Classifying Brain Tumors with Deep Neural Networks and Image Preprocessing

Abstract

One of the most serious conditions affecting the nervous system is brain tumors, and a patient’s successful treatment depends on a correct diagnosis. A radiologist’s manual processing of MRI images takes a lot of time and can result in some incorrect diagnoses, particularly when the tumor’s boundaries are unclear. In order to address this issue, this paper suggests an automated system for brain tumor detection that combines deep learning algorithms with OpenCV-based image processing methods. First, the MRI pictures are preprocessed using methods for region of interest extraction, noise reduction, and contrast enhancement. A convolutional neural network is then fed the preprocessed images in order to identify and detect tumors. It is anticipated that the suggested system will increase tumor detection accuracy without sacrificing efficiency. The experimental results indicate that the combination of traditional image processing techniques with deep learning algorithms can improve the performance of the system and provide a useful tool for computer-aided medical diagnosis.