Machine Learning Based Framework for Medical Image Classification in Early Disease Detection
Contributors
Snehlata Kapil Wankhade
Ganesh Khekare
Keywords
Proceeding
Track
General Track
License
Copyright (c) 2026 Sustainable Global Societies Initiative

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Abstract
The importance of the timely and precise classification of MRI brain tumors entails accurate diagnosis, planning of treatment, and better patient prognosis in clinical practice. The current research paper suggests a new hybrid architecture comprising deep feature extraction with the pre-trained ResNet-50 network and traditional machine learning classifiers, namely, Random Forest (RF) and Logistic Regression (LR), and a soft voting ensemble to make predictions homogenous. The suggested methodology takes advantage of the hierarchical nature of the feature representations learned by ResNet-50 and combines them with the advantage of robustness and interpretability of classical classifiers. The provided experimental findings prove that the ensemble system is superior to a standard fine-tuned ResNet-50 model in the accuracy, the capability to generalize as well as the ability of high-resilience, and this fact supports the idea that a combination of deep learning-based feature extraction and classical machine learning combined with the ensemble approach is beneficial in the context of medical image analysis.