From ftm
- CataRT ... data-driven concatenative sound synthesis in real time, based on unit selection from large databases
- Gesture Follower ... gesture recognition and online-alignement (following) using HMMs
For simpler examples using FTM see here.