AI Paper Atlas

Research map / Point clouds / Shape and surface completion

RoofDiffusion: Constructing Roofs from Severely Corrupted Point Data via Diffusion

Kyle Shih-Huang Lo, Jorg Peters, Eric Spellman

ECCV 2024 · Point clouds › Shape and surface completion · approach: Point cloud diffusion generation · 3 citations (OpenAlex)

This paper tried a common change: More diffusion sampling steps

increase inference steps from 60 to 500 · PoznanRD s99 i80 · RMSE (m): 1.123 -> 0.893 Improved

What 457 accepted papers report for this change

See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.

Read the paper

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 RoofDiffusion (footprint-guided) vs best interpolation baseline on PoznanRD with high sparsity and incompletenessheight map completion · PoznanRD s99 i80MAE (m): 0.827 -> 0.603Improved
use RoofDiffusion on sparsity-only BuildingNet s99 i0height map completion · BuildingNet s99 i0RMSE (m): 0.540 (IDW) vs 0.550No clear effect
use No-FP RoofDiffusion vs CU-Net on PoznanRD s99 i60 and BuildingNet s99 i0no-footprint completion · RMSE on PoznanRD s99 i60 and BuildingNet s99 i0RMSE (m): 2.600 vs 2.554; 0.962 vs 0.829Mixed
preprocess sparse point clouds with RoofDiffusion before City3DCity3D · PoznanRD s99 i30RMSE (m): 0.352 -> 0.244Improved
add tree augmentation in training (ablation)RoofDiffusion · PoznanRD, 3 treesRMSE (m): 0.523 -> 0.356Improved
add tree augmentation in training when there are no trees (ablation)RoofDiffusion · PoznanRD, 0 treesRMSE (m): 0.203 -> 0.208No clear effect
increase inference steps from 60 to 500 (ablation)RoofDiffusion · PoznanRD s99 i80RMSE (m): 1.123 -> 0.893Improved
increase inference steps beyond 500 to 2000 (ablation)RoofDiffusion · PoznanRD s99 i80RMSE (m): 0.893 -> 0.922No clear effect
apply RoofDiffusion to KITTI depth completionRoofDiffusion · KITTI depth completion valRMSE: 917.76 (CU-Net) vs 1631.98Hurt
apply mean shift offset to KITTI predictions (ablation)RoofDiffusion on KITTI · KITTI depth completion valMAE: 491.61 -> 448.97Improved

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