jancio / deep-virtual-rapport-agent Goto Github PK

Hello.

Thanks for your code. I'm currently is trying to evaluate your head gesture detection code on the webcam. Is it possible? Could you give me some advices where to start?

Main question is, i guess, what kind of data should be given to your model? AS i see in evaluate_hgd, you are loading npz file (which i dont have of course) and using X-part of it as an input. Since i don't have the datasets, i can't check what exactly lies in npz files. Is it just a frame or series of frames?

THanks in advance for the answer.

In file head_gesture_detector/hgd_annotate_frames.ipynb, there are 2 rows using function unique_fusion. However it's undefined.

Do you have any idea about this function?

new_df['head_gesture_unique_NS'] = new_df.apply(unique_fusion, args=('_NS',), axis=1)
new_df['head_gesture_unique'] = new_df.apply(unique_fusion, args=('',), axis=1)