UniHOI: Unified Human-Object Interaction Understanding via Unified Token Space
This work addresses the challenge of comprehensive interaction understanding in computer vision, offering a unified approach that enhances generalization for applications like robotics and scene analysis, though it is incremental in combining existing tasks.
The paper tackled the problem of separate modeling for human-object interaction detection and generation by proposing UniHOI, which uses a unified token space to jointly model both tasks, achieving state-of-the-art performance with a 4.9% accuracy improvement in detection and a 42.0% boost in generation metrics.
In the field of human-object interaction (HOI), detection and generation are two dual tasks that have traditionally been addressed separately, hindering the development of comprehensive interaction understanding. To address this, we propose UniHOI, which jointly models HOI detection and generation via a unified token space, thereby effectively promoting knowledge sharing and enhancing generalization. Specifically, we introduce a symmetric interaction-aware attention module and a unified semi-supervised learning paradigm, enabling effective bidirectional mapping between images and interaction semantics even under limited annotations. Extensive experiments demonstrate that UniHOI achieves state-of-the-art performance in both HOI detection and generation. Specifically, UniHOI improves accuracy by 4.9% on long-tailed HOI detection and boosts interaction metrics by 42.0% on open-vocabulary generation tasks.