1 paper
Cho-Ying Wu, Jialiang Wang, Michael Hall +2
The majority of prior monocular depth estimation methods without groundtruth depth guidance focus on driving scenarios. We show that such methods generalize poorly to unseen comple…