You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In the phalp/visualize/postprocessor.py, in the run_lart() function, there is this code:
final_visuals_dic = joblib.load(phalp_pkl_path)
Where the function read the tracks generated by PHALP from videos. By my understanding, in the phalp_pkl_path file there are all features of all tracks in the target video. However, some lines of code later, the post_process() function is called, where it calls:
self.phalp_tracker.pose_predictor.smooth_tracks()
where all the tracks are smoothed.
What is the difference between smoothed tracks and non smoothed tracks? Can I use the features of non smoothed tracks in my application?
Thanks
The text was updated successfully, but these errors were encountered:
Right now, the smooth_tracks() function only applies action recognition, it does not smooth the tracks. Track smoothing requires a different checkpoint and a different config. Depending on your application you can directly use the features of the smpl parameters or the features for the action recognition if it more related to action understanding.
Hi,
In the phalp/visualize/postprocessor.py, in the run_lart() function, there is this code:
final_visuals_dic = joblib.load(phalp_pkl_path)
Where the function read the tracks generated by PHALP from videos. By my understanding, in the phalp_pkl_path file there are all features of all tracks in the target video. However, some lines of code later, the post_process() function is called, where it calls:
self.phalp_tracker.pose_predictor.smooth_tracks()
where all the tracks are smoothed.
What is the difference between smoothed tracks and non smoothed tracks? Can I use the features of non smoothed tracks in my application?
Thanks
The text was updated successfully, but these errors were encountered: