Web-Based AI System for Medical Image Segmentation

Image segmentation is a crucial step in the diagnosis of brain tumours, and machine learning has emerged as a promising tool for tumour characterisation from medical imaging data. Despite their enormous potential in automatic segmentation of brain tumours from complex MRI scans, the implementation and use of machine learning algorithms can often present practical challenges to medical imaging researchers. This paper introduces a web-based GUI application designed to integrate all the components needed in deep learning workflows, allowing medical imaging researchers to seamlessly train and infer on data stored on in-house servers or on local machines. Our platform simplifies the process of training and inferring on MRI data using state-of-the-art models, supports integration with XNAT servers, and incorporates powerful tools for visualizing inference results.