AcT icon indicating copy to clipboard operation
AcT copied to clipboard

Published 20 hours ago • verified publisher icon PIC4SeR

Inference

Open Cavan1Ed1s0n opened this issue 10 months ago • 3 comments

Could you show me how to infer your pretrained with video?

Cavan1Ed1s0n avatar May 01 '24 09:05 Cavan1Ed1s0n

Hi @Cavan1Ed1s0n, We used the PoseNet inference demo code by Google Coral and added AcT in cascade with the pose extraction model. You can adapt to your needs using a CPU. I hope this helps!

simoneangarano avatar May 01 '24 14:05 simoneangarano

I have read and run your code in this repository. It seems that you only test accuracy from extracted feature, right? Could you please provide complete pipeline from input video -> keypoints detector -> transformer -> action recognition -> result visualization.

Cavan1Ed1s0n avatar May 02 '24 01:05 Cavan1Ed1s0n

Hi @Cavan1Ed1s0n ! I will try to do it as soon as possible, but I cannot give you an ETA now, sorry!

simoneangarano avatar May 13 '24 15:05 simoneangarano