Inside Point-n-Capture: How Wound Border Detection and Ambient Scribing Work Together
product
A technical look at Point-n-Capture — how border detection, dimensional measurement, and ambient SOAP drafting combine into a single bedside capture.
Most wound documentation tools bolt imaging onto a note editor. Point-n-Capture inverts that: the image, the measurement, and the note are the same event. Here's how the pieces fit together.
Border detection: segmentation, not contrast
Older wound imaging depended on color contrast — the wound had to look different enough from surrounding skin. That fails on darker skin, on macerated tissue, and on wounds with diffuse edges. Point-n-Capture uses trained segmentation to identify the wound bed and perimeter as objects, not as contrast gradients.
Dimensional measurement without a fiducial
Some tools require you to place a reference sticker or ruler in-frame. Point-n-Capture derives scale from device sensors and depth cues, so the clinician points and shoots — no sticker, no re-shoot when the sticker falls off in a home visit.
Ambient scribing tied to the capture
While the image is being processed, the AI medical scribe is listening to the encounter. The transcript isn't a raw dump — it's parsed into structured wound fields (location, tissue type, exudate, undermining) alongside the free-text SOAP narrative. The measurement from the image populates the same record.
Coding suggestions from what was captured
Because the record is structured — not prose — ICD-10, CPT, and HCPCS suggestions are derived from actual documented findings, not inferred from a paragraph. Debridement depth, wound size, and location drive the code, and the code is tied back to the evidence in the note.
Offline-first for mobile routes
Captures don't require live connectivity. The image, transcript, and measurement queue on-device and sync when signal returns. Nothing about the bedside workflow changes based on where the visit happens.
Healing trajectory as a byproduct
Because every visit produces a structured, measured record, the Healing dashboard can trend area, depth, and tissue composition over time without any extra data entry.
The point of the architecture: documentation, measurement, and coding are outputs of the visit, not tasks performed after it.