Enhanced Brain Tumor Detection in MRI Using Advanced CNNs

Abstract

Artificial Intelligence (AI), a branch of computer science, focuses on developing intelligent systems capable of performing tasks that require human-like cognitive abilities, such as speech recognition, learning, planning, and problem-solving. Deep learning, a subfield of machine learning, utilizes algorithms to detect patterns in data for complex tasks. This thesis explores sophisticated deep learning methods for Magnetic Resonance Imaging (MRI)-
based brain tumor detection and classification. The primary goal is to develop an effective model that assists medical professionals in diagnosing brain cancers with speed and precision. The World Health Organization reports that brain cancer mortality rates are high in Asia, which emphasizes the need for early identification. To address this challenge, the study introduces an enhanced approach using Convolutional Neural Networks (CNNs), specifically
YOLOv7, integrated with advanced components such as the Convolutional Block Attention Module (CBAM) and Spatial Pyramid Pooling Fast+ (SPPF+). The project employs a dataset of 10,000 MRI images, encompassing nontumor cases, meningiomas, pituitary tumors, and gliomas,
to develop a CNN model that improves accuracy, reduces false positives/negatives, and offers detailed tumor segmentation and feature extraction. The model’s robustness is validated through diverse datasets and cross-validation techniques, demonstrating its potential for integration into
clinical practice. The project also addresses The capacity to analyze deep learning models and ethical considerations regarding patient data privacy, emphasizing the importance of transparency and responsible AI deployment in healthcare.