International Journal of Advanced Research in Computer and Communication Engineering

Abstract: Brain tumours are a significant health concern, and timely and accurate diagnosis is crucial for patient care. Magnetic Resonance Imaging (MRI) is a widely used non-invasive diagnostic tool for brain tumour detection. However, there are challenges in accurately classifying brain tumours from MRI images, including: Image Variability in MRI images can vary in terms of resolution, contrast, and acquisition parameters, making it challenging to develop a consistent classification method.

There have been too many methods developed in recent years to diagnose brain tumour. Heterogeneity of Brain tumours come in various types (e.g., glioblastoma, meningioma) and grades (low-grade, high-grade), each requiring different treatment strategies. Accurate classification must account for this heterogeneity. From this study it has been found that identifying and extracting relevant features from MRI images that can discriminate between different tumour types and healthy brain tissue is a complex task. Limited Training Data for the availability of labelled MRI data for brain tumour classification is often limited, and collecting large datasets can be time-consuming and costly. Interpretability, ability to interpret the decisions made by machine learning models in the context of brain tumour classification is crucial for medical professionals to trust and use these tools.Therefore, there is a need to develop a robust and accurate machine learning system that can effectively classify brain tumours from MRI images by addressing the challenges of image variability, heterogeneity, feature selection, limited data, and providing interpretable results.

Keywords: Brain tumour detection, machine learning, MRI, Heterogeneity.

Downloads: | DOI: 10.17148/IJARCCE.2025.14824