Imagimob Edge Makes TensorFlow AI Models Edge Device-Ready at the Click of a Button

Imagimob, a company based in Stockholm, Sweden, developing artificial intelligence products for edge devices, announced a new SaaS solution that accelerates time to market for edge devices by converting h5 formatted AI models into ultra-efficient C-code in a matter of seconds — a process that once took weeks to complete. The new Imagimob Edge solution requires no programming skills and targets processing and interpreting sensor data in real time using trained models, a critical application in wearables and all sorts of connected devices.

Swedish companies Acconeer, Imagimob and Flexworks decided to join forces in a project to develop gesture-controlled in-ear headphones. Using the profound expertise of the three companies, the project will be a unique combination of research in advanced radar technology, expertise in AI (deep learning) and expertise in designing and manufacturing consumer electronics products. A first working prototype is expected within the first year of the project.

By combining Edge AI software from Imagimob, a company specializing in artificial intelligence development, with Pulsed Coherent Radar technology from Acconeer, the radar sensor specialists, the two Swedish companies have created an innovative platform for gesture control that is perfect for consumer electronics products. Scheduled to be demonstrated at CES 2020 (January 7-10) in Las Vegas, the solution will be implemented for application with wireless headphones, trained to recognize five hand gestures with high accuracy and close to zero latency.