Cam Cube

Hand-Held Camera Cube Reference Design Enables AI at the Edge

July 21, 2021
MAXREFDES178# camera cube executes low latency AI vision and hearing inferences on a coin cell power budget with reduced cost and size.

Maxim Integrated's MAXREFDES178# camera cube reference design serves Artificial intelligence (AI) applications that previously were limited to those with the resources and deep pockets. The camera cube puts AI Vision in space- and cost-constrained applications like battery-powered edge devices, giving ultra-low-power IoT solutions hearing and vision.  Highlighting the MAX78000 low-power microcontroller with neural network accelerator for audio and video, it also has the MAX32666 ultra-low power Bluetooth microcontroller and two MAX9867 audio CODECs. Provided in a compact form factor for AI applications such as facial identification and keyword recognition, the camera can also be used in low-power, cost-sensitive products like wearables.

he MAXREFDES178# camera cube enables time- and safety-critical applications to operate on even small batteries, using up to 1,000x less power for vision and hearing applications, compared to other embedded solutions. AI inferences running on the MAXREFDES178# show significant latency improvements, over 100x faster than on embedded microcontrollers. The 1.6in x 1.7in x 1.5in (41mm x 44mm x 39mm) solution is claimed to be up to 50% smaller than the next-smallest GPU-based processor, and doesn't require memories or complex power supplies.

To learn more about the MAXREFDES178# visit 
To learn more about the MAX78000 visit
To learn more about the MAX32666 visit 
To learn more about the MAX9867, visit 


To join the conversation, and become an exclusive member of Electronic Design, create an account today!