search

nianticlabs / manydepth

[CVPR 2021] Self-supervised depth estimation from short sequences
Other
618 stars 83 forks source link

The dataset GT-Calculation #71

Closed wangjiyuan9 closed 1 year ago

wangjiyuan9 commented 1 year ago

Dear: First of all, thank you for your amazing paper ！ I encountered a problem when I read your public code： I‘d like to calculate the gt_depth of the image for the training part (not all). Do you still have the Calculation code ? Thank you for your time and help!

JamieWatson683 commented 1 year ago

Hi sorry for the delay - Cityscapes provide intrinsics, camera baseline and disparity map, so to recover the depth map you simply do: depth = focal length [in pixels] x baseline [in metres] / disparity