Capture
A guided phone workflow helps acquire a usable image and flags problems with framing, distance, or image quality.
Highlighted research · Clinical imaging AI
Turning a bedside photograph into a reliable measurement.
Pressure injuries change over time, but their record is often reduced to a categorical stage and an inconsistent image. The question is whether routine clinical capture can yield a reproducible measure of wound state.
Read the rationale
Rationale
The bedside is not a benchmark dataset.
Real photographs vary in lighting, distance, framing, device, wound presentation, and surrounding skin. Models trained only on small, curated image sets may perform well in familiar data while failing in the clinical environment where they are needed. Models have to be trained where care actually happens.
Hypothesis
Combining real-world images with depth, infrared, and clinical context can produce a reliable digital biomarker of wound extent, tissue, and change.
RGB imageInfraredDepth / LiDARClinical context
What we’re building
From capture to clinical feedback.
A bedside workflow designed around the nurse and the patient—not around a curated research image.
Quantify
Image, spatial, thermal, and clinical signals are combined to characterize wound extent, tissue, depth, and change.
Assist
The system returns staging support and longitudinal comparison, with every output open to clinician review.
Staging decisions stay with clinicians. The system contributes a consistent measurement they can inspect, compare over time, and use at the bedside.