Error in extract_faces.py

[chinmaynehate]

• I first ran the index_dfdc.py file from which a .pkl file was generated
• Next, I ran the extract_faces file by providing the paths of source, videodf, facesfolder, facesdf and checkpoint directories. I got the following error. I haven't modified the code.
  

[CrohnEngineer]

Hey @chinmaynehate ,

thanks for checking out our code!
It seems you found a little bug in our pipeline: indeed, the extract_faces.py script looks for a field class in the video DataFrame created with index_*.py, but this field is created only for the FF++ dataset (if I'm not mistaken, it indicates the different categories of the manipulated videos, which is an info not available at the time we wrote the code for the DFDC dataset).
I'll fix it in the next committ, but it might take me a while as I might need to re-run the pipeline.
In the meantime, you could try this temporary "patch".
The line giving you this error is the 243 of extract_faces.py, where a DataFrame containing the info of the faces extracted from a video N is created. You should remove the fields 'class', 'quality' and 'source', and everything should run fine.
Something along this line:

face_dict = {'facepath': str(face_path.relative_to(facedestination_dir)), 'video': idx,
                    'label': record['label'], 'videosubject': subj_id, 'original': record['original']}

Let me know if that works, and thank you for spotting this!
Bests,

Edoardo

[CrohnEngineer]

Hey @chinmaynehate ,

with this last commit we should have fixed the bug you found.
If you haven't tried the patch I suggested you yet, try cloning the master branch and repeat the faces extraction process with the updated extract_faces.py.
Let us know if you have any other problem, and thanks again for the catch!
Bests,

Edoardo

[chinmaynehate]

Hey,
Thank you for your help!
I ran the fixed code of extract_faces.py from the master branch on some part of the dfdc dataset. Your latest commit has fixed the issue that I had raised earlier. I want to bring to your attention another small issue.

After running the latest extract_faces.py:

• As expected, all the extracted faces were stored in the path given in --facesfolder argument which is the faces output directory.
• But the Output DataFrame of faces which were supposed to be saved in the path given in --facesdf(for me it was icpr2020dfdc/facesdf folder). But the Output DataFrame( i.e. .pkl file) got saved in the icpr2020dfdc/ folder itself( which is not the folder given as --facesdf argument)

Thank you

[nicobonne]

Since the original problem seems to be solved, let's discuss this in a new issue