AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Φeat: Physically-Grounded Material Feature Representation

Giuseppe Vecchio, Adrien Kaiser, Claudia Cuttano, ROUFFET Romain, Rosalie MARTIN, Elena Garces, Tamy Boubekeur

ECCV 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 1 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 single-render material supervision to DINOv3 (ablation)DINOv3 · material k-NN classificationaccuracy: 66.9 -> 34.5Hurt
add single-render material supervision to DINOv3 (ablation)DINOv3 · DuMaS material selectionmIoU: 59.7 -> 69.4Improved
add multi-render scheme (ablation)DINOv3 + single render · material k-NN classificationaccuracy: 34.5 -> 51.3Improved
add multi-render scheme (ablation)DINOv3 + single render · DuMaS material selectionF1: 78.1 -> 79.2Improved
add contrastive objective (ablation)DINOv3 + multi render · material k-NN classificationaccuracy: 51.3 -> 71.8Improved
add contrastive objective (ablation)DINOv3 + multi render · DuMaS material selectionmIoU: 69.9 -> 72.0Improved
material pretraining Φeat vs DINOv3DINOv3 · ImageNet linear probeaccuracy: 83.5 -> 80.1Hurt
material pretraining Φeat vs DINOv3DINOv3 · ADE20K linear probemIoU: 51.5 -> 46.2Hurt
material pretraining Φeat vs DINOv3DINOv3 · Illumination robustnesspairwise Hamming distance: 0.240 -> 0.221Improved
material pretraining Φeat vs DINOv3DINOv3 · Geometry robustnesspairwise Hamming distance: 0.365 -> 0.305Improved

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