Haoxiang Yang

UE Plugin Integrating AI Automatic Segmentation and 3D Reconstruction of Liver MRI Images for VR

Summary

This project presents an AI-powered Unreal Engine 5 plugin for automatic segmentation and 3D reconstruction of liver MRI images, enabling immersive visualization in virtual reality.

Key features include:

  • Deep Learning Models (SegResNet, U-Net) for liver, vessel, and tumor segmentation.
  • 3D Reconstruction using the Marching Cubes algorithm, generating meshes from MRI data in real time.
  • Unreal Engine 5 Plugin supporting interactive 2D/3D visualization, synchronized views, and seamless backend integration.
  • FastAPI Server for scalable AI inference with support for NIfTI and DICOM formats.

This system provides an end-to-end pipeline from AI model training to VR-based visualization, supporting medical education, research, and diagnostic assistance.

Thesis: UE Plugin Integrating AI Automatic Segmentation and 3D Reconstruction of Liver MRI Images for VR

GitHub Repository: https://github.com/xiangq9/LiverAIforUEVR

Demo Video:

Contact