CAD-Driven Approach with Enhanced Imaging and Deep Learning for Accurate Diagnosis of Brain Tumor

Recently, the biomedical research field has become popular due to the use of digital image processing as the diagnosis of clinical patients has become more accurate and efficient with the help of Computer-Aided Diagnosis (CAD). Identifying diseases promptly and arranging treatment accordingly can improve the life quality and life expectancy of patients with brain tumors. Different tools have been designed to detect brain tumors, but the current diagnosis system utilizing Magnetic Resonance Imaging (MRI) scanning devices is expensive and often yields low accuracy and efficiency. To address this, a novel methodology utilizing CAD and different kinds of algorithms is proposed in this method to predict brain tumors more effectively. The proposed method uses a 2D Adaptive Bilateral Filter algorithm for image restoration to improve image quality. Additionally, it employs an Adaptive Histogram Adjustment algorithm to enhance brightness and contrast. For segmentation of the Region of Interest for brain tumors, uses the U-net algorithm and calculates various features with the convolutional neural network. With the aid of a Deep Convolutional Neural Network, it classifies disease images and stages using a Deep Learning approach. The proposed approach results in superior accuracy and efficiency in diagnostic decision-making compared to existing systems