3D Wound Scanner
Android app for clinical wound assessment using ARCore and Raw Depth API. Captures and reconstructs 3D wound models without the need for depth sensors, enabling depth, area, and volume measurements on a wide range of Android devices.
Clinical Measurement Challenges
The app was originally developed for iOS and Android to assist clinicians in wound analysis using 3D reconstruction. While depth measurements were accurate on LiDAR- or ToF-enabled devices, the majority of Android devices without depth sensors produced poor results—limiting accessibility for clinical deployment.
The goal was to make the app widely compatible while preserving 3D measurement quality, especially on consumer Android devices without dedicated depth hardware.
Camera-Based 3D Reconstruction
To address poor depth results on Android devices without ToF sensors, a custom scanning pipeline was implemented using ARCore and the Raw Depth API. Instead of relying on a single frame, the system aggregates depth data from multiple frames, starting from a high-quality initial image.
Each frame is projected into a common coordinate system using device motion and camera intrinsics. The resulting dense point clouds capture wound shape and surface detail with significantly improved accuracy.
Testing included scanning arms and faces of volunteers, followed by visualisation in MeshLab to validate geometric and visual fidelity. The result is a robust reconstruction method optimised for human tissue, requiring no special hardware beyond a supported ARCore device.
This innovation expanded support to 100x more Android phones by eliminating the dependency on ToF hardware. Users are guided through the scanning process via a UI that ensures stable capture, enabling precise 3D wound measurement for clinical and personal health tracking.