AI Paper Atlas

Research map / 3D vision / Depth and stereo

Boosting Monocular Depth Estimation With Lightweight 3D Point Fusion

Lam Huynh, Phong Nguyen, Jiří Matas, Esa Rahtu, Janne Heikkilä

ICCV 2021 · 3D vision › Depth and stereo · approach: Self-supervised monocular depth · 28 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 confidence predictor (ablation)Point-Fusion · NYU-Depth-v2RMSE: 0.097 -> 0.090Improved
add confidence predictor (ablation)Point-Fusion · NYU-Depth-v2REL: 0.015 -> 0.014Improved
increase Fusion-Nets from 2 to 5 (ablation)Point-Fusion · NYU-Depth-v2RMSE: 0.105 -> 0.090Improved
increase Fusion-Nets from 5 to 6 (ablation)Point-Fusion · NYU-Depth-v2RMSE: 0.090 -> 0.093Hurt
add 3D point convolution branch (ablation)Point-Fusion · NYU-Depth-v2RMSE: 0.196 -> 0.090Improved
use FKAConv instead of PCC (ablation)Point-Fusion · NYU-Depth-v2RMSE: 0.096 -> 0.090Improved
use FKAConv instead of PCC (ablation)Point-Fusion · NYU-Depth-v2#params: 9.1M -> 8.7MImproved
sample input points from dense GT depth instead of COLMAP (ablation)Point-Fusion · NYU-Depth-v2RMSEImproved
add 32 sparse 3D points to the networkPoint-Fusion · NYU-Depth-v2RMSE: 0.505 -> 0.319Improved
use Point-Fusion instead of NLSPN with 2 pointsPoint-Fusion vs NLSPN · NYU-Depth-v2RMSE: 1.152 -> 0.470Improved

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