very nice presentation! What I dont understand is the training part. What incentive is there for the network to look at the depth map from a specific point of view? Do you in practice view the colored 3D image from a known camera position that corresponds to the target image camera position?
Cool approach.
are you just raymarching at 7:23 ?
very nice presentation! What I dont understand is the training part. What incentive is there for the network to look at the depth map from a specific point of view? Do you in practice view the colored 3D image from a known camera position that corresponds to the target image camera position?
Can I run the Github Code without Nvidia graphics card?
great!!!!