About the implementation of normalisation

[mucunzhuzhu]

Hi author, I found that you performed normalisation operations on the true and predicted joints before calculating the FID and Diversity, but I found that the normalisation you implemented is not the same as the normalisation method that I have seen in the code of other papers.

You are normalising the true and predicted joints separately, but the other papers are calculating the mean and variance of the true joints first, and then the predicted joints are normalised using the mean and variance of the true joints. ( The first figure shows your implementation of normalisation, the second figure shows the implementation of normalisation in other papers. )

The results calculated by these two methods are not the same, can I ask you the difference between these two normalisation methods?