AI Paper Atlas

Research map / 3D vision / Depth and stereo

CasaMaestro: Multi-View Panoramas for House-Scale 3D Reconstruction

Yuzhou Ji, Xiaotian Yang, Zhipeng Zhang

ECCV 2026 · 3D vision › Depth and stereo · approach: Cost-volume multi-view stereo · 0 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 panoramic cross-view pose decoder (vs linear head) (ablation)CasaMaestro DINO-base · Realsee poseAUC@30: 0.568 -> 0.831Improved
add panoramic cross-view pose decoder (vs linear head) (ablation)CasaMaestro DINO-large · Realsee poseAUC@30: 0.598 -> 0.871Improved
add ERP rotation augmentation (ablation)CasaMaestro DINO-base, linear decoder · Realsee poseAUC@30: 0.568 -> 0.772Improved
add ERP augmentation on top of pose decoder (ablation)CasaMaestro DINO-large · Realsee poseAUC@30: 0.871 -> 0.927Improved
add ERP augmentation (linear decoder) (ablation)CasaMaestro DINO-base · Realsee depthAbsRel: 0.092 -> 0.111Hurt
scale backbone from DINO-base to DINO-large (linear decoder, no aug) (ablation)CasaMaestro · Realsee poseAUC@30: 0.568 -> 0.598No clear effect
use panoramic pose decoder rather than ERP augmentation alone (ablation)CasaMaestro · Realsee poseAUC@30: decoder 0.831/0.871 vs aug 0.772/0.815Improved
panoramic multi-view model vs pinhole VFMs on sparse house panoramasCasaMaestro · Realsee-RealAUC@30: 0.510 -> 0.903 vs Pi3-FinetuneImproved
panoramic multi-view model vs best previous on syntheticCasaMaestro · Realsee-SynAUC@30: 0.507 -> 0.927 vs Pi3-FinetuneImproved
fine-tune VGGT/Pi3 on Realsee training setVGGT, Pi3 · Realsee poseAUC@30: VGGT 0.388 -> 0.411; Pi3 0.450 -> 0.510Improved
multi-view panoramic depth vs single-view panoramic depth methodsCasaMaestro · zero-shot Replica/PanoSUNCG/HM3DAbsRel: e.g. Replica 0.058 vs 0.070 best priorImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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