Uloženo v:
| Hlavní autor: | |
|---|---|
| Médium: | Recurso digital |
| Jazyk: | |
| Vydáno: |
Zenodo
2026
|
| On-line přístup: | https://doi.org/10.5281/zenodo.19711590 |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
Obsah:
- <p>Brain tumor detection using Magnetic Resonance Imaging (MRI) is a critical task in medical diagnosis, as early and accurate identification significantly improves patient survival rates. This study presents a deep learning-based approach for automated classification of brain tumors into four categories: glioma, meningioma, pituitary tumor, and no tumor.</p> <p>Transfer learning is applied using multiple pretrained convolutional neural network models, including DenseNet121, EfficientNetB0, ResNet50, and MobileNetV2. A comparative evaluation of these models is conducted, and MobileNetV2 is selected as the final model due to its optimal balance between classification accuracy and computational efficiency.</p> <p>The model is initially trained using frozen base layers and further improved through fine-tuning with a reduced learning rate to enhance generalization capability. Experimental results demonstrate that the proposed approach achieves consistent and reliable classification performance across all tumor categories. The model performs particularly well in identifying “no tumor” and “pituitary” classes, while minor misclassification occurs between glioma and meningioma due to their visual similarity.</p> <p>Overall, the proposed system provides an efficient and practical solution for automated brain tumor classification and has the potential to assist medical professionals in faster and more accurate diagnosis.</p>