Thank you for providing this fantastic new dataset!
In the paper, the authors mentioned "We use 7/8 of the images for training and 1/8 of the images for testing" while evaluating the sota methods on the dl3dv benchmark. How are the 7/8 image sampled? Are they sampled randomly? Or uniformly?
Hi,
Thank you for providing this fantastic new dataset!
In the paper, the authors mentioned "We use 7/8 of the images for training and 1/8 of the images for testing" while evaluating the sota methods on the dl3dv benchmark. How are the 7/8 image sampled? Are they sampled randomly? Or uniformly?
Thank you, Ziwen