vinvino02 / GLPDepth

GLPDepth PyTorch Implementation: Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth
191 stars 44 forks source link

depth prediction is not absolute depth #42

Open gliehu opened 1 year ago

gliehu commented 1 year ago

Hello, I have a question about the predictions made by your network. I printed the predicted depth values and found that they do not correspond to the true distance. Does the network predict relative depth? The labels in the NYUv2 dataset are absolute depth, so why did the network predict relative depth? Could you please advise me on how to modify the network to predict true depth?