Developers from the CEA-LETI Institute from French Grenoble presented the world’s first autonomous image scanner for smartphones, smart home appliances and cars. A small chip with the sides of about one centimeter can work five years from the battery type “Tablet”, activating only when a person’s face falls in his field of view. The algorithm built into the chip with the highest accuracy recognizes personality to fulfill its requirements.
The autonomous image-scanner μWAI (Micro-Way) is equipped with a new reading and processing architecture developed in conjunction with an optimized algorithmic conveyor in which recognition is the result of the sequence of elementary algorithms. The algorithm ensures the modes of awakening with ultra-low energy consumption, reacting, first of all, to move in the field of view of the sensor.
According to the developers, this is the first intelligent image sensor that provides automatic exposure for all lighting conditions in the dynamic range of 88 dB together with motion detection. Also, the system retrieves the specified functions from the image and using an artificial intelligence launches highly handed-based identification. The accuracy of identification of the personality reaches 95%, while the consumption of the device is ten thousand times less than in the case of modern “typical” embodiments.
From one CR1025 type battery, the sensor can continuously work for five years. The sensor power consumption in the internet of things is 3-6 μW. Such sensors promise to be used in everyday household appliances, in sports and entertainment institutions, in the recognition systems in cars, in parking lots and many where else, where contactless control of electronics or admission to any special functions will be convenient and appropriate.