Computer Vision incorporates vision and machine-learning technology to allow for a machine to gain understanding of its surroundings through visual recognition. Using optical sensory as an input, a machine equipped with machine-learning capabilities allows it to extract and classify its surrounding, so that it can better “react” to the environment that it is in. This is part of the Artificial Intelligence (AI) initiative, where machines are trained to interpret sensory inputs, so that it can behave and perform human-like tasks.

With an increasing number of AI applications operated on the EDGE, the ability to create low power deep learning models becomes crucial in a battery operated device. Similarly, visual sensors is no exception to this power constraints. In order to fuel the next wave of always-on sensing needs on this so-called edge devices, optical sensors needs to also support the notion of always-on visual sensing.

PixArt’s Ultra-Low Power (ULP) Computer Vision Sensor products are developed with unique power-saving architecture. With power consumption as low as 600uW at 30fps, these ULP sensors are designed to provide visual means to AI machine-learning devices. 
Ultra Low Power Computer Vision Sensor
Part No. Type
Package Dimension
Lens FOV
(Array Size)
Shutter Type
Max. Frame Rate
Avg. Operating Power Consumption
More Info
Camera Module
90 degrees
QVGA (320x240)
Global Shutter
1,400µW@30fps QGA; 500µW@30fps QQVGA