How to apply the grid sampling strategy (GS) for eq (9)
Hi, thanks for your interesting work.
Could you provide more details about the process of eq (9)? Specifically, how to apply the grid sampling strategy (GS) here?
Thanks a lot.
For the grid-sample strategy, we first uniformly sample the m rays, and the m rays match the m corresponding pixel colors. (https://github.com/YifYang993/CR-NeRF-PyTorch/blob/bb0828915543dd1230544d933762230808198403/datasets/phototourism_mask_grid_sample.py#L242-L266)
here we get the index of the m pixels https://github.com/YifYang993/CR-NeRF-PyTorch/blob/bb0828915543dd1230544d933762230808198403/datasets/phototourism_mask_grid_sample.py#L272
we then use the index to do the same grid sample to the estimated visibility mask on transient objects https://github.com/YifYang993/CR-NeRF-PyTorch/blob/bb0828915543dd1230544d933762230808198403/train_mask_grid_sample.py#L175