Called µWAI (microway) and sized as small as a 1-€ coin, the autonomous imager features a novel readout and processing architecture co-designed with an optimized algorithmic pipeline, in which the recognition results from a sequence of elementary algorithms, to provide ultralow-power wake-up modes and compact silicon implementation to keep costs down. 

It is the first smart image sensor jointly featuring auto-exposure for all lighting conditions and 88 dB dynamic range, as well as motion detection and feature extraction for event-based functioning, and AI-based object recognition that triggers highly reliable identification. These key features also enable highly reliable decision taking for a few tens of pJ/pixel/frame, which outperforms existing off-the-shelf systems. The pJ/pixel/frame measures the energy spent by each pixel for each single image within a frame of images. A typical implementation requires about 10,000 times more energy than µWAI.

The autonomous imager is the world’s first highly efficient, compact and ultralow-power, smart-awaken system designed for everyday small appliances. It also includes: energy efficiency: consuming 10k less than low-power camera plus processor set; privacy-compliant, AI-based recognition: nearly human-detection performance (95 percent); wide operating lighting sensitivity to ensure accurate recognition in extremely variable conditions; five-year, always-on CR1025 battery lifetime, and 3 – 6 µW operation, which is required for the Internet of Things and can work with a button cell that lasts five years.

Applications and functions include automatic switching and face identification in mobile devices, contact-less smart switching of household appliances and sport-and-entertainment devices in smart homes. µWAI also provides face recognition, people counting, alarm triggering in smart buildings, vehicle-interior situation awareness, driver identification, parking-situation awareness and a smart-unlocking system in automobiles.

“The recognition engine is optimized to recognize faces when movement is detected. CEA-Leti’s team is working hand-in-hand with STMicroelectronics to develop specific smart-imager products as we consider extending the technology to other use cases,” said Antoine Dupret, CEA-Leti's industrial partnership manager. “We target adapting the recognition engine as IP embedded in various cameras and optimizing the performance of the imager to the requirements of our partner's customers.” (Source: CEA-Leti)