Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

the motion features Fk are taken as a concatenation of the output features and the downsampled warping flow #2

Open
chengzihua opened this issue Nov 12, 2022 · 2 comments

Comments

@chengzihua
Copy link

hello,output features shape is b,128,H/8,W/8, downsampled warping flow shape is b,2,H/8,W/8 ,but Fk shape is b,128,H/8,W/8.What is the connection method? how to get the Fk?

@fh2019ustc
Copy link
Owner

Hi~ Thanks for your attention to our work!
The shape b,128,H/8,W/8 can be first reshaped to b, 882,H/8,W/8 and then permuted to b, 2, H/8, 8, W/8, 8.
In this way, we can get a flow with shape b * 2 * H * W.
Hope this can help.

@chengzihua
Copy link
Author

Hi~ Thanks for your attention to our work! The shape b,128,H/8,W/8 can be first reshaped to b, 8_8_2,H/8,W/8 and then permuted to b, 2, H/8, 8, W/8, 8. In this way, we can get a flow with shape b * 2 * H * W. Hope this can help.

about motion feature generator, How is the output feature of the convolutional layer zθ merged with downsampled warping flow fm(k-1),and get the motion features Fk?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants