Using the pre-trained models

[ymohamed08]

Hello,
I just wanted to ask if you cant provide some instructions on how to use the pre-trained models on new videos to extract the action units.
I would like to feed in videos and output the action units. do you i need to retrain the model. what kind of format should the model be fed? Is there already a script you used to input video and output the action units?

[akiratsuraii]

bump,i knew there is a photo detection demo, but I also want to know how to feed videos.
I look forward to hearing any answer

[ymohamed08]

@akiratsuraii would you point me towards the demo please? Can't seem to find it either?

[akiratsuraii]

ME-GraphAU/OpenGraphAU/demo
description zone have the command to run with photos

[zapan-669]

I'm interested with Videos too

[Andreas-UI]

You might want to checkout my repo, I have implemented ME-GraphAU on a video in my project. No changes been made just minor refactor and uses their model to predict the frame when reading the video.

https://github.com/Andreas-UI/ME-GraphAU-Video

[zapan-669]

Appreciate brother, I’m going to test it soonOn May 2, 2024, at 8:52 PM, Andreas Susanto ***@***.***> wrote: You might want to checkout my repo, I have implemented ME-GraphAU on a video in my project. No changes been made just minor refactor and uses their model to predict the frame when reading the video. https://github.com/Andreas-UI/ME-GraphAU-Video —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: ***@***.***>