There are as many aspects to the solution sets created to address Cloud- and IoT-oriented applications as the markets they serve. Devices empowered by face and/or pattern recognition require many layers of tech infrastructure. This includes the RF systems behind the Cloud to the sensor suite used to create the user interface. One of the ways to reduce cost and footprint while adding functionality is to replace, consolidate, or augment existing components in the device.
CEA-Leti recently released an autonomous imaging device called µWAI (micro-WAY) to address this opportunity in personal electronics. Provided in a package as small as a quarter, the autonomous imager has a readout and processing architecture with optimized algorithmic pipeline. Recognition results are created from a sequence of elementary algorithms, to provide highly efficient sleep/wake modes. Presented as the first smart image sensor with features s auto-exposure, motion detection, and feature extraction for event-based functioning, along with AI-based object recognition.
Providing highly reliable identification, the functionality can also make a highly reliable decision using a few tens of pJ/pixel/frame, which measures the energy spent by pixel in every single image. Using 3-6µW in operation, the µWAI approach is up to 10,000x more efficient than legacy multi-component solutions. A cost-effective low-power smart-wakeup system intended for mainstream applications. Features include privacy-compliant, AI-based recognition with nearly 95% human-face detection success, and a wide light sensitivity to ensure accurate capture in variable conditions,
Addressing automatic switching and face identification in mobile devices, contact-less smart switching of household appliances, and other user devices in a smart home, µWAI can also count people, trigger alarms, monitor vehicle interiors and identify driver, as well as other observational tasks. The technology behind µWAI will be introduced during CEA-Leti’s flagship digital event, Leti Innovation Days, June 22 and 23, 2021.
“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.”
