AI Paper Atlas

Research map / 3D vision / Depth and stereo

Self-Supervised Monocular Depth Estimation for All Day Images Using Domain Separation

Lina Liu, Xibin Song, Mengmeng Wang, Yong Liu, Liangjun Zhang

ICCV 2021 · 3D vision › Depth and stereo · approach: Dense depth prediction and completion · 91 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use GAN-generated paired night images instead of unpaired real day/night images (ablation)domain-separated network · Oxford RobotCar night (40m)Abs Rel: 0.573 -> 0.429Improved
use paired instead of unpaired images (ablation)domain-separated network · Oxford RobotCar day (40m)Abs Rel: 0.280 -> 0.117Improved
add reconstruction loss (ablation)P · Oxford RobotCar night (40m)Abs Rel: 0.429 -> 0.357Improved
add reconstruction loss (ablation)P · Oxford RobotCar day (40m)Abs Rel: 0.117 -> 0.131Hurt
add direct feature orthogonality loss L_f (ablation)PR · Oxford RobotCar night (40m)Abs Rel: 0.357 -> 0.251Improved
add direct feature orthogonality loss L_f (ablation)PR · Oxford RobotCar day (40m)Abs Rel: 0.131 -> 0.108Improved
add Gram matrices orthogonality loss L_g (ablation)PRF · Oxford RobotCar night (40m)Abs Rel: 0.251 -> 0.231Improved
add Gram matrices orthogonality loss L_g (ablation)PRF · Oxford RobotCar day (40m)Abs Rel: 0.108 -> 0.109No clear effect
add similarity loss with day depth as pseudo label (ablation)PRFG · Oxford RobotCar night (40m)Abs Rel: 0.231 -> 0.233Mixed
add similarity loss with day depth as pseudo label (ablation)PRFG · Oxford RobotCar day (40m)Abs Rel: 0.109 -> 0.109No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy