xLungs — Medical Imaging Inference Application
Java Spring and React application coordinating CT image processing, model inference and radiology reports through RabbitMQ on GPU-powered Kubernetes.
What did the xLungs application do?
xLungs is a medical imaging research project from MI².AI at Warsaw University of Technology. I created the Java and React application that connected the model’s inference pipeline to an interface for radiologists.
My role
My work covered the Java Spring backend, React TypeScript frontend, inference integration, deployment and testing on a GPU-powered Kubernetes cluster during 2024–2025.
Architecture and implementation
The application coordinated CT image processing, organ segmentation and report generation through a multistep RabbitMQ inference pipeline. The backend and frontend connected that processing workflow to the radiologist-facing application.
Technologies
Java, Spring, React, TypeScript, RabbitMQ and Kubernetes with GPU infrastructure.
Research evidence
I co-authored the ISD2025 poster Radiomic Medical Data Transformation for Radiologists Support. It describes image conversion, segmentation, feature extraction and report rendering. The paper reports 89.09% DICE across five organs and processing in under five and a half minutes. These are results of the published research system, rather than separate measurements of my web application.