Medical Robotics · Computer Vision & Medical Imaging · AR/XR · Real-Time Systems
Mingxu Liu
Computer Science Ph.D. Student · Johns Hopkins University
Seeking 2027 internship opportunities
I develop intelligent systems for surgical intervention, with the goal of improving procedural safety, efficiency, and reducing radiation exposure for both patients and surgical teams. My work integrates artificial intelligence with medical image analysis, computer vision, medical robotics, and augmented reality. Research spans MRI, CT, X-ray, and fluoroscopic image processing; multimodal image registration; real-time robotic perception; image-guided navigation; continuum robotics; and 3D reconstruction.
Research & Engineering
Featured Projects
Medical Image Analysis · Registration
MRI-to-Fluoroscopy 2D–3D Registration
End-to-end multimodal registration combining MRI segmentation, MRI-to-CT synthesis, DRR generation, and multi-view fluoroscopic registration.
2.4 ± 1.0 mmtranslation error1.6 ± 0.8°rotation error
Published
Medical Imaging · Deep Learning
MRI Segmentation & Localization for Osteonecrosis
Segmentation and registration for localizing MRI-visible osteonecrotic lesions in intervention-compatible imaging.
0.93femur Dice score0.91pelvis Dice score
Published
Medical Imaging · Benchmarking
PENGWIN Pelvic Fracture AI Benchmark
A multimodal CT/X-ray pelvic-fracture segmentation benchmark and MICCAI challenge series.
48,600simulated X-rays (2024 benchmark)500CT cases (expanded challenge)
Published
AR/XR · Tracking
Device-Agnostic Multi-Sensor Tracking for AR
Multi-view tracking and pose optimization using HoloLens 2, optical trackers, dynamic scene graphs, and pose-graph optimization.
~80%reduction in translational error30%reduction in tracking-loss rate
Published
Computer Vision · AR/XR
AR Arthroscopy Reconstruction & Guidance
Vision-based arthroscopic reconstruction and AR guidance using localization, depth estimation, and 3D Gaussian Splatting.
2.21 mmreconstruction RMSE32.86PSNR
Published
Medical Robotics · Real-Time AI
Real-Time Perception and Autonomous Control for Continuum Surgical Robots
Integrated surgical robotics framework combining real-time AI perception, multimodal sensing, medical image analysis, and robot control.
Ongoing
Background
Education
Ph.D., Computer Science
Johns Hopkins University
Member of the Laboratory of Biomechanical and Image-Guided Surgical Systems (BIGSS) at Johns Hopkins University and the Institute for Integrative & Innovative Research (I³R) at the University of Arkansas. Research focuses on medical robotics, medical image analysis, computer vision, image-guided intervention, and augmented reality.
M.S.E., Robotics
Johns Hopkins University
Member of BIGSS, JHU.
B.Eng., Mechanical Engineering
University of Leeds
Research focused on autonomous driving, imitation learning, and computer vision.
Publications
Selected Publications
X-GuideAR: An Augmented Reality Framework to Mitigate Radiation Exposure during Fluoroscopic Guidance
International Journal of Computer Assisted Radiology and Surgery
Benchmark of Segmentation Techniques for Pelvic Fracture in CT and X-Ray: Summary of the PENGWIN 2024 Challenge
IEEE Transactions on Medical Imaging
End-to-End 2D/3D Registration from Pre-Operative MRI to Intra-Operative Fluoroscopy for Orthopedic Procedures
International Journal of Computer Assisted Radiology and Surgery
Towards Visualizing Early-Stage Osteonecrosis Using Intraoperative Imaging Modalities
Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization