Paper Overview
- Field: cs.CV
- Authors: Homanga Bharadhwaj, Yash Jangir
- Published: 2026-07-21
- arXiv: 2507.15491
- Interaction anticipation is cast as object-centered 3D motion forecasting with no assumptions on object properties.
- Pseudo-ground-truth 3D tracks come from a dense tracker built on a pretrained video model; the forecaster sees only observed frames.
- Future RGB and geometry are replaced by learned mask latents; a lightweight adapter handles prediction while large components stay frozen.
- With 40k videos and no language or auxiliary inputs, it generalizes to out-of-distribution objects, environments, viewpoints, and interactions, beating larger models trained on 1M+ videos.
Abstract
Humans can infer how objects are likely to move from passive observation: a cup may be lifted, a drawer may slide, and a lid may rotate shut. Such predictions expose the physical consequences of interaction needed to act in the real world. We study how to learn this anticipation from ordinary monocular videos of human-object interaction. Given a short observed video context, MotionForesight predicts future 3D trajectories for points on the manipulated object. This casts interaction prediction as object-centered 3D motion forecasting without any assumptions on the object properties.Our key insight is that video prediction models already encode rich priors about how objects move during human interactions. We redirect these priors from pixel prediction toward future 3D scene flow. We start from a dense 3D tracker built on a pretrained video model, generate pseudo-ground-truth tracks from complete clips, and train the forecaster using only the observed frames. We replace future RGB and geometry with learned mask latents and train a lightweight adapter to turn the retrospective tracking representation into a forward predictor, while freezing the large video and tracking components.
Using just 40k human videos and no auxiliary inputs such as language, MotionForesight generalizes across diverse out-of-distribution objects, environments, viewpoints, and interactions. It also outperforms substantially larger models that use over a million training videos. These results show that we can efficiently re-purpose video priors into explicit geometric forecasts for embodied intelligence.