{"slug": "single-query-person-centric-bimanual-hand-object-interaction-detection", "title": "Single-Query Person-Centric Bimanual Hand-Object Interaction Detection", "summary": "Researchers posting to arXiv (paper 2609.12155v1) proposed a person-centric formulation for bimanual hand-object interaction detection in which a single query predicts a structured output for one person, including the human box, body pose, hand boxes and states, and interaction targets. The method introduces part-aware deformable attention across human, hand, and pose-specific reference regions and a hand-to-query relationship matrix that lets each hand select its interaction target from the detected query set plus a learnable off token. Experiments with a transformer-based detector on a new COCO-based dataset with person-centric bi-manual interaction annotations showed the formulation improves person-level bi-manual interaction parsing and provides a unified framework for joint detection, pose estimation, and hand reasoning.", "body_md": "arXiv:2609.12155v1 Announce Type: new \nAbstract: Understanding person-level bi-manual interactions requires not only detecting hands, but also identifying which two hands belong to the same person and what each hand interacts with. Existing hand--object interaction methods are mostly hand-centric: they treat each hand as an independent instance, which can lead to ambiguous ownership in multi-person scenes.\n  We propose a person-centric formulation in which a single query predicts a structured output for one person, including the human box, body pose, hand boxes and states, and interaction targets. We introduce part-aware deformable attention to allocate attention across human, hand, and pose-specific reference regions, enabling one query to capture the full person structure. We further unify detection and interaction reasoning with a hand-to-query relationship matrix, where each hand selects its interaction target from the detected query set plus a learnable off token, directly recovering the target's box and class without separate object regression.\n  We build a COCO-based dataset with person-centric bi-manual interaction annotations and define structured metrics for evaluating hand states and complete hand--object tuples. Experiments with a transformer-based detector show that our formulation improves person-level bi-manual interaction parsing and provides an effective unified framework for joint detection, pose estimation, and hand reasoning.", "url": "https://wpnews.pro/news/single-query-person-centric-bimanual-hand-object-interaction-detection", "canonical_source": "https://arxiv.org/abs/2609.12155", "published_at": "2026-09-14 04:00:00+00:00", "updated_at": "2026-09-14 04:27:21.289428+00:00", "lang": "en", "topics": ["computer-vision", "machine-learning", "ai-research"], "entities": ["arXiv", "COCO"], "alternates": {"html": "https://wpnews.pro/news/single-query-person-centric-bimanual-hand-object-interaction-detection", "markdown": "https://wpnews.pro/news/single-query-person-centric-bimanual-hand-object-interaction-detection.md", "text": "https://wpnews.pro/news/single-query-person-centric-bimanual-hand-object-interaction-detection.txt", "jsonld": "https://wpnews.pro/news/single-query-person-centric-bimanual-hand-object-interaction-detection.jsonld"}}