AI Paper Atlas

Research map / 3D vision / Depth and stereo

Gaussian Belief Propagation Network for Depth Completion

Jie Tang, Pingping Xie, Jian Li, Ping Tan

ECCV 2026 · 3D vision › Depth and stereo · approach: Dense depth prediction and completion · 1 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 probability-based loss to the optimization baseline (ablation)V1 baseline · NYUv2RMSE (mm): 342.70 -> 340.84Improved
add MRF with fixed local edges solved by serial GBP (ablation)V2 · NYUv2RMSE (mm): 340.84 -> 108.29Improved
use attention-only backbone instead of conv (ablation)V3 · NYUv2RMSE (mm): 108.29 -> 109.54Hurt
combine conv and attention in global-local unit (ablation)V3 · NYUv2RMSE (mm): 108.29 -> 107.01Improved
add dynamic MRF parameters (ablation)V5 · NYUv2RMSE (mm): 107.01 -> 103.92Improved
add dynamic non-local edges with parallel propagation (ablation)V6 · NYUv2RMSE (mm): 103.92 -> 101.42Improved
increase local edges from 4 to 8 (ablation)V7 · NYUv2RMSE (mm): 101.42 -> 100.95Improved
increase GBP iterations from 3 to 5 (ablation)V8 · NYUv2RMSE (mm): 100.95 -> 100.69Improved
add dynamic non-local edges (delta 1.02) (ablation)V6 · NYUv2delta 1.02 (%): 84.69 -> 85.19Improved
add serial GBP over MRF (delta 1.02) (ablation)V2 · NYUv2delta 1.02 (%): 25.23 -> 83.71Improved
filter lowest 1% confidence pixels using predicted precisionGBPN · NYUv2RMSE (mm): 85.51 mm -> 26.89 mmImproved
GBPN versus BP-Net in zero-shot VOID 150 pointsGBPN · VOID 150 points (zero-shot from NYUv2)RMSE (m): 0.847 -> 0.830Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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