Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study introduces the Infant Care Video Dataset (ICVD), which consists of 4,144 videos across 12 intervention classes aimed at automating documentation in neonatal intensive care units (NICUs). By employing video transformer architectures like TimeSformer and MotionFormer, the authors achieve impressive baseline accuracies of 93.97% and 93.17%, respectively, while demonstrating the critical importance of temporal modeling over framewise approaches, which only achieved 23.17% accuracy. The findings highlight the potential of the ICVD to significantly alleviate the documentation burden on NICU staff and enhance care practices through automated systems.
A staggering 70.80% performance gap reveals the crucial role of temporal modeling in automating neonatal care documentation.
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.