search

DegardinBruno / human-self-learning-anomaly

Code for the paper "Human Activity Analysis: Iterative Weak/Self-Supervised Learning Frameworks for Detecting Abnormal Events", IJCB 2020
Other
42 stars 13 forks source link

normalize_notes is not working as expected #11

Open bashayerAlsalman opened 2 years ago

bashayerAlsalman commented 2 years ago

Hello Bruno Degardin, Really appreciate the amazing effort in the paper and in the code as well.

I tried to run the code but I faced an issue (I am just beginner in the AI field, pardon my silly questions).

We I ran: 1- video_frames -> there was no modification needed 2- normalize_videos -> I had to change frames path, where will read and partition the images. 3- normalize_notes -> I did understand what is the output, because when I run this command python3 utils/normalize_notes.py --root_csv data/UBI_FIGHTS/annotation --dest_csv data/annotation --fps 30 --duration 16

Could not run, because it is asking for annotation csv files which I did not understand from where to get them, could you please explain it to me?

Thank you a lot

DegardinBruno commented 2 years ago

Hi @bashayerAlsalman !

normalize_notes.py will normalize each video annotation (corresponding video csv file) into fixed length annotations. In your case fps x duration, which will be 480 frames per video. So it takes as input the path folder to the annotation files, you can use UBI-Fights for that. And will generate multiple annotation files for each video.