AI Paper Atlas

Research map / 3D vision / Depth and stereo

When Epipolar Constraint Meets Non-Local Operators in Multi-View Stereo

Tianqi Liu, Xinyi Ye, Weiyue Zhao, Zhiyu Pan, Min Shi, Zhiguo Cao

ICCV 2023 · 3D vision › Depth and stereo · approach: Cost-volume multi-view stereo · 62 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 intra- and cross-epipolar augmentation (IEA and CEA) (ablation)MVSTER · DTU evaluation setOverall (mm): 0.318 -> 0.297Improved
add local augmentation (LA) convolution after IEA and CEA (ablation)ET · DTU evaluation setOverall (mm): 0.297 -> 0.291Improved
add CEA to IEA and LA (ablation)ET · DTU evaluation setOverall (mm): 0.294 -> 0.291Improved
add IEA to CEA and LA (ablation)ET · DTU evaluation setDepth error: 5.982 -> 5.754Improved
use line-to-line search instead of point-to-line search (ablation)ET · DTU evaluation setinference time (s): 3.06 -> 0.46Improved
use line-to-line search instead of point-to-line search (ablation)ET · DTU evaluation setOverall (mm): 0.290 -> 0.291No clear effect
use Epipolar Transformer instead of FMT (ablation)MVSTER feature encoder · DTU evaluation setOverall (mm): 0.303 -> 0.291Improved
use Epipolar Transformer instead of DCN or ASPP (ablation)MVSTER feature encoder · DTU evaluation setOverall (mm): 0.309 -> 0.291 (DCN); 0.320 -> 0.291 (ASPP)Improved
embed ET into CasMVSNet feature encoderCasMVSNet · DTU evaluation setOverall (mm): 0.355 -> 0.301Improved
embed ET into CasMVSNet feature encoderCasMVSNet · Tanks and Temples AdvancedMean F-score: 31.12 -> 35.65Improved
use ET-MVSNet versus Gipuma on accuracyET-MVSNet · DTU evaluation setAccuracy (mm): 0.283 (Gipuma) vs 0.329No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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