Automated Surgical Guide Generation for Microtia Reconstruction using Artificial Intelligence and Computer Vision.
AMGG is a computer program designed to democratize high-precision surgical planning for Microtia reconstruction. By leveraging AI and Computer Vision, this tool converts standard 2D smartphone photos of a healthy ear into a biomechanically accurate 3D surgical guide (STL format) ready for 3D printing.
This project addresses the limitations of traditional manual 2D tracing methods (which lack depth) and the high cost/radiation risks of CT scans, providing a low-cost, accessible solution for surgeons in developing regions.
AI-Powered Segmentation: Utilizes Meta SAM (Segment Anything Model) and YOLOv8 to automatically isolate the healthy ear from complex backgrounds without manual tracing.
No CT Scan Required: Generates 3D topology from a single 2D image, reducing cost and radiation exposure for pediatric patients.
Nagata Technique Compliance: Automatically applies biomechanical standards:
Base Plate: 2.0 mm thickness (Simulating Ribs 6 & 7).
Helical Projection: 5.0 mm total height (Simulating Rib 8).
Contralateral Mirroring: Automatically flips the healthy ear geometry to create a guide for the affected side (supporting the 60-80% of cases that are Right-sided Microtia).
Automated Suture Windows: Generates 1.2mm x 2.7mm anchor holes for surgical needle guidance.
Watertight STL Output: Produces files ready for standard SLA/Resin 3D printers.
Language: Python 3.10
Segmentation: Meta SAM (Segment Anything Model), Ultralytics YOLOv8
Image Processing: OpenCV (cv2)
3D Mesh Generation: numpy-stl
Math/Logic: NumPy
Clone the repository
git clone [https://github.com/TrisesaSadewa/AutoMicrotiaGuideGenerator_AMGG.git](https://github.com/TrisesaSadewa/AutoMicrotiaGuideGenerator_AMGG.git)
cd AutoMicrotiaGuideGenerator_AMGG
Create a virtual environment (Recommended)
python -m venv venv source venv/bin/activate
# On Windows use `venv\Scripts\activate
Install dependencies
pip install -r requirements.txt