search

Tanveer81 / BoxVOS

Other
4 stars 1 forks source link

Questions about Semi-VOS #2

Open 9p15p opened 1 year ago

9p15p commented 1 year ago

Hi, Tanveer,

I have two questions.

  1. When I evaluate or train the model. It has not use any reference information(e.g. mask or box). How does the model know what to segment for Semi-VOS. https://github.com/Tanveer81/BoxVOS/blob/521e534001630e4a104b4d509e8cd54ce36d765c/adet/modeling/condinst/condinst.py#L194

  2. Are there any codes for user to create the results for evaluation on DAVIS-2017(semi) or YoutubeVOS(semi). It needs to assigning colors according to reference information. The demo.py doesn't seem to have that function.

Tanveer81 commented 1 year ago

The model only predicts objects per frame. We use PRemVOS to create the results for evaluation. We employ both optical flow and object re-identification as PReMVOS. We needed to adapt the code base of PReMVOS but did not change any underlying design. We will add the modified code and instructions soon.

9p15p commented 1 year ago

Hi, Tanveer, Could you please offer the Precomputed Results. Thank you.