search

w2kun / RNW

This is the official repository for Regularizing Nighttime Weirdness: Efficient Self-supervised Monocular Depth Estimation in the Dark (ICCV 2021).
MIT License
84 stars 13 forks source link

about pre-process depth map #7

Open mzy97 opened 2 years ago

mzy97 commented 2 years ago

Hi, about processing depth, I have two questions

  1. In script above, how _RAW_TEST_TIMESTAMPS generated, and what chunk is used for ?
  2. How to choose the _TIME_SPAN, when the value is large, some occluded pixel will appear when project accumulated point cloud to image, like the visualization I ran below ( the red house suppose to hide behind the wall, but large _TIME_SPAN will make it appear, and it is a invalid GT depth image) image @w2kun

Originally posted by @mzy97 in https://github.com/w2kun/RNW/issues/2#issuecomment-1001391207

mzy97 commented 2 years ago

And It will be a great help if you could provide the already processed data used in this paper. @w2kun

Originally posted by @mzy97 in https://github.com/w2kun/RNW/issues/2#issuecomment-1001406960

agenthong commented 2 years ago

Hi, @mzy97!

Do you know what 'raw test timestamps' is? I can't find the corresponding file.