Inside Point-and-Capture Wound Imaging: No Ruler, No Marker
aiengineering
A technical look at how depth-sensor imaging, auto-framing, and computer-vision edge detection replace the sticker, the ruler, and the redo photo.
Most wound imaging tools still require a calibration marker in the frame — a sticker next to the wound so software has a known-scale reference. Point-and-capture imaging removes that step. Here is how the pieces fit together.
Depth sensing does the scale work
Modern phone depth sensors (LiDAR on iOS, ToF on some Android devices) report a distance for every pixel. Given the camera intrinsics and a per-pixel depth map, the system can convert pixel coordinates into real-world millimeters without any object of known size in the frame. That is what makes the sticker unnecessary — the full explanation is in point-and-capture wound imaging without calibration markers.
Auto-framing catches operator drift
On-screen guidance keeps the phone in the range where depth data is clean and the wound fills enough of the frame for edge detection to work. The shutter fires automatically when framing, distance, and stability are all in tolerance. Clinicians stop retaking blurry, tilted shots.
Computer vision traces the border
A segmentation model outlines the wound edge and separates granulation, slough, and eschar. If the trace is off, a clinician drags a vertex to correct it. The full CV pipeline is walked through in inside automated wound edge detection.
The 3D reconstruction
Multiple depth frames are fused into a mesh, which is what preserves undermining and tunneling that a flat photo would flatten. Volume changes over time — not just area — become a chartable trajectory in the wound healing software.
The outcome: one hand, one motion, measurements you can defend.