After holding Unsupervised human video generation subject, I decided to solve landmark detection problems with unsupervised or self-supervised method
This is my proposal for solving this topic
Current status is just for survey previous research and deriving ideas to solve and contribute, there are some topics based on previous research
1) Find landmarks from videos in unsupervised / self-supervised method
2) Reshape objects from detected landmarks
3) Reformat 3D shape from 2D images with detected landmarks
Related research
- Unsupervised landmark detection
- Landmark detection with generative model
- Unsupervised Part Representation with optical flows