OrthoShapeLab is an interactive toolkit for 3D Statistical Shape Modeling (SSM) and visualization of anatomical shape variation.
It allows users to load a set of aligned meshes and explore shape variation using Principal Component Analysis (PCA) through an interactive 3D viewer.
The project is designed for medical imaging, orthopedics, and computational anatomy research.
Example of the interactive PCA viewer used to explore variation in 3D anatomical meshes.
- Interactive 3D mesh visualization
- Principal Component Analysis (PCA) for statistical shape modeling
- Real-time shape deformation using sliders
- Display variance explained for each PCA mode
- Random shape generation
- Reset to mean shape
- Load meshes directly from a folder
- Support for STL and PLY mesh formats
Future versions will include:
- Principal Geodesic Analysis (PGA)
- Multi-bone statistical shape models
- Additional preprocessing tools
Install using pip:
pip install orthoshapelabgit clone https://github.com/prabhat-twr/orthoshapelab.git
cd orthoshapelab
pip install -e .Run the application from the terminal:
orthoshapelabThe viewer will open and allow you to:
- Adjust PCA sliders
- Visualize shape variation
- Generate random shapes
- Reset to the mean shape
The meshes must satisfy the following conditions:
- Meshes should be rigidly aligned
- Meshes must have point-to-point correspondence
- All meshes must have the same number of vertices
.stl.ply
OrthoShapeLab has been developed and tested on Ubuntu Linux.
- OS: Ubuntu 20.04 / 22.04
- Python: 3.9+
- Rendering: Open3D GUI backend
Windows support has not yet been fully tested.
Because the viewer relies on the Open3D rendering engine, some systems may encounter OpenGL or WGL related issues depending on GPU drivers and graphics configuration.
Windows users may still attempt installation:
pip install orthoshapelabIf problems occur, running the software on Ubuntu Linux is recommended.
- Python
- Open3D
- NumPy
- Scikit-learn
OrthoShapeLab can be used for:
- Femur statistical shape modeling
- Shoulder joint modeling (scapula + humerus)
- Orthopedic implant design
- Anatomical variability studies
- Medical imaging research
Planned improvements:
- PGA implementation
- Multi-bone SSM support
- Landmark-based modeling
- Improved visualization tools
- Dataset preprocessing utilities
Prabhat Tiwari
Developer working in:
- Computer Vision
- Medical Imaging
- Statistical Shape Modeling
GitHub:
https://github.com/prabhat-twr
