AI Paper Atlas

Research map / 3D vision / Depth and stereo

Learned Multi-Patch Similarity

Wilfried Hartmann, Silvano Galliani, Michal Havlena, Luc Van Gool, Konrad Schindler

ICCV 2017 · 3D vision › Depth and stereo · approach: Cost-volume multi-view stereo · 119 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 concatenation instead of averaging of Siamese branch outputs (ablation)OUR (5-view) · DTU CANmedian accuracy error (mm): 0.477 -> 0.478No clear effect
use averaging instead of concatenation of Siamese branch outputs (ablation)OUR (5-view) · DTU FLOWERmedian accuracy error (mm): 0.843 -> 0.771Improved
compute proposed similarity over 5 views directly instead of consensus of 2-view scores (ablation)OUR · DTU (20 objects)mean completeness error (mm): 2.964 -> 2.126Improved
compute proposed similarity over 5 views directly instead of consensus of 2-view scores (ablation)OUR · DTU (20 objects)median completeness error (mm): 1.391 -> 0.868Improved
compute proposed similarity over 5 views directly instead of consensus of 2-view scores (ablation)OUR · DTU (20 objects)mean accuracy error (mm): 1.339 -> 1.356No clear effect
replace ZNCC with learned multi-view similarityplane-sweep stereo · DTU (20 objects)median accuracy error (mm): 0.577 -> 0.472Improved
replace ZNCC with learned multi-view similarityplane-sweep stereo · DTU (20 objects)mean accuracy error (mm): 1.237 -> 1.356Hurt
replace LIFT descriptor distance with learned multi-view similarityplane-sweep stereo · DTU FLOWERmedian accuracy error (mm): 1.420 -> 0.771Improved
replace LIFT descriptor distance with learned multi-view similarityplane-sweep stereo · DTU CANmedian accuracy error (mm): 1.207 -> 0.478Improved
increase number of input views from five to nine at test time without retraining (ablation)OUR · DTUdepth map qualityImproved
decrease number of input views from five to three at test time without retraining (ablation)OUR · DTUdepth map qualityHurt

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