The face cannot be obtained.

[GNIJZ]

I used the sample video emotion detection and got the following error:
Dropped 20 frames because no landmarks detected by mediapipe in the frame.
No face detected in the provided video
No face detected in the video, get_facial_attributes fails.
Since no face is detected, can't save the file

[Ration-Lee]

I got the same error, did you figure that out?

[Ration-Lee]

I reinstalled the conda environment in Linux according to the commands in the README, and that solved the issue.

[quantumdotsss]

927096

I reinstalled the conda environment in Linux according to the commands in the README, and that solved the issue.

Have you tried your own video?

[quantumdotsss]

I used the sample video emotion detection and got the following error: Dropped 20 frames because no landmarks detected by mediapipe in the frame. No face detected in the provided video No face detected in the video, get_facial_attributes fails. Since no face is detected, can't save the file

downgradeyour Pillow

pip install Pillow==9.2.0 is all you need