Low-Power Intelligent Object Detection

With today’s popularity of artificial intelligence (AI), AI technology is now finding its way into smaller IoT devices, even from the sensor perspective. Performing AI calculation on the sensor delivers great benefits including bandwidth saving, computation-cost saving, power saving, faster response rate, and data security reinforcement. PixArt’s Low-Power Intelligent Object Detection (LIOD) technology is specially designed for applications that require detection capabilities with the least power consumption. 

PixArt's LIOD product consists of an ultra-low-power CMOS image sensor and an ultra-low-power image processing chip. The sensor is able to detect both objects and motions, and therefore can be applied to a variety of applications, including the detection of human faces and bodies, as well as report their positions.

Below are the key advantages of the LIOD sensors:
1. Provide motion detection functionalities with a power consumption rate of less than 100μW, making them applicable on always-on devices
2. Integrated with a Hard-Wired Deep Learning Engine that helps to save a large amount of computing cost on the host side
3. Instead of outputting bulky image data that occupies data bandwidth, the sensor can directly deliver readily usable information (e.g. object position, size, etc) in real-time
Low-Power Intelligent Object Detection
Part No. Type
Package Size
Report Rate
Control Interface
Image Output
Power Consumption under Human Sensing Mode
Power Consumption under Human Detection Mode
More Info
Motion detection, human/face detection
WLCSP: 4.58*4.58
Parallel 8-bit, JPEG