Hi all,
After the unfortunate deletion of some conversations, I wanted to repost my recently published video tutorials on the practical steps to create a statistical shape model.
The series is available on youtube:
https://www.youtube.com/watch?v=D4W2sotakYk&list=PLPmOaPsuoeltjdVRHqDXfN5tcUxZRrQ7j In the video series I cover everything from:
- Shape model data requirements (aligned and in point-correspondence)
- Different methods for rigid alignment
- How to design the template/reference mesh for the shape model
- How to make the template deformable with kernels
- How to perform fitting (non-rigid registration) using the Gaussian Process Morphable Model (GPMM) framework and Gaussian Process Regression (GPR)
And then extra videos on:
- Model evaluation (specificity, generalization, compactness)
- Model visualization
Let me know what you think. And also let me know if something is missing which I should make a video/tutorial about.
Best,
Dennis