Dashboard
Standardize the image.
Preserve the anatomy.
V5 adds live camera capture, active acquisition coaching, baseline reproduction scoring, browser-native face detection when available, model-ready landmark/pose hooks, longitudinal sessions, and auditable geometry-safe comparison.
Recent cases
Patient-linked longitudinal photo records
Guided Capture
Choose a patient and case to begin.
V5 metrics are computed from the actual image. Native browser face detection is used only when available; backdrop isolation remains a controlled-background heuristic until the production segmentation model is integrated.
Case library
Longitudinal sessions with baseline and follow-up comparison.
Before / after PNG
1800 × 900 comparison with geometry-safe alignment.
Clinical case manifest
Patient ID, protocol, sessions, image slots, integrity metadata, and alignment values.
Quality + reproduction report
Per-image acquisition QA, detector status, and baseline reproduction scores.
Source inventory
Count original and processed image records stored on this device.
Default background
Black portfolio, white clean, neutral gray clinical.
Local persistence
Metadata uses localStorage and photos use IndexedDB. This is a device-local prototype, not hosted clinical infrastructure.
Production security boundary
Production requires authenticated tenancy, encryption, access controls, audit events, retention/deletion controls, secure backups, and appropriate HIPAA contracting where applicable.
Production model layer
Dedicated neural person segmentation, 468-point facial landmarks, yaw/pitch/roll estimation, camera-distance calibration, occlusion detection, and automatic protocol compliance. V5 exposes the adapter points without pretending those production models are already bundled.