People who cannot use their hands with full efficiency may find it extremely difficult to interact with touchscreen devices. Zack Freedman thus came up with the Hypervisor, which is a wearable device that tracks the user’s eye movements and pupil positions to move a mouse cursor around on a device. This tool enhances the accessibility of electronics for those who cannot use their hands effectively. Zack Freedman entered the Hackaday Prize contest with the device to serve a non-profit organization “United Cerebral Palsy”.
The hypervisor runs on a Raspberry Pi Compute Module 3+ installed on a StereoPi baseboard. Sensors on the device all relay data to Raspberry Pi and it processes the data input. It has a pair of infrared transmitters and receivers which process device information and gaze data. Two CSI (Camera Serial Interfaces) give environmental data along with eye data to aid computer vision processing. The user gets to know about the state of the machinery, whether it’s working or not etc. with some visible spectrum LEDs.
IR CSI cameras feed a continuous video stream to the Raspberry Pi Compute Module 3+ running OpenCV, which is a library focused on real-time-computer-vision to determine where you are looking and it uses another camera to advance the accuracy of several things like where your eye pupil is pointing at and where you are looking at.
The device then determines where you are looking at and places the cursor there after the completion of all these calculations. The device allows you to click with a blink, squint to drag and drop and blink twice in fast succession to scroll up and down. These devices are connected to a dedicated receiver that brings in data from the headset and it knows where to position the cursor. The infrared camera of Hypervisor pointed at your eye, an IR transceiver and a full-colour camera on your forehead identify exactly where you’re looking and information about the device you are using.
The Hypervisor is an eye-tracking headband powered by a Raspberry Pi which is competent enough to control an entire room of computers, tablets and phones just by looking at them.