Closed sontung closed 5 years ago
Hi,
we are just resizing images to 224x224 resolution. Depth maps and labels are resized using nearest-neighbor interpolation and depth maps are normalized to [0, 255].
Do you use any colorization to fill in missing values for depth maps, like in the NYU dataset?
Yes, we use NYUv2 kit for it.
Is there a Python code for it? Or only MATLAB?
No it is only in MATLAB.
Hi, Can you also publish the code for preprocessing the SUNRGB dataset? I have trained your model on SUNRGB and now I want to test the model on my own images. Thanks.