Use data to represent certain pose is defenitely reasonable. The ideal data would be representative yet reasonably small number of features. But how do we capture those features and how shoud we decide how much to use for one pose (simple ones and complex ones should demand differently)?
Use data to represent certain pose is defenitely reasonable. The ideal data would be representative yet reasonably small number of features. But how do we capture those features and how shoud we decide how much to use for one pose (simple ones and complex ones should demand differently)?