search

raphaelmemmesheimer / skeleton-dml

Skeleton-DML: Deep Metric Learning for Skeleton-Based One-Shot Action Recognition
24 stars 6 forks source link

How to convert skeleton sequences into image representations? #6

Closed charlietie closed 2 years ago

charlietie commented 2 years ago

Hi: Your work is very cool!

But I want to know how to convert the raw skeleton sequence into image representations. As there is no relevant code. If possible, I want to use this approach on other skeleton datasets, e.g. kinetics skeleton.

Could you provide the detailed code! I will appreciate for your help!

Thanks!

raphaelmemmesheimer commented 2 years ago

Hey @charlietie,

example code for the image representation generation is here:

https://github.com/raphaelmemmesheimer/skeleton-dml/tree/main/representation

You slightly need to adapt it to the kinetics skeletons.

charlietie commented 2 years ago

Thanks! I will try to adapt it.

guedesopaulo commented 1 year ago

Hi, good job man. This conversion code of yours also applies to the NTU RGB+D 120 dataset? What I really wanted to know is how to get to the Precalculated Representations that you shared in this github.

raphaelmemmesheimer commented 1 year ago

Yes, the skeletons are very similar. The same code can be used for the NTU-120 dataset.