search

sconlyshootery / FeatDepth

This is the offical codes for the methods described in the "Feature-metric Loss for Self-supervised Learning of Depth and Egomotion".
MIT License
247 stars 28 forks source link

inference results in your Table 2 in your paper #45

Closed sunnyHelen closed 3 years ago

sunnyHelen commented 3 years ago

Hi, thanks for your great work. I am a little confused about the test results reported in your Table 2. Why the results of DORN and BTS are much worse than that in their papers? Are the test splits used in their papers are different with yours and monodepth2? Please help me to figure it out. Looking forward to your reply~ Thanks a lot~

sconlyshootery commented 3 years ago

The number is from Table 2 in BTS paper, which is its version 4. The link is https://arxiv.org/pdf/1907.10326v4.pdf. Since BTS paper has been through 5 versions, and its content changes a lot, so you must refer to its old version to find the number.

sunnyHelen commented 3 years ago

In BTS's latest version there is a huge difference regarding to the test results of DORN and their own results. Maybe there's something wrong in the old version. It's too confusing.

sconlyshootery commented 3 years ago

When I writed this paper, BTS is at version 4, I have no idea about the upcoming change. Since the BTS' code is released, there is a way to test its authenticity.

sunnyHelen commented 3 years ago

I don't know. Maybe there are different test splits or using different ground truth version. Too confusing~ Most paper cited DORN's results as 0.072, but in some papers it is a little worse.

sconlyshootery commented 3 years ago

I will close this issue, you can reopen it if you have further question.