A hybrid EfficientNet-DbneAlexnet for brain tumor detection using MRI images.
Vasavi G1, Vaddadi Vasudha Rani2, Sreenu Ponnada3
1Department of CSE (Cyber Security), School of Engineering, Malla Reddy University, Hyderabad, India.
Computational Biology and Chemistry
|December 4, 2024
Summary
This study introduces EfficientNet-DbneAlexnet, a novel method for brain tumor (BT) detection using enhanced Magnetic Resonance Images (MRI). The approach achieves high accuracy in identifying cancerous growths, improving diagnostic capabilities.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Brain tumors (BTs) pose a significant health risk due to rapid abnormal cell growth.
- The heterogeneity in shape, size, and location of BTs complicates their identification.
- Magnetic Resonance Images (MRI) are crucial for detecting malignant tumors.
Purpose of the Study:
- To develop and evaluate an advanced deep learning model for accurate brain tumor detection.
- To enhance the efficiency and precision of brain tumor identification from MRI scans.
Main Methods:
- An image enhancement technique using Piecewise Linear Transformation (PLT) was applied.
- Skull stripping was performed using Fuzzy Local Information C Means (FLICM).
- Tumor segmentation utilized a Projective Adversarial Network (PAN), followed by feature extraction and detection via the EfficientNet-DbneAlexnet model.
Main Results:
- The EfficientNet-DbneAlexnet model demonstrated a sensitivity of 90.36%, accuracy of 92.77%, and specificity of 91.82%.
- The proposed method effectively detects brain tumors from enhanced MRI scans.
Conclusions:
- The developed EfficientNet-DbneAlexnet model offers a promising and accurate approach for brain tumor detection.
- This method can aid in improving the diagnostic accuracy of brain tumors using MRI.
