Brain Tumor Detection System Using CNN with Alexnet Extraction

The Brain Tumor Detection, localization, and area calculation system serves as a hybrid AI-based diagnostic framework that integrates deep learning with classical image processing techniques to improve brain MRI analysis . The system automates tumor detection, reducing manual effort and improving accuracy and consistency at the initial stage, MRI scans are classified as tumor-positive or tumor-negative using Alex Net, a pre-trained convolutional neural network fine – tuned on brain MRI data. The steps serve as a filter , ensuring that only relevant images proceed to further analysis. The subsequent stage involves skull stripping to isolate brain tissue from surrounding structures .otsu’s is applied to adaptive thresholding, followed by morphological operations to improve the quality brain mask and eliminate non-brain artifacts.Next, tumor region is performed using a multi-thresholding approach to manage MRI intensity variations. Morphological smoothing and connected component analysis was used to remove noise and accurately define tumor boundaries. The proposed localizes and quantifies the tumor by calculating its centroid and area as a percentage of total brain volume. Additional shape and texture features support tumor characterization. The results are presented in an interactive report with annotated images and numerical summaries, making the system efficient, reliable, and applicable for hospital deployment.