Project No.: ELKH-POC-2022-009, Title: Visual Monitoring of Premature Babies

Newborn and preterm infant care is of utmost importance in hospital settings, especially in neonatal units where continuous attention and care are essential. In addition to monitoring vital life functions, it is also crucial to observe sleep-wake cycles, behavioral states, and both circadian and ultradian rhythms, as these provide valuable information about the infants’ development and overall condition. Currently, accurately tracking infant activity and behavioral states requires significant staff effort, as it mainly relies on direct, in-person observation.

Our solution offers a system capable of recording video, audio, and other data around the clock from multiple incubators in a clinical environment. We have developed a camera specifically designed for incubators, which are in active use in three clinics, as well as a central computer application that has enabled us to collect and process thousands of hours of recordings. The cameras are equipped with infrared illumination and Wi-Fi streaming, featuring custom design and electronics. The central computer is used to control multiple cameras simultaneously and securely record the data.

Based on these recordings, we have developed artificial intelligence algorithms that can accurately identify sleep states and care events. The system and its associated database provide a reliable foundation for long-term clinical research and collaborative development.