{"slug": "infant-care-video-dataset-for-classification-of-interventions-using-transformers", "title": "Infant Care Video Dataset for Classification of Interventions Using Transformers", "summary": "Researchers introduced the Infant Care Video Dataset (ICVD), a collection of 4,144 videos spanning 12 simulated intervention classes, to enable automated documentation in neonatal intensive care units (NICUs). Using video transformer architectures TimeSformer and MotionFormer, they achieved top-1 accuracies of 93.97% and 93.17%, respectively, compared to 23.17% for a framewise approach, demonstrating a 70.80% performance gap that validates the need for temporal modeling. The dataset aims to reduce the clinical burden of documentation, where nurses spend about 25% of their time on record-keeping and up to 60% of interventions go undocumented.", "body_md": "arXiv:2608.23838v1 Announce Type: new\nAbstract: Healthcare documentation in the neonatal intensive care unit (NICU) presents significant challenges, with nurses spending approximately 25\\% of their time on record-keeping, while up to 60\\% of interventions remain undocumented. Motivated by the need to detect interventions from video automatically, we present the Infant Care Video Dataset (ICVD), a collection of 4,144 videos spanning 12 simulated intervention classes designed for developing automated documentation systems. Our manikin-based approach systematically varies conditions, such as camera angle and clinician skin tone, while ensuring privacy compliance. Using video transformer architectures (TimeSformer and MotionFormer), we establish strong baseline performance (93.97\\% and 93.17\\% top-1 accuracy) among the 12 infant care classes. Our ablation study comparing temporal models with a framewise approach (23.17\\% accuracy) demonstrates a 70.80\\% performance gap, validating the need for temporal modeling. The ICVD provides a foundation for developing automated documentation systems to reduce clinical burden in neonatal care environments and improve existing practices.", "url": "https://wpnews.pro/news/infant-care-video-dataset-for-classification-of-interventions-using-transformers", "canonical_source": "https://arxiv.org/abs/2608.23838", "published_at": "2026-08-26 04:00:00+00:00", "updated_at": "2026-08-26 04:14:06.807855+00:00", "lang": "en", "topics": ["computer-vision", "machine-learning"], "entities": ["Infant Care Video Dataset", "TimeSformer", "MotionFormer"], "alternates": {"html": "https://wpnews.pro/news/infant-care-video-dataset-for-classification-of-interventions-using-transformers", "markdown": "https://wpnews.pro/news/infant-care-video-dataset-for-classification-of-interventions-using-transformers.md", "text": "https://wpnews.pro/news/infant-care-video-dataset-for-classification-of-interventions-using-transformers.txt", "jsonld": "https://wpnews.pro/news/infant-care-video-dataset-for-classification-of-interventions-using-transformers.jsonld"}}