Skip to content

README.md

Latest commit

 

History

History
24 lines (19 loc) · 1.53 KB

README.md

File metadata and controls

24 lines (19 loc) · 1.53 KB

AxisAlign: Precision Surgical Planning Suite 🦵

AxisAlign is a "Human-in-the-Loop" (HITL) computer vision application designed for orthopedic surgeons to perform pre-operative planning for osteotomies (HTO/DFO).

🚀 Key Features

  • Spatial Calibration: Corrects for radiographic magnification using fiducial markers.
  • Dynamic Simulation: Uses affine transformation matrices to rotate anatomical segments and predict Post-Op HKA (Hip-Knee-Ankle) alignment.
  • Precision UX: Integrated ROI (Region of Interest) Loupe to ensure sub-millimeter landmarking accuracy.
  • Patient-Specific Morphometrics: Calculates required opening wedge height based on real-time Euclidean distance measurements of the bone.

🛠️ Technical Stack

  • Frontend: Streamlit (Reactive State Management)
  • Geometry: NumPy & Python Math (Trigonometric Identities & Vector Analysis)
  • Computer Vision UI: Streamlit-Drawable-Canvas
  • Imaging: Pillow (ROI Extraction)
  • Visualization: Matplotlib (Vector Overlays)

📐 The Math Behind the Medicine

The application calculates the Opening Wedge Height ($H$) using the Tangent Identity: $$H = W \times \tan(\theta)$$ Where $W$ is the measured bone width and $\theta$ is the planned correction angle.

👨‍💻 AI Engineer Perspective

This tool serves as a high-fidelity data labeling interface. The manual landmarks acquired provide the Ground Truth dataset required to train future Deep Learning models (CNNs) for automated anatomical landmark detection and automated surgical planning.