While there are already a variety of systems that monitor how your baby sleeps, most of them require you to fasten some sort of wearable to your infant's body. Nanit, however, is different. It's a contactless system that utilizes an overhead camera to observe your little one, using machine learning to teach itself their sleep patterns.

The hardware end of the system consists simply of an HD night vision camera on a floor stand, which looks down into the baby's crib. It also has a built-in night light, can play soothing nature sounds or white noise, and measures room data such as temperature and humidity.

Using computer vision, it processes what it sees, analyzing how much the infant moves throughout the night, and in what ways. By comparing any one night's patterns to what it's learnt is normal for that particular baby, Nanit is reportedly able to determine how well they slept.

Parents receive a daily report of the previous night's sleep every morning, on an accompanying iOS/Android app. This includes metrics such as sleep patterns, parent interventions, sleep onset, and total hours of sleep – the same information is available for daytime sleep, too. Time-lapse video is provided, showing moments when the baby may have been fussing or was active.

The app additionally lets parents view real-time footage from the camera via a secure Wi-Fi connection, and can send alerts when the baby is crying or moving.

Announced this Wednesday, Nanit can now be preordered for US$279, and will ultimately retail for $349. Users will also need to subscribe to an analytics service, for $10/month or $50 for a year.

Source: Nanit

View gallery - 2 images