A data generation pipeline for creating semi-realistic synthetic multi-object videos with rich annotations such as instance segmentation masks, depth maps, and optical flow.
I am having trouble understanding how the forward and backward flows for MOVi-A scenes are rendered. Specifically, I would like to know why flows corresponding to shadows are not generated (see attached gifs). Are the forward and backward flows motion fields and not optical flows?
Hi,
I am having trouble understanding how the forward and backward flows for MOVi-A scenes are rendered. Specifically, I would like to know why flows corresponding to shadows are not generated (see attached gifs). Are the forward and backward flows motion fields and not optical flows?