Published Paper Details:

Brain cancer is a critical health condition that poses a great threat to human life due to its rapid progression and high mortality rate. Accurate and timely detection plays a pivotal role in enabling effective treatments and improving patient's survival outcomes. Recently deep learning has raised as a great tool for medical imaging analysis, particularly in the domain of automated cancer diagnosis. This study presents an efficient brain cancer classification framework based on the LeNet-5 CNN architecture, optimized for analysis of magnetic resonance imaging (MRI) scans. The proposed model leverages the inherent capability of CNNs to automatically extract hierarchical features from medical images, thereby reducing reliance on manual feature engineering. The research employs a publicly available brain MRI dataset encompassing multiple tumor types as well as non-tumorous cases. Data preprocessing steps like normalizing, resize, and augmentation were applied to enhance model generalization and mitigate over fitting. The modified LeNet-5 architecture was trained and validated using stratified value splits to ensure class representation. Model's performance evaluated using standard classification metrics, such as Accuracy, Precision, Recall, and F1-Score. Results demonstrated that the proposed algorithm achieves competitive accuracy compared to contemporary deep learning models, while maintaining lightweight architecture suitable real-time clinical applications. The findings suggest that an optimized LeNet-5 model can work as a cost effective and computationally efficient solution for brain cancer-detection, potentially assisting radiologists in early diagnosis and treatment planning. This work contributes to the growing field of AI-driven healthcare by illustrating how classical CNN architectures, when appropriately adapted, can deliver reliable diagnostic performance in medical imaging tasks.