NV-Raw2Insights-US
Raw Ultrasound Processing
Ultrasound beamformers usually assume one speed of sound, but tissues carry sound at different speeds, which blurs the image. NV-Raw2Insights-US estimates a spatially varying speed-of-sound map from raw, pre-beamformed IQ channel data so the beamformer can correct for local tissue, like autofocus for ultrasound. A compact two-stage CNN (2.3M parameters) outputs a 32×32 map in m/s, trained against the physics-based DBUA solver.
No agent skill for this model yet. Use it directly via
HuggingFace
or GitHub.
Engineering & research use only — not a medical device, not for diagnosis or clinical decision-making. Skills follow the open Agent Skills spec.