Skip to content

Releases: DebadityaHait/WoundDoc

WoundDoc v1.0.0 — First Public Release (Android APK)

Choose a tag to compare

@DebadityaHait DebadityaHait released this 23 Mar 22:14

🎉 WoundDoc v1.0.0 is here!

Photograph once. Track healing forever.

The first official release of WoundDoc — a production-grade AI-powered mobile platform for clinicians to monitor chronic wounds and diabetic foot ulcers with calibrated measurements and intelligent insights.

✨ Key Features

  • 16-class AI wound classification (EfficientNet-B0 backbone, 92.33% accuracy on Medetec dataset)
  • 30+ tissue segmentation with color-coded overlays (necrotic, granulation, slough, epithelial, etc.)
  • ArUco-calibrated size detection — perspective-corrected area in cm² (robust to angle & lighting)
  • Longitudinal tracking — sparkline trends + stacked tissue history charts
  • On-demand Gemini AI clinical analysis — summary, key findings, healing trajectory (improving/stable/worsening), concerns & recommendations
  • 3-tab image comparison — Original | Annotated | Rectified (bird’s-eye view)
  • Fully offline-first — all data & images stored locally (no cloud required)

Research-Backed Technology

Built on peer-reviewed methods:

  • Semi-supervised learning pipeline from DFUTissueSegNet (DSC improved to 87.64%)
  • EfficientNet-B0 + ReLU from Eff-ReLU-Net
  • Phone-based workflow inspired by DFUCare (F1 0.80 localization, 94.81% ischemia accuracy)

Installation

  1. Download WoundDoc-v1.0.0.apk below
  2. Install on any Android device (Android 8.0+)
  3. Open with Expo Go or directly (allow "Unknown sources" if needed)

Screenshots

Full gallery available in the README


⚠️ Medical Disclaimer
This is a research and demonstration tool only. All AI outputs are advisory. Not a substitute for professional clinical judgment. Use at your own risk. Always verify with a qualified clinician.


First release — March 2026