BCMG AI — Objective Wound Measurement & Tracking

Ruler-free wound area, dimensions, and serial trend from an ordinary image — measured by BCMG AI's proprietary metrology, reviewed by the clinician, and kept entirely on-premises.

Wound care runs on measurements, yet everyday documentation still relies on paper rulers, tracings, and visual estimates — methods that vary by observer and visit. BCMG AI converts a simple photograph into a calibrated, reproducible measurement: wound area, greatest dimension × width in millimetres, tissue composition, and a registered map of change since the last visit. Everything runs locally — an on-prem, HIPAA-compliant workflow in which no images or patient data leave the clinic.

Printed BCMG T-FID reference tag beside a wound on the forearm — **Capture with any phone.** A printed, single-use BCMG T-FID reference tag placed beside the wound. Its coded markers define a known plane and absolute scale, so a 1–2 second video clip or a still image from any recent phone becomes a millimetre-accurate measurement — in clinic, or captured at home by the patient to support documented telehealth review between visits.

From image to measurement

A promptable vision model delineates the wound; the boundary is projected into the metric frame to compute area, greatest dimension, perpendicular width, and the wound-bed tissue breakdown.

The clinician stays in control. Every automated result can be accepted, rejected, or corrected with a single click-and-drag re-segmentation, and any measurement outside strict quality bounds is flagged rather than silently reported.

Automated wound delineation with greatest-dimension axis — **Automated delineation.** Wound boundary (green) and greatest dimension (red) on the scale-rectified frame.

Serial wound mapping — measuring change, not just size

Wound area and dimension trajectory across six weeks — **Quantified trend.** Area and greatest dimension across visits with automatic resolution detection — this course closes from 1.02 cm² to healed in six weeks.

Registered inter-visit change map — **Registered change map.** Each visit is registered to the last and every point coloured by how it moved — warm = advancing margin, cool = healed. *(Illustrative.)*

Wound-bed tissue composition map — **Tissue composition.** The wound bed segmented into necrotic / granulation / epithelial zones, tracked alongside size across the same visits.

advancing (warm) no change healed (cool)

objective area & dimensions

registered

pixel-level change between visits

reviewable

clinician accepts, rejects, or corrects

on-prem

offline · HIPAA-compliant · no PHI egress

The BCMG AI Wand — 3-D capture, no reference tag

A handheld, factory-calibrated stereo imaging device that standardises wound capture in the clinic — operated by any trained staff member, fully on-premises.

The Wand pairs two matched, synchronised cameras in an engineered handheld enclosure. Because the pair is stereo-calibrated, absolute scale comes from the 3-D geometry itself — no reference tag on the skin. Its role in the clinic is consistency: the same framing, working distance, and imaging conditions at every visit, so serial measurements stay comparable regardless of who captures them.

Live wand view of a test wound on the forearm — **What the Wand sees.** Live capture of a printed test wound during development. On-screen guidance keeps the operator inside the validated working band (~140–320 mm) and warns when lighting or texture would degrade the measurement — capture problems are caught *while capturing*, not discovered after the patient has left.

Why a wand

No fiducial on the skin. Calibrated stereo recovers millimetre scale directly — nothing is placed near the wound.
Standardised, comparable imaging. Guided distance and framing every visit; measurements track the wound, not the photographer.
Any trained operator. Point, follow the on-screen reticle, sweep — the system selects the best frames and measures.
Surface shape where it matters. The same 3-D geometry yields wound depth and volume where the surface allows — dimensions a flat photograph cannot provide.
On-prem and offline. The device and its workstation form a closed loop inside the clinic.

Exploded CAD view of the wand stereo enclosure — **Engineered enclosure.** The stereo module and optical cover of the handheld unit (development CAD, exploded view).

Calibrated like an instrument

Before clinical use the Wand's stereo geometry is characterised against reference targets, and every capture is measured against those calibrated parameters.

Corner reprojection residual vectors on a calibration target — **Sub-pixel geometry.** Per-corner reprojection residuals on the rectified target — the stereo model resolves the geometry to a **1.25 px** RMS error.

Fiducial pose-coverage series across the sensor — **Full-frame coverage.** The calibration target tracked across many poses spans the entire sensor, so scale holds wherever the wound sits in view.

Reconstructed stereo depth field of a scene — **Metric depth.** A development scene and its reconstructed depth field (warm = nearer) — the mechanism that replaces the reference tag; recovered scale tracks a known reference to **+0.62%**.

Phone + tag and Wand captures feed the same proprietary measurement, review, and reporting pipeline, so either route contributes to one longitudinal record. Device validation and safe-use protocols are designed and overseen by a CAP Fellow (FCAP) and CAP accreditation inspection team lead.

Wound segmentation uses an on-device promptable vision model (SAM 3); stereo figures are from the system's calibration QA. Investigational — for evaluation; not a substitute for clinical judgement.