AI-Powered Detection Systems for Precision Neuro-Oncology: From Tumor Diagnosis to Mobile Deployment

This preprint presents an AI-powered detection system for precision neuro-oncology, 
focused on brain tumor classification from MRI scans. Three machine learning 
approaches are compared — Random Forest, Support Vector Machine (SVM), and a custom 
Convolutional Neural Network (CNN) — achieving up to 97% classification accuracy. 
A mobile deployment framework powered by MobileNetV3 enables real-time inference on 
consumer-grade hardware, targeting low-resource clinical settings. A live, interactive 
web-applet is hosted on Google AI Studio for testing and clinical validation.

Code: https://gist.github.com/kumarasmita26/7a64caceafc36b1eb650e1e472e9f862
Live demo: https://ai.studio/apps/drive/1qOH6sD0Sgk1XZOD8xUrRe1u3CtCI_AiZ