Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data

Benchmark Model Rank Results
monocular-depth-estimation-on-eth3dDepth Anything#4Delta < 1.25: 0.882absolute relative error: 0.0127
monocular-depth-estimation-on-kitti-eigenDepth Anything#9absolute relative error: 0.046RMSE: 1.896Sq Rel: 0.121
monocular-depth-estimation-on-nyu-depth-v2Depth Anything#10absolute relative error: 0.056RMSE: 0.206log 10: 0.024
semantic-segmentation-on-cityscapesDepth Anything#7Mean IoU (class): 84.8%
semantic-segmentation-on-cityscapes-valDepth Anything#8mIoU: 86.2