Production-ready hardware for the intelligent OR.

The operating room of the future will not run on reference designs. At COMPUTEX 2026, YUAN High-Tech unveiled its Smart Healthcare hardware — edge AI platforms anchored by NVIDIA IGX Thor and the NVIDIA Holoscan Sensor Bridge that carry real-time inference, multi-source medical video integration, and connected surgical environments from blueprint to deployment-ready hospital infrastructure. As an OEM hardware partner in NVIDIA's medical AI ecosystem, YUAN builds the production layer that system integrators and ISVs deploy.

A surgical video platform for the connected OR

The first of two showcase systems is a Smart Surgical Video Integration and Streaming Platform, purpose-built for intelligent operating rooms and telemedicine. It brings together 4K endoscopes, PACS, ultrasound, and remote medical devices, delivering real-time display, recording, and IP-based streaming inside and beyond the OR. It processes 4K30 2D and 3D imagery with Single View, PIP, and Multi-View modes routed simultaneously to multiple monitors; records flexibly to USB, NAS, or FTP; and streams at low latency over RTSP, RTMP, and WebRTC for remote consultation and cross-hospital collaboration.

CoDoctor AI Endovia: detection during the procedure

The second is CoDoctor AI Endovia, a real-time AI polyp and lesion detection system for colonoscopy, delivered in collaboration with Foxconn and validated with extensive endoscopy data at Kaohsiung Medical University Chung-Ho Memorial Hospital. It runs on YUAN's NEAR T7X platform, built on NVIDIA IGX Thor, delivering functional-safety-grade inference during live clinical procedures. Alongside detection it provides AI-assisted bowel preparation scoring and withdrawal-time monitoring to reduce missed findings, and zero-click structured clinical report generation powered by NVIDIA Nemotron — with multi-source video integration and remote consultation built in.

What the Sensor Bridge actually does

While IGX Thor handles secure inference, the Holoscan Sensor Bridge is the layer that makes the OR's data usable. It aggregates heterogeneous sources — HD and 4K endoscopic video, physiological signals such as heart rate and SpO2, device telemetry, external cameras — into one coherent pipeline before inference begins. Its stream-first architecture processes data as it arrives, with GPU/CPU co-processing and zero-copy transfer for continuous streaming inference. And it keeps time: multi-source timestamp alignment and frame-level synchronisation give AI models the temporal context they need to detect lesions and clinical events reliably. The full path runs medical devices, through the Sensor Bridge and Holoscan pipeline, to IGX Thor inference and Nemotron reporting — delivered as one medical-grade system.

The OEM partner behind the application

For ISVs and integrators building on NVIDIA, YUAN offers production-ready hardware designed against NVIDIA Reference Architectures, pre-integrated sensor and streaming pipelines, functional-safety design experience for medical environments, a clinically validated co-development record, and long-term roadmap and customisation support for OEM and ODM engagements. YUAN absorbs the hardware, safety, and supply complexity so partners can concentrate on the AI applications that differentiate them.

Visit YUAN at COMPUTEX 2026 — TWTC Hall 1, 1F, Robotics Zone, Booth A0925a.