Brain Tumor Classification of Medical Images based on Transfer Pre-trained Learning Models

Identifying brain tumor types is crucial for accurate diagnoses and appropriate therapeutic planning in medical imaging. The study aims to develop a robust and accurate system for categorizing medical images of brain tumors, utilizing transfer learning from pre-trained neural network models. The purpose is to optimize the efficiency and precision of brain tumor diagnosis, ultimately facilitating prompt and precise medical intervention by clinicians. This endeavor holds immense importance in promptly diagnosing and treating brain tumors. Transfer learning greatly facilitates the classification of different forms of brain tumors visible in medical imaging, enabling the achievement of this target. Utilizing pre-trained neural network models capitalizes on their comprehension of visual attributes and patterns by adjusting to the specific activity with efficiency and accuracy. Through this strategy, the training process is accelerated, and the model is generalized. Medical professionals can rapidly and reliably diagnosing brain tumors from medical imaging via this technique. he results demonstrate that the AlexNet model showed a remarkable performance, with an accuracy rate of 95.60% and high sensitivity and specificity rates.