Related Experiment Video
Updated: Jan 9, 2026

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Real-Time 3D Ultrasound Reconstruction Using U2-Net Segmentation for Cardiovascular Interventional Surgery
None:
In current cardiovascular interventional surgeries, physicians primarily rely on observing 2D X-ray images for surgical guidance. This approach exposes both physicians and patients to significant X-ray radiation and lacks intuitive 3D information. To address this issue, we propose a real-time 3D ultrasound reconstruction technique based on U2-Net segmentation and electromagnetic tracking. The segmented blood vessels are reconstructed in real-time at a rate of 60 frames per second using volume rendering, and the real-time ultrasound images are fused with preoperative CT scans. The RMS error of ultrasound-to-navigation coordinate registration is only 1.375±0.369 mm, providing physicians with intuitive, radiation-free 3D surgical guidance.Clinical Relevance- This technique enables cardiovascular interventional surgeries to be performed without reliance on X-ray guidance.

