Related Experiment Video
Updated: May 23, 2025
06:48
Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images
Published on: January 7, 2019
8.8K
An automated deep learning framework for brain tumor classification using MRI imagery
Muhammad Aamir1, Ziaur Rahman2, Uzair Aslam Bhatti3
1School of Computer Science and Artificial Intelligence, Huanggang Normal University, Huanggang, 438000, Hubei, China. aamirshaikh86@hotmail.com.
Scientific Reports
|May 21, 2025
Summary
This study presents an automated method for detecting brain tumors in MRI scans, significantly improving diagnostic accuracy and efficiency. The novel approach enhances image clarity and uses deep learning for precise tumor identification and classification.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Neurology
Background:
- Accurate and timely brain tumor diagnosis is critical for patient outcomes.
- Manual detection of brain tumors in MRI is challenging, time-consuming, and prone to errors.
- There is a need for automated, robust, and efficient methods for brain tumor identification.
Purpose of the Study:
- To develop an automated method for detecting and classifying brain tumors in MRI images.
- To enhance image quality and improve the accuracy of tumor segmentation.
- To reduce reliance on manual interpretation in the diagnostic process.
Main Methods:
- Image enhancement using guided filtering and anisotropic Gaussian side windows (AGSW).
- Morphological analysis for excluding non-tumor regions.
- Deep neural networks with an attention module for segmentation and feature extraction.
- Ensemble model for multi-class brain tumor classification.
Main Results:
- Achieved 99.94% accuracy on the BraTS2020 dataset and 99.67% on the Figshare dataset.
- The method demonstrates superior automation and robustness compared to existing techniques.
- Successfully identified and classified brain tumors with high precision.
Conclusions:
- The proposed automated approach significantly enhances brain tumor diagnosis in MRI.
- The integration of advanced image processing and deep learning offers a reliable diagnostic tool.
- This method has the potential to improve patient recovery and treatment planning.
