which makes accurate detection harder. Because of recent progress in machine learning, in this research I proposed a new model called DE-CNN-Gray for automatic detection of MS lesions from brain MRI images. In my method, I first designed and implemented a Convolutional Neural Network (CNN). Then, I used the Differential Evolution (DE) algorithm to optimize the structure of the network, such as the number of layers and the main parameters. The goal was to make the model both efficient and accurate for MS detection. I tested the model on MRI data using 5-fold cross-validation. The results were very strong: 99.4% accuracy, 97.8% sensitivity, 98.9% specificity, and 98.9% F1-score, which is much better than many existing methods. The findings show that the DE-CNN-Gray model can be a useful tool for faster and more accurate detection of MS lesions. The DE algorithm also proved to be an effective way to improve CNN performance. This research can help future work on intelligent diagnostic systems and also support doctors in better understanding brain lesions. Keywords: Multiple Sclerosis, Deep Learning, Convolutional Neural Network, Differential Evolution Algorithm, Brain Lesion Detection, MRI, DE-CNN-Gray