Detection and Classification of Brain Tumor and 3D Reconstruction of MRI

Brain tumors pose a significant global health concern, with millions affected worldwide, often resulting in death. In the United States, approximately 70,000 people have primary brain tumors, making brain cancer the 10th most common tumor in India. Accurate diagnosis is crucial for initiating appropriate treatment, and improving patient survival. Treatment varies based on tumor type, size, location, and aggressiveness. Deep learning methods, like transfer learning-based techniques using traditional classifiers, are increasingly used for tumor detection in MRI images. Our study explores VGG-16 and VGG-19 transfer learning methods, leveraging 3D Slicer’s adaptable algorithm for reconstructing MRI scans of brain tumors. By utilizing robust software components like Tcl/Tk, VTK, and ITK, alongside an intuitive application layer, we enable advanced functionalities such as image-guided surgery and brain mapping. Our work highlights 3D Slicer’s alignment with research goals, despite challenges like prolonged diagnosis times due to workload pressures, potentially leading to errors. This predictive models enhance patient outcomes in this critical medical domain, with improved ROC-AUC score >0.9 and classification accuracy of $\sim{99\%}$.