How can FasterSeg be used to perform inference on a video?

[Gaussianer]

Hello,

how can I run the inference of FasterSeg on a video?

I saw that you created the gif in the readme using a script. Did you use the tester.py script for this? If so, how do you use this script?

Best regards

[chenwydj]

Hi @Gaussianer,

Thank you for your interest in our work!

You can generate the RBG predictions by setting show_prediction=True in train.py.

For the gif, I just made predictions to all frames of the video, and concatenated them into a gif.

[Gaussianer]

Thanks @chenwydj for your reply!

I need a script that does the inference on a video. Do I have to write my own script for this?

Best regards

[chenwydj]

Hi @Gaussianer,

Sorry I did not save that command. As I remember I used ffmpeg command line tool to chunk the video into frames and to make the gif.

[Gaussianer]

Hi @chenwydj ,

All right, thanks. I would like to apply the inference of FasterSeg to a video camera in real time. So I think I will have to write my own script for this?

The gif does not really play a role in my project. I just thought that you used a video for this. That would have saved me a little work.

[chenwydj]

Hi @Gaussianer,

You are right, and unfortunately, we did not prepare scripts for videos.

[Gaussianer]

Hi @chenwydj,

All right. Thanks for your quick answers :-)