CareOS, the health-beauty platform for the connected bathroom of the French group Barracoda, goes one step further with its Artemis mirror. The bathroom smart screen now incorporates the Edge AI TensorFlow Lite, Google's framework for deploying artificial intelligence models to mobile devices and connected objects. The announcement took place on the occasion of Google I / O 2019, May 7th to 9th Mountain View.
Awarded an Award at CES 2019, the Artemis connected mirror has many features. Beyond the voice command to consult the weather and choose his favorite music or facial recognition, an onboard 360 ° camera can scan the physical evolutions (weight, wrinkles, moles …) of the user and to prevent it in case of possible problem. Thanks to augmented reality, the user can also test and visualize different makeup looks. And to go even further, the mirror embeds tutorials to guide the user on the right gestures beauty (makeup, hair …) or health (brushing teeth, skin care …).
A mirror boosted to AI
The mirror connects objects and services to offer customization features, thanks to artificial intelligence, all securely through a private network and "personal safe and confidential". With TensorFlow Lite, CareOS wishes to go even further: "TensorFlow Lite's state-of-the-art artificial intelligence gives us a wealth of power to provide useful and engaging information to improve preventative care, beauty routines and hygiene rituals while retaining all local data."", says Chloé Szulzinger, co-founder and Head of Marketing-Communication CareOS who already works with many brands such as Wella, Withings, iHealth, Romy, Legrand …
"The Edge AI is often considered valuable for its speed, but CareOS shows that the localized performance of TensorFlow Lite opens up possibilities that go far beyond reducing time"In turn," said Tim Davis, product manager at Google AI, in a statement.CareOS is able to change health outcomes by providing people with instant feedback on their practices and by suggesting preventative care exactly where they are when such questions arise. This is an eloquent example of the very practical way that TensorFlow Lite makes possible new models. "