Year: 2025
Type: article
Abstract: Brain tumor detection is a critical aspect of medical diagnostics, as early and accurate identification significantly improves treatment outcomes. This study explores the application of advanced image segmentation techniques for automated brain tumor detection in medical imaging, such as magnetic resonance imaging (MRI) scans. The proposed approach leverages deep learning models, particularly convolutional neural networks (CNNs) and U-Net architectures, to achieve precise tumor segmentation. By preprocessing MRI data and enhancing image quality, the method ensures accurate delineation of tumor boundaries. The system evaluates performance based on metrics such as Dice similarity coefficient (DSC), precision, recall, and accuracy. Results demonstrate that the model effectively distinguishes between healthy tissue and tumor regions, providing reliable diagnostic support to clinicians. This approach holds promise for reducing manual effort, enhancing diagnostic speed, and improving patient outcomes through accurate and timely detection of brain tumors. Future work focuses on integrating multimodal imaging and expanding the dataset to enhance robustness and generalizability.
Language: English
Subfield: Neurology
Field: Neuroscience
Domain: Life Sciences
Open Access status: hybrid
OpenAlex ID: W4409195249