AI-assisted ultrasound analysis on VPP6N0-S NX.

Challenge

Ultrasound interpretation requires trained clinicians and depends on image quality, anatomy, acquisition technique, and clinical context. YUAN developed an edge-AI system intended to assist with image review during an examination.

Solution

The system combines a VPP6N0-S NX, a low-latency 4K60 HDMI 2.0 M.2 capture card, 70/100 TOPS of compute, and a medical-image model. YUAN reports that the model was trained on more than 6,000 images annotated by medical experts and measured up to 87 percent accuracy when distinguishing normal from abnormal tissue.

Key benefits

The model returns an analysis result during the examination so a clinician can decide whether further review is warranted. Reported performance does not establish results for every patient, probe, operator, or clinical setting. The system is an assistance tool and does not replace clinical judgment, validation, or applicable medical-device requirements.