YvanYin/VNL_Monocular_Depth_Prediction

Inpainting method used on dataset.

Closed this issue · 3 comments

mengtan00 commented

I notice that the vacuity in ground truth of NYU dataset has been filled. I'd like to know the method you use on inpainting dataset and whether it helps the performance of the network. Also, is the same method used to inpaint KITTI dataset? Thank you!

YvanYin commented

Hi, I used the inpainted depth for all the experiments, including the NYU and KITTI. Many previous methods also take the inpainted depth. I have never done a comparison of using the raw depth and inpainting depth.

mengtan00 commented

Thanks for your response. Could you please give me more details about the inpainting method you use, like papers which describe this method?

  • 👍1
YvanYin commented

Hi, sorry for replying late. I directly use the inpainting method provided by NYU dataset.