Maxim has teamed up with Xailient to create the fastest, lowest power IoT face detection solution

     Maxim Integrated Products has announced a partnership with Xailient Inc, which specializes in artificial intelligence (AI) edge computing, Xailient uses Maxim Integrated's MAX78000 ultra-low power neural network microcontroller in its patented neural network solution Detectum to detect and lock faces in videos and images. Compared with traditional embedded solutions, Xailient's neural network technology solution reduces power consumption by 250 times (only 280 microjoules), and only 12 ms per detection operation, supporting real-time network operation at a rate higher than the most efficient edge computing face detection solutions on the market today.

     Home improvement cameras, industrial-grade intelligent security cameras, and face recognition terminals in the retail industry are mostly battery-powered, and similar AI systems put strict requirements on power consumption to ensure that the battery supports the longest working time after charging. In addition to supporting stand-alone operation, Maxim Integrated microcontrollers work with Xailient's neural network technology to significantly improve overall system power efficiency and extend battery life in hybrid edge/cloud computing systems. Such applications employ a low-power "listen" mode that awakens more complex system operations when a face target is detected.

     Xailient's Detectum neural network solution, which includes focusing, zooming and visual awakening technologies, is 76 times faster than traditional software solutions when detecting and locking faces in video and images, and has the same or even higher detection accuracy. In addition, its flexible networking capabilities can be extended to applications beyond facial recognition, such as livestock storage and monitoring, parking space occupancy, and inventory level monitoring.