Closed weberhen closed 3 years ago
I obtain the depth value by sending email to the author of 'deep parametric indoor lighting estimation'. Then the depth map can be processed similarly to obtain the ground truth for regression.
And you can also delete the depth part directly, and run the model, which is equivalent to EMLight.
I see, thanks!
And from the depth map, you actually only need the depth on the (128) anchor points right?
I see, thanks!
And from the depth map, you actually only need the depth on the (128) anchor points right?
In fact, we average the depth of pixels assigned to an anchor points, but it's not a big difference.
Hi,
congrats for the great work!
I'm trying to run the training code and I'm getting the following error:
I generated the dataset using
distribution_representation.py
, but I could not find anywhere the depth being added to the list.Thanks for the help!