On May 25, French research agency CEA-Leti announced the launch of the first automatic imager that activates smartphones and small appliances through facial recognition or other specific modes, as well as driver identification.
The automatic imager is called µWAI (micro-WAY) and is the size of a 1 euro coin. In addition, the imager features a read and processing architecture co-designed with optimized pipeline algorithms to support ultra-low-power wake-up modes and compact silicon usage to reduce cost. In pipelined algorithms, the recognition results come from a series of basic algorithms.
(Image credit: CEA-Leti)
In addition, the imager adopts a new reading and processing architecture, and synchronously optimizes the design of pipeline algorithms, using a series of basic algorithms for recognition, which in turn provides ultra-low power wake-up modes and compact silicon applications, reducing costs.
The µWAI is the first smart image sensor of its kind with automatic exposure in all lighting conditions and 88dB dynamic range, motion detection and feature extraction for event-based functions, and AI-based object recognition that triggers highly reliable recognition.
These key features also enable highly reliable decision making at tens of pJ/pixel/frame, outperforming existing off-the-shelf systems. pJ/pixel/frame measures the energy consumed per pixel for each individual image in a frame. A typical application requires around 10,000 times the energy of µWAI.
According to the agency, the automatic imager is the world’s first efficient, compact, ultra-low-power smart wake-up system. It also includes high energy efficiency: consumes 10k less than low-power cameras and processor sets; complies with privacy protection requirements, AI-based recognition: close to human detection performance (95%); Accurate identification under changing conditions; the CR1025 battery can be used for five years when left on, and meets the 3-6µW operating power required by the Internet of Things, and can also work with a 5-year life coin cell battery.
Applications and functions include automatic switching and facial recognition in mobile devices, contactless smart switching in home appliances, and sports entertainment devices in smart homes. ? WAI also provides facial recognition, people counting, alarm triggering in smart buildings, vehicle interior situation awareness, driver recognition, parking location awareness, and smart car unlocking systems.
The imager is optimized to recognize faces when motion is detected. Antoine Dupret, Industry Partnerships Manager at CEA-Leti, said: “We are considering extending the technology to other application use cases, so the CEA-Leti team is working with ST to develop a specific smart imager product. The instrument is tuned to IP embedded in various cameras to optimize performance according to the needs of our partner customers.”
The Links: CM150TX-24S LM150X08
0 Comments for “CEA-Leti unveils first automatic imager for cars”