Applications

Milliseconds matter in the OR.

From remote 3D surgery to AI ultrasound in rural clinics, YUAN builds the capture and edge-compute layer of modern medicine — with Foxconn, National Taiwan University Hospital, and NVIDIA's medical AI ecosystem.

<50 ms: glass-to-glass latency for remote 3D surgery
15 µs: HDMI/SDI DMA capture latency
87%: accuracy distinguishing normal from abnormal tissue
50: iGood Liver AI ultrasound systems deployed across rural Taiwan

Remote surgery

Fifty milliseconds, glass to glass.

Telesurgery through remotely controlled 3D medical robots promises to put a specialist's hands in any operating room on earth. The constraint is brutal: to feel like in-person surgery, the delay from camera to surgeon's eyes must stay within 70 milliseconds.

YUAN's integrated solution, built on the NVIDIA Jetson Orin platform, gets 3D camera images into the system through low-latency HDMI/SDI DMA capture in as little as 15 microseconds. AV1 encoding preserves image quality while conserving bandwidth, and YUAN's SkyLink X transmission technology brings glass-to-glass latency well below 50 milliseconds.

The same platform's 275 TOPS of AI computation assists physicians in assessing local tissue in real time and computes image depth on the fly — sharpening the performance of the 3D robotic arm mid-procedure.

Diagnostics

A second pair of eyes on every ultrasound.

Reading a tumour on ultrasound takes concentration, years of experience, and continuous training. YUAN's diagnostic assistance system — built on the VPP6N0-S NX with a 4K60 HDMI2.0 M.2 capture card and 70/100 TOPS of compute — was trained on more than 6,000 images annotated by medical experts, and distinguishes normal from abnormal tissue with accuracy up to 87%.

That capability left the lab. Developed with National Taiwan University Hospital and the Good Liver Foundation, YUAN's iGood Liver AI system — powered by the NVIDIA Jetson Orin platform — has deployed 50 AI-assisted ultrasound systems to rural and underserved clinics across Taiwan, spotting early-stage suspicious liver lesions when they are most treatable.

The point is equity: the same standard of preventive care in a remote township as in a metropolitan medical centre.

A second pair of eyes on every ultrasound.

With Foxconn

CoDoctor AI, from colonoscopy to the smart hospital.

In collaboration with Foxconn, YUAN delivers CoDoctor AI Endovia — real-time AI polyp and lesion detection for colonoscopy, anchored by YUAN's NEAR T7X platform on NVIDIA IGX Thor and validated through extensive endoscopy data at Kaohsiung Medical University Chung-Ho Memorial Hospital.

The system scores bowel preparation, monitors withdrawal time to reduce missed findings, and generates zero-click structured clinical reports powered by NVIDIA Nemotron — functional-safety-grade edge AI at the point of care.

The collaboration reaches further into the operating room: a scrub nurse collaborative robot, jointly developed by Foxconn, Kawasaki Heavy Industries, Taichung Veterans General Hospital, and YUAN, uses the NVIDIA Isaac for Healthcare Rheo Blueprint and Isaac GR00T VLA architecture for multimodal perception and surgical scene understanding.

CoDoctor AI, from colonoscopy to the smart hospital.

The architecture

Every signal in the OR, one AI pipeline.

YUAN's Smart Surgical Video Platform integrates the operating room's many sources — 4K endoscopes, PACS, ultrasound, remote medical devices — with 4K30 2D and 3D processing, Single View, PIP, and Multi-View modes routed simultaneously to multiple OR monitors, recording to USB, NAS, or FTP, and low-latency IP streaming over RTSP, RTMP, and WebRTC for cross-hospital consultation.

Beneath it, NVIDIA Holoscan Sensor Bridge aggregates the heterogeneous streams an OR produces — endoscopic video in HD and 4K, physiological signals like heart rate and SpO2, surgical device telemetry, external camera feeds — and synchronises them into a single AI pipeline, while IGX Thor handles secure, functional-safety-grade inference at the edge.