An Optimized Deep Learning Model with Transfer Learning-Based Feature Extraction for Brain Tumor Classification

Abstract

The early and accurate detection of brain tumors is critical for effective treatment and patient survival. In this paper, we propose a novel approach for brain tumor detection that integrates various preprocessing and deep learning techniques to enhance diagnostic accuracy. The method begins by applying a Median Filter (MF) for noise reduction, ensuring the removal of image artifacts while preserving important structural details. Contrast Limited Adaptive Histogram Equalization (CLAHE) is then used to enhance contrast, improving the visibility of critical features within brain CT images. For feature extraction, we employ EfficientNet, a state-of-the-art convolutional neural network known for its balance between performance and efficiency. The extracted features are passed through a custom deep learning model designed for tumor classification. The ADAM optimizer is used to fine-tune the hyperparameters, achieving optimal training performance. Model evaluation is performed using accuracy, precision, recall, and F1-score metrics, providing a comprehensive assessment of the model's effectiveness. Experimental results demonstrate that the proposed system achieves high accuracy in detecting brain tumors, showcasing its potential as a reliable tool for clinical diagnosis.