Please enable JavaScript.
Coggle requires JavaScript to display documents.
Human Action Synthesis - Coggle Diagram
Human Action
Synthesis
Datasets
Descriptions
Depth sensing cameras vs
pose estimation methods
What is an ideal dataset
for human motion synthesis?
Availability of datasets
Class to motion
Language to motion
Human locomotion
Multi person interaction
Pros/cons with the
available datasets?
Evaluation Methods
non-Generative models
Generative models
Approaches
Traditional Motion
predictions Motion prediction,
motion tweening etc.
Motion Retargeting
Conditioned on Class labels
Conditioned on natural language
Generative models
Embedding methods(CNN vs RNN vs
GCN vs attention based methods)
Introduction (Optinal
in the mind-map)
What is Human
motion synthesis
Different types of
human motion synthesis
Why Human
motion synthesis?
Why it is difficult
Representation
RGB based representations
Skeleton based representations
3D joint positions
Joint Rotation based
representation
Other
Future Research
Direcctions
Conclusion
Key research problems
in this fields