MAMHOI: Factorizing Scene-Aware Human-Object Interaction through Affordances
MAMHOI factorizes scene-aware human-object interaction generation through motion affordances without paired scene data.
MAMHOI is an affordance-mediated method for generating human-object interactions inside complex 3D scenes. A scene-conditioned model predicts where and how an interaction can be executed, and an affordance-conditioned model then synthesizes the human-object motion. The split lets scene understanding and interaction dynamics be learned from separate datasets. Indoor experiments report reduced object-scene penetration while preserving interaction quality.
- Splits scene feasibility prediction from human-object motion synthesis.
- Trains without scarce paired human-object-scene supervision.
- Indoor tests report less object-scene penetration with preserved interaction quality.
From the original article
1 min read at huggingface.co
Generating realistic human-object interactions (HOI) in complex 3D scenes requires two complementary capabilities: reasoning about interaction feasibility in the environment and synthesizing realistic human-object motion. However, supervision for these capabilities is rarely available jointly at scale.
By Mingyuan Lei, Yoonchang Sung, Tat-Jen Cham · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.