Related Experiment Video
Updated: Feb 18, 2026
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
43.6K
Brain tumor classification using optimized ResNet50 with dynamic precision optimization for enhanced speed and
Vahid Mehrdad1, Reza Talebzadeh2, Negin Fazaeli2
1Department of Electrical Engineering, Faculty of Engineering, Lorestan University, Khorramabad, Iran. mehrdad.v@lu.ac.ir.
Scientific Reports
|February 16, 2026
Summary
This study introduces an optimized ResNet model for brain tumor detection, achieving 99.69% accuracy. The efficient system reduces computational load and enhances diagnostic speed for clinical applications.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computer Science
Background:
- Brain tumor detection and classification are challenging due to complex morphology.
- Accurate and efficient diagnostic tools are crucial for timely treatment.
Purpose of the Study:
- To develop an optimized deep learning model for accurate and efficient brain tumor detection and classification.
- To improve diagnostic accuracy and reduce computational complexity for clinical deployment.
Main Methods:
- Optimized a ResNet-based model with architectural modifications and specialized layers.
- Employed Transfer Learning (TL), Fine-Tuning, and intelligent computational precision management.
- Integrated a Random Forest (RF) classifier for enhanced result interpretability.
Main Results:
- Achieved up to 99.69% diagnostic accuracy on a dataset of 7023 brain MRI images.
- Demonstrated 100% precision for glioma, pituitary, and healthy cases; 98.71% for meningioma.
- Reduced computational parameters, training time, and increased inference speed.
Conclusions:
- The proposed model offers high efficiency, computational optimization, and clinical applicability for automated brain tumor diagnosis.
- The system balances model complexity with operational accuracy, suitable for resource-limited hardware.
- Enhanced interpretability through RF integration aids physician decision-making.