DeepNeuroScan: AI-Powered Brain Tumor Detection Using Machine and Deep Learning Algorithms
Shantanu Shahi, Veeramalai Sankaradass, Vaibhava Vasantrao Desai, M. Amina Begum, N Shilpa, Anita Soni
Abstract
Cancer is the most deadly disease since it destroys healthy tissue and drastically reduces a person's lifespan. When compared to the features recovered using more conventional methods, those generated by deep convolutional layers are more robust and easily distinguishable. Brain tumours can be classified and diagnosed using AlexNet and ResNet-18 in conjunction with the support-vector-machine (SVM). The average filter method is used to improve MRI scans of brain tumours. After that, we use deep convolutional layers and other deep learning (DL) approaches to extract robust and significant deep features. SoftMax and SVM are then used to categorise these features. There are a total of 3,060 pictures in the MRI dataset, with three classifications representing cancers and one representing normal tissue. The greatest results are shown by a hybrid approach involving AlexNet and SVM, which achieves 96.2% accuracy, 96.65% sensitivity, and 98.60% specificity.