AI Paper Atlas

Research map / 3D vision / Depth and stereo

"Smoothness, Synthesis, and Sampling: Re-thinking Unsupervised Multi-View Stereo with DIV Loss"

Alex Rich, Noah Stier, Pradeep Sen, Tobias Hollerer

ECCV 2024 · 3D vision › Depth and stereo · approach: Cost-volume multi-view stereo · 4 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 clamped 2nd-order depth smoothness (ablation)DIV-MVS baseline loss · DTUOverall (mm): 0.361 -> 0.346Improved
clamp the 1st-order smoothness gradient (ablation)DIV-MVS baseline loss · DTUOverall (mm): 0.361 -> 0.366Hurt
use 2nd-order smoothness without clamping (ablation)DIV-MVS baseline loss · DTUOverall (mm): 0.361 -> 0.365Hurt
add learned reference-view synthesis (no occlusion masks) (ablation)DIV-MVS with clamped 2nd-order · DTUOverall (mm): 0.346 -> 0.339Improved
add occlusion masking without learned synthesis (ablation)DIV-MVS with clamped 2nd-order · DTUOverall (mm): 0.346 -> 0.351Hurt
add occlusion masking to learned synthesis (ablation)DIV-MVS · DTUOverall (mm): 0.346 -> 0.335Improved
add supervision-view sampling beyond MVS input views (ablation)DIV-MVS · DTUOverall (mm): 0.335 -> 0.330Improved
replace pipeline loss with DIV lossDIV-RC (RC-MVSNet) · DTUOverall (mm): 0.350 -> 0.333Improved
replace pipeline loss with DIV lossDIV-CL (CL-MVSNet) · DTUOverall (mm): 0.330 -> 0.321Improved
replace pipeline loss with DIV lossDIV-MVS · DTUAbs. depth error (mm): 19.34 -> 16.32Improved
replace pipeline loss with DIV lossDIV-MVS · DTUTraining memory (GB): 10.50 -> 10.52No clear effect
train DIV-MVS on DTU only and test zero-shot on Tanks and Temples intermediateDIV-MVS · Tanks and Temples intermediateF-score: 37.03 -> 38.35 (vs CL-MVSNet)Improved

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