@hoangtnm Only the metric model's (model finetuned on kitti dataset) output is metric, meaning in meters. The other models output relative disparity (inverse depth) upto an unknown scale and unknown shift which needs to be computed using GT for example

LiheYoung / Depth-Anything

[CVPR 2024] Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data. Foundation Model for Monocular Depth Estimation

https://depth-anything.github.io

Apache License 2.0

6.95k stars 532 forks source link

@hoangtnm Only the metric model's (model finetuned on kitti dataset) output is metric, meaning in meters. The other models output relative disparity (inverse depth) upto an unknown scale and unknown shift which needs to be computed using GT for example #233

Open ysz-leo opened 1 week ago

ysz-leo commented 1 week ago

          @hoangtnm Only the metric model's (model finetuned on kitti dataset) output is metric, meaning in meters. The other models output relative disparity (inverse depth) upto an unknown scale and unknown shift which needs to be computed using GT for example

Originally posted by @kishore-greddy in https://github.com/LiheYoung/Depth-Anything/issues/72#issuecomment-1930222283

ysz-leo commented 1 week ago

What is the difference between the true depth calculated by relative depth and the true depth output by fine-tuning metric model? Is the result of fine-tuning the model more accurate than the calculated result?