2D Photo-Based Burn Assessment — Research Build

Get Started with BurnCare

How to request access, and how to run a case from the web portal or the Android app.

1. Get an account

Access is restricted to authorized clinical and research personnel, so every new user starts with a request rather than a self-serve sign-up.

Submit the request an account form with your name and role.
An administrator reviews and approves the request from the admin dashboard.
You'll be notified once approved; you can check the request state any time on the check application status page.
Once approved, sign in from the user login page.

2. Walk through the web portal

This is the actual screen-by-screen flow — from signing in to reading a finished TBSA & fluid estimate — exactly as it looks in the portal.

BurnCare sign-in screen with username and password fields

Step 1 · Sign in

Sign in with your approved account

Enter your username and password. "Remember this device" keeps you signed in for 30 days — leave it off on a shared workstation.

Patient Stats and Volume dashboard with case counters and the sidebar navigation

Step 2 · Dashboard

Land on Patient Stats & Volume

Your home screen: total cases, new-this-week, active jobs, and anything flagged "needs review." The sidebar (Upload Patient Data, AI Burn Analysis, Case History, Upload History) is how you get everywhere else.

AI Burn Analysis page with Add Patient and Existing Patient buttons

Step 3 · Start a case

Open "AI Burn Analysis" and pick a patient

Choose + Add Patient for a new patient, or Existing Patient to run another photo against a case already on file. (The same two buttons also appear under Upload Patient Data if you just want to archive a photo without running the AI pipeline.)

Register the patient form with name, age, gender, weight, height and vitals fields

Step 4 · Register the patient

Fill in the patient's medical details

Name, age, gender, weight and height are required — weight is what drives the Parkland fluid estimate later. Medical history, allergies and vitals are optional and can be filled in now or later. This record becomes the case every photo, note and result attaches to.

Workspace screen with the Upload, Background, Burn Detection and Burn Degree pipeline stages, ready to process a photo

Step 5 · Open the workspace

Choose a photo and a processing target

On the right, pick Single Image or Batch Photos, then Choose from Device (or reuse a prior upload). Under "Processing Target," select TBSA and Burn to run the full pipeline, then Run TBSA and Burn.

Results panel with the background-removed image, burn-skin-detected overlay, computed TBSA and fluid requirement

Step 6 · Read the result

Review the AI output and the TBSA & fluid estimate

The pipeline shows the background-removed photo next to the burn-region overlay, then "Computed Metrics" (burn % of image, burn % of visible skin, estimated TBSA) and the Parkland Fluid Requirement estimate below it. If a mask looks wrong, use Correct Foreground Mask or Correct Burn Mask right above the metrics — see the calibration walkthrough below.

Every button in the workspace panel

The right-hand panel in the workspace (step 5 above) looks a little different depending on Single Image or Batch Photos mode — here's both, and what every button in them does.

Single Image upload panel with Choose from Device, Use Existing Upload, Processing Target chips and Run TBSA and Burn

Single Image mode Choose from Device picks one photo. Use Existing Upload (only shown once a case already has an upload on file) reuses it instead of picking a new file. All four targets are selectable here, and the run button reads Run TBSA and Burn / Run Fluid Requirement / Run Burn Detection depending which chip is active.

Batch Photos upload panel with Choose Photos or Videos, only Detect Burn enabled among Processing Target chips, and Batch: Run Burn Detection

Batch Photos mode Choose Photos or Videos accepts multiple JPG/PNG/WEBP photos and MP4/MOV/AVI/MKV/WEBM videos at once (video frames are extracted automatically). TBSA and Burn and Fluid Requirement gray out here — both need the single-photo scale-lock dialog — so only Detect Burn is selectable, and the run button becomes Batch: Run Burn Detection.

Processing Target	Modes	What it does
TBSA and Burn	Single only	Full pipeline: background removal → burn detection → reference-scale TBSA estimate.
Fluid Requirement	Single only	Runs the same TBSA and Burn flow, then adds a 24-hour Parkland fluid estimate from the TBSA and the patient's recorded weight.
Detect Burn	Single + Batch	Burn-region detection only — background removal and the burn overlay, no TBSA number.
Detect Burn Degree	—	Disabled — still a prototype heuristic, not yet validated for use.
Treatment	—	Disabled — planned after decision-support recommendations are integrated.
3D Modeling	—	Disabled — coming soon.

Stop — only appears once a job is running; cancels it mid-pipeline.
Reset — clears the current selection and results, back to idle.
Correct Foreground Mask / Correct Burn Mask — below each result image; opens the point-and-polygon editor from the calibration walkthrough below on that image's mask.
Calculate TBSA — appears only after a mask correction is saved; recomputes the TBSA % from the corrected mask.
Export Referral PDF — opens the report modal to generate a PDF for this case/run.
Download Overlay — downloads the burn-overlay image file directly.

3. Calibrate the scale and correct a mask

Every photo needs a real-world scale to turn a pixel ratio into a TBSA percentage. This is the dialog behind Correct Foreground Mask / Correct Burn Mask, and it also covers what to do when no ruler or surgical tape is visible in the shot. (The photo itself is blurred below — these are from a real test capture — but every control and number is untouched.)

TBSA Calculation dialog step 1, entering a reference length and tapping two points on the photo

1. Set the reference length Tap the two ends of a known length — a ruler or surgical tape in the photo — enter its real length in cm, and lock the scale. No ruler in frame? Use Calculate TBSA Without Scale instead; you still get a burn/body percentage, just not an absolute cm² area.

TBSA Calculation result screen showing the automatic burn overlay and an estimated TBSA percentage

2. Get the automatic result The dialog returns an estimated TBSA from the AI's own foreground and burn masks, with "Show Calculation Steps" to see exactly how the number was derived.

Editing the foreground mask by drawing a correction polygon over the patient outline

3a. Correct the foreground mask If the patient outline is wrong, tap Add and click points around the correct region (edges draw automatically) — Close Polygon, repeat for more regions, then Preview and Apply Edit.

Editing the burn mask with the same point-and-polygon correction tool

3b. Correct the burn mask Same tool, same controls, now tracing the burn region itself instead of the patient outline.

TBSA result after correction, showing an updated estimated TBSA percentage

4. See the corrected estimate The TBSA recalculates from your corrected mask immediately, and that corrected figure is what carries through to the case's results panel and any exported report.

4. Find a case again

Case History

Every case uploaded by any signed-in staff member

Search by name, case ID, or uploader, and filter by Active / Needs Review / Closed. Open a case to resume analysis, export its PDF report, or delete it.

Upload History

Every archived photo, video and report

A record of everything uploaded under Upload Patient Data — photos, videos, day-progression images and referral reports — independent of whether the AI pipeline ever ran on them.

Every case, upload, and result stays tied to your account — only you and administrators can see your patients' photos and reports.

5. Using the Android app

The Android app mirrors the web portal workflow for bedside or ward use. It is currently a research-build debug APK (not yet listed on Google Play), so it is installed directly from the APK file rather than a store link.

Install the debug APK on an Android device (minimum Android 7.0 / API 24). Since it isn't from the Play Store, you'll need to allow installs from this source when prompted.
Open the app — it connects to the BurnCare server automatically on first launch.
Sign in with the same account used on the web portal.
Add or select a patient, then capture or upload a photo to run the same AI pipeline.
Review the TBSA result on-device — via the automatic estimate, or the manual Rule-of-Nines / Lund-Browder calculator — right after upload.

The app and web portal share one account and one case history, so a patient started on the ward in the app shows up immediately in the web portal, and vice versa.