how to get the figure 6?

Question

how to get the figure 6?

Closed this issue 2 years ago · 6 comments

Hi,
Recently i'm trying to get the correspondences of the rest frames using the first frame of my own video, but the results are bad.
Could you open source the codes that can get the correspondences like figure6 in your paper?
Thanks in advance.

Liesy commented 2 years ago

thx

Answer 1 · 2022-05-07T18:18:37.000Z

You can use the code here to get started: https://github.com/hkchengrex/Mask-Propagation/blob/main/try_correspondence.py

Answer 2 · 2022-05-07T18:21:20.000Z

The logic is the same but you would have to change some code to fit it to STCN. I look at the attention map and manually put the symbols in Figure 6.

Answer 3 · 2022-05-08T03:53:45.000Z

thank you very much! I'll try it later.

Answer 4 · 2022-05-22T09:05:18.000Z

Does the MaskRGBEncoder have the same effect as RGBEncoder in corr_network.py#L40 when getting the memory key in try_correspondence.py#L60 ?

Answer 5 · 2022-05-22T17:56:42.000Z

Memory keys are not symmetric in MiVOS so the keys you get from the two encoders are different. This problem does not exist in STCN.