AI Paper Atlas

Research map / 3D vision / 3D generation

GenMatter: Perceiving Physical Objects with Generative Matter Models

Eric Li, Arijit Dasgupta, Yoni Friedman, Mathieu Huot, Vikash Mansinghka, Thomas O'Connell, William T. Freeman, Joshua B. Tenenbaum

CVPR 2026 · 3D vision › 3D generation · approach: Learning physics from video · 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 cluster layer (hierarchy) over particles, versus fixed particles K=5 (ablation)GenMatter · Random dot kinematogramscorrelation r^2: 0.35 -> 0.86Improved
add cluster layer over adaptive particles K=500 (ablation)GenMatter · Random dot kinematogramscorrelation r^2: 0.41 -> 0.86Improved
add cluster-level variables (ablation)GenMatter · TAP-Vid DAVIS (SAM init)matter-weighted Jaccard: 0.69 -> 0.79Improved
add monocular depth input (ablation)GenMatter · TAP-Vid DAVIS (SAM init)matter-weighted Jaccard: 0.69 -> 0.79Improved
add monocular depth input (ablation)GenMatter · Gestalt camouflaged videosper-pixel accuracy: 0.89 -> 0.94Improved
use ground-truth mask instead of SAM2 for initializationGenMatter · TAP-Vid DAVISmatter-weighted Jaccard: 0.79 -> 0.77Hurt
use ground-truth mask instead of SAM for initializationCoTracker3 · TAP-Vid DAVISmatter-weighted Jaccard: 0.78 -> 0.78No clear effect
replace FlowSAM with GenMatterFlowSAM · Gestalt camouflaged videosJaccard: 0.67 -> 0.72Improved
replace SegAnyMo with GenMatterSegAnyMo · Gestalt camouflaged videosJaccard: 0.26 -> 0.72Improved
subsample data points to 1/128 of full resolutionGenMatter · TAP-Vid DAVISmatter-weighted Jaccard: 0.76 -> 0.79No clear effect
subsample data points to 1/512 of full resolutionGenMatter · TAP-Vid DAVISmatter-weighted Jaccard: 0.76 -> 0.56Hurt
subsample data points from full resolution to 1/128GenMatter · TAP-Vid DAVISFPS: 0.80 -> 9.8Improved

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