Related Experiment Video
Updated: Jan 17, 2026
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
43.4K
An Efficient Neuro-framework for Brain Tumor Classification Using a CNN-based Self-supervised Learning Approach with
Paripelli Ravali1, Pundru Chandra Shaker Reddy1,2, Pappula Praveen1
1School of Computer Science and Artificial Intelligence, SR University, Warangal, Telangana, India.
Current Neuropharmacology
|September 22, 2025
Summary
This study introduces a deep learning framework for glioma brain tumor classification from MRI scans, achieving over 93% accuracy. The method enhances diagnostic speed and reliability for clinical decision-making.
Area of Science:
- Artificial Intelligence
- Medical Imaging Analysis
- Computational Biology
Background:
- Accurate glioma brain tumor grading from MRI is challenging due to limited labeled data and complex clinical evaluation.
- Traditional Convolutional Neural Networks (CNNs) face limitations in feature extraction and data scarcity for glioma classification.
- Developing robust and efficient deep learning frameworks is crucial for improving non-invasive tumor grading.
Purpose of the Study:
- To develop a robust and efficient deep learning framework for improved glioma classification using MRI images.
- To overcome data scarcity and limited feature extraction issues inherent in traditional CNNs.
- To enhance the accuracy and efficiency of glioma grading for clinical applications.
Main Methods:
- A multi-stage framework utilizing SimCLR-based self-supervised learning for representation learning.
- Deep Embedded Clustering for effective feature extraction and grouping, followed by EfficientNet-B7 for initial classification.
- A weighted ensemble of EfficientNet-B7, ResNet-50, and DenseNet-121, with hyperparameters optimized by a Differential Evolution-optimized Genetic Algorithm.
Main Results:
- EfficientNet-B7 achieved approximately 88-90% classification accuracy.
- The weighted ensemble model improved classification accuracy to approximately 93%.
- Genetic optimization further enhanced accuracy by 3-5% and reduced training time by 15%.
Conclusions:
- The proposed framework effectively addresses data scarcity and feature extraction limitations in CNNs for glioma classification.
- The combination of self-supervised learning, clustering, ensemble modeling, and evolutionary optimization yields improved performance and robustness.
- The framework offers an accurate, scalable solution for glioma classification, supporting faster clinical decision-making and promising real-world diagnostic applications.
