Maxim Integrated Products is looking to make battery-powered cameras more efficient through a new partnership with Xailient. To that end, Maxim has integrated Xailient’s Detectum neural network into its own MAX78000 ultra-low power microcontroller for use in edge devices.
The Detectum neural network allows a camera to detect faces and other objects that appear in its field of view. It needs only 12 milliseconds to lock in on a face, and can focus and zoom in on that face in real time once it has done so.
The key selling point, however, may be Detectum’s power-saving capabilities. The solution draws only 280 microJoules while it is in use, a figure that is 250 times lower than comparable solutions. As a result, it extends the battery life of cameras that have been outfitted with the tech, which in turn means that venue operators will need to charge their devices less frequently. Detectum itself also offers a ‘listening’ mode utility, which shuts down more complex functions until a face is detected to further conserve battery life.
The MAX78000 microcontroller is intended for use in security cameras and other edge devices. Detectum will let users set up a visual wake word, and the system can be configured to detect objects other than faces. For example, store managers could use the system to track inventory, while an entertainment venue could use it to count empty parking spots.
“With Detectum, the MAX78000 is capable of both classification and localization, so in addition to seeing faces in the image or video you can also determine where those faces are in the image’s field of view,” said Maxim Fellow and MAX78000 architect Robert Muchsel.
“AI is on track to be the second largest carbon emitting industry,” added Xailient CTO Shivy Yohanandan. “Replacing legacy Internet protocol cameras that use traditional cloud AI with edge-based cameras with Detectum has the equivalent carbon impact of taking one gasoline powered car off the road.”
Maxim is listing the MAX78000 on its website at $8.50. The company recently integrated B-Secur’s HeartKey algorithms with its MAX30003 ECG sensor and MAX32660 microcontroller to enable heart rate detection in wearable devices. It also released a new blood-oxygen saturation solution in February of 2020.
(Originally posted on Mobile ID World)