dontLoveBugs / DORN_pytorch

PyTorch implementation of Deep Ordinal Regression Network for Monocular Depth Estimation
299 stars 67 forks source link

Question about paper #16

Closed abeyang00 closed 5 years ago

abeyang00 commented 5 years ago

what is the main reason for dividing channel into 2 sections? I can't seem to understand why they would use 142 channels and then divide into 2 sections of 71 channels.