Accurate classification of brain tumors1 is a critical step in medical imaging, as it enables timely diagnosis and supports the design of effective therapeutic plans. This study explored a fusion of traditional machine learning methods and hybrid deep learning strategies to classify brain tumors from MRI scans. The Kaggle Brain Tumor Dataset was used consisting of 253 MRI images, including 155 tumor and 98 non-tumor samples. E The study concentrated on the preprocessing actions like resizing, normalization, and augmentation to optimize model performance. Several Machine learning models (Logistic Regression, Random Forest, SVM, KNN, Gradient Boosting, and XGBoost)2 and proposed Attention-based CNN were trained and tested with the help of model accuracy. The findings exhibited that the Attention- Based CNN was able to outperform all other models in showing that it yielded the best accuracy of 94.2 percent, thus revealing its effectiveness in its ability to focus attention on tumor-specific features. This paper indicated the effectiveness of proposed Attention-based CNN method to produce a certain solution to classify brain tumors with a high percentage of certainty. The advancements of AI in Neuro-Oncology represent a noteworthy breakthrough with substantial clinical impact.
Moondra et al. (Tue,) studied this question.