search

TadasBaltrusaitis / OpenFace

OpenFace – a state-of-the art tool intended for facial landmark detection, head pose estimation, facial action unit recognition, and eye-gaze estimation.
Other
6.71k stars 1.82k forks source link

[Emergency] Question about how to extract specific feature via algebr/openface:latest #1021

Open EricKani opened 1 year ago

EricKani commented 1 year ago

Sorry to bother you, it's really urgent for me to solve this issue! Thanks in advance!

I want to extract 3d facial landmarks and gaze vector via OpenFace from my own videos(image sequences). And I have implemented the recommended docker Image (algebr/openface:latest), and it worked very well, which can extract all the features out of the face and save the outputs and visualization into the disk.

And here is the problem, my study project is very urgent, and I only need the output(csv file) of 3d facial landmarks and gaze vector features. Saving all feature and visualization to the disk is very time-consuming for me. How can I speed up this process using this docker environment. I found the executable file seems is a compile-finished file.

Thank you very much!

Pengbo

EricKani commented 1 year ago

I found the "build/bin/FaceLandmarkVidMulti -f /home/pbliu/openface_workspace/230_1.mp4" can process video file directly. But it doesn't output features for all frames, e.g., it output for frame_1,2,3,4,50,51,52,60,200,... . And it doesn't output gaze feature. Are there some problems in my terminal command?

Best, Pengbo