D-X-Y/landmark-detection

Pretrained Model of SRT?

CodingMice opened this issue · 3 comments

Is there any pretrained models of SRT? I am looking forward to test it.

D-X-Y commented

Yes, we have. I need some time to download them from the server and then upload to Google Drive.

I'm anxious to see how the increase in temporal precision impacts video data collected from mobile devices.
My work uses a landmark detection system to provide a framework for specific types of behavioral classification.
Currently, I'm in the process of writing my dissertation, and I'd love to be able to include a comparison and future perspective using your model's localization.
I realize you're quite busy, so I definitely don't want to be a pest, but it would be excellent if you could find the time to upload these models!

D-X-Y commented

@Tetomach Thanks for your interest. I just uploaded some pre-trained models at here: https://drive.google.com/drive/folders/1uGRf74VecXwNzwMzxxlkVfvreIwGoOHa?usp=sharing

The models included the regression model pre-trained on 300-W, 300-W with (SBR augmented) 300VW, 300-W with (SBR augmented) VOX, 300-W with (SBR augmented) PanopticFace, and 300-W with (SRT augmented) PanopticFace.

Having fun!