AI Paper Atlas

Research map / 3D vision / Depth and stereo

Adaptive Confidence Thresholding for Monocular Depth Estimation

Hyesong Choi, Hunsang Lee, Sunkyung Kim, Sunok Kim, Seungryong Kim, Kwanghoon Sohn, Dongbo Min

ICCV 2021 · 3D vision › Depth and stereo · approach: Self-supervised monocular depth · 40 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
add learned confidence thresholding with L_T supervision (ablation)DepthNet · KITTI Eigen splitAbs Rel: 0.108 -> 0.099Improved
add learned confidence thresholding with L_T supervision (ablation)DepthNet · KITTI Eigen splitRMSE: 4.552 -> 4.266Improved
add explicit supervision L_T for the threshold (versus Tonioni et al. learned threshold) (ablation)DepthNet · KITTI Eigen splitAbs Rel: 0.101 -> 0.100Improved
use the full soft-threshold learning versus Tonioni + L_T (ablation)DepthNet · KITTI Eigen splitRMSE: 4.390 -> 4.266Improved
add RefineNet with PAC-based probabilistic refinement (ablation)Ours · KITTI Eigen splitAbs Rel: 0.099 -> 0.096Improved
add the uncertainty map sigma to depth refinement (ablation)RefineNet · KITTI Eigen splitAbs Rel: 0.099 -> 0.096Improved
add the uncertainty map sigma to depth refinement (ablation)RefineNet · KITTI Eigen splitRMSE: 4.298 -> 4.201Improved
apply threshold learning to PackNet-SfM architecture (ablation)PackNet-SfM · KITTI Eigen splitAbs Rel: 0.111 -> 0.105Improved
add refinement to PackNet (D+R) (ablation)PackNet-SfM · KITTI Eigen splitAbs Rel: 0.105 -> 0.100Improved
use self-supervised ThresNet instead of supervised ThresNet (ablation)Ours (D+R) · KITTI Eigen splitAbs Rel: 0.096 -> 0.098No clear effect
use alternative pseudo labels from synthetic-trained meta-learning model (ablation)Ours (D) · KITTI Eigen splitAbs Rel: 0.099 -> 0.102Mixed
apply threshold learning (w/tau) to LAFNet confidence estimation (ablation)LAFNet · KITTI 2015 C-SGMAUC: 1.680 -> 1.587Improved

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