Gesture recognition to get cheaper, more private with STMicroelectronics sensors
Gesture sensors using laser rather than cameras and computer vision are becoming cheaper and less energy intensive. STMicroelectronics has launched an affordable turnkey solution with free engineering software to get its gesture sensors into more devices.
Demand for touchless gesture control is rising in part thanks to the COVID-19 pandemic, as with biometric technologies produced by the firm. STMicroelectronics is now tackling affordability and privacy issues with the turnkey STGesture recognition system for touchless controls.
The firm’s FlightSense time-of-flight (ToF) multizone ranging sensor can track a hand front of it in three dimensions. Using invisible infrared light to illuminate a scene for an array of lasers, the sensor works by detecting presence and movement across an 8×8 grid of squares. Users can focus this on a 4×4 grid for faster gestures.
The algorithm interprets movement across the grid to determine gestures which correspond to controls. Waving, swiping, tapping, double tapping, pushing and pulling can all be detected.
Use cases span everything from vending machines and laptops for giving presentations to industrial robots to understand humans better. The sensors are more popular in situations where users’ hands are dirty or there are dangerous objects, but the low costs and low power needs could see the sensors adopted more widely.
With no cameras capturing images, privacy is protected. With no radar and camera set up, data quantities are much reduced and there is no requirement for external illumination. No AI is required for learning gestures as they can be pre-programmed, also reducing processing load.
The engineering software has been made free of charge and a library of example code snippets reduces design time for integration. The software also allows for integration with other hardware such as webcams, possibly allowing for biometric recognition alongside gesture.