X-Brain: Explainable Automated Recognition of Brain Tumors using Robust Deep Attention CNN

---

Automated brain tumor recognition is crucial for swift diagnosis and treatment in healthcare, enhancing patient survival rates but manual recognition of tumor types is time-consuming and resource intensive. Over the past few years, researchers have proposed various Deep Learning (DL) methods to automate the recognition process over the past years. However, these approaches often lack convincing accuracy and rely on datasets consisting of limited samples, raising concerns regarding real-world efficacy and reliability. Furthermore, the decisions made by black-box AI models in healthcare, where lives are at stake, require proper decision explainability. To address these issues, we propose a robust and explainable deep CNN-based method for effective recognition of brain tumor types, attaining state-of-the-art accuracies of 99.81%, 99.55%, and 99.30% on the training, validation, and test sets, respectively, surpassing both the previous works and baseline models. Moreover, we employed three Explainable AI techniques: Grad-CAM, Grad-CAM++, and Score-CAM for explainability analysis, contributing towards the development of trustworthy and reliable automation of healthcare diagnosis.

• The article has been published in Biomedical Signal Processing and Control and can be accessed from here

Result Summary

Dataset	Accuracy	Loss
Train set	0.9981	0.0589
Validation set	0.9955	0.0739
Test set	0.9930	0.0763

Class-wise Performance over the Test set

Class	Precision	Recall	F1-Score	Support
glioma	0.994141	0.984526	0.989310	517
meningioma	1.000000	0.989031	0.994485	547
no_tumor	0.992352	1.000000	0.996161	519
pituitary	0.985689	0.998188	0.991899	552
accuracy			0.992974	2135
macro avg	0.993045	0.992936	0.992964	2135
weighted avg	0.993022	0.992974	0.992971	2135

XAI Visualization

Glioma	Meningioma
No Tumor	Pituitary

Citation

@article{TONMOY2025106988,
title = {X-Brain: Explainable recognition of brain tumors using robust deep attention CNN},
journal = {Biomedical Signal Processing and Control},
volume = {100},
pages = {106988},
year = {2025},
issn = {1746-8094},
doi = {https://doi.org/10.1016/j.bspc.2024.106988},
url = {https://www.sciencedirect.com/science/article/pii/S1746809424010462},
author = {Moshiur Rahman Tonmoy and Md. Atik Shams and Md. Akhtaruzzaman Adnan and M.F. Mridha and Mejdl Safran and Sultan Alfarhood and Dunren Che},
keywords = {Brain tumor, Medical imaging, XAI, Attention mechanism, Computer vision}
}

---