Specialized keyboards for paralyzed patients might use solutions like eye-tracking systems, sip-and-puff switches, and mouth/head sticks. Here, I've attempted to make a keyboard that uses nose placement and head tilt to select letters, then text to speech to hear the results. Hover your nose over a key (it will turn brighter when selected) and tilt your head temporarily to the right to confirm your selection of the letter/add it to your message. Do the same to select the delete last, delete all, and play buttons. I tried to build in some safeguards to prevent "keysmashing," but it can get tricky! You might try moving farther away from the camera or make sure you are holding your head upright and as still as you can between selections. It might take some debugging on the user's part to figure out what lighting conditions, distance from the camera, etc., yield the best results. But I like that it makes us think about how we deal with the limitations of our options or the limitations of the user. I went with head tilt to select so the user wouldn't need to use their hands or a physical keyboard or mouse at all. But loudness or some other input might be fun to try in a remix! What are the pros and cons of different approaches?
See inside for where I've added some comments. Thanks to Scratcher user testers, who provided some great points and constructive feedback in the comments below! I love the helpfulness of the community! Like this? You may also want to check out my project "Face Sensing and micro:bit Pairing Experiment" https://scratch.mit.edu/projects/1232277018 or "Dem Bones" https://scratch.mit.edu/projects/1229361456 When you use Face Sensing blocks, only your computer can sense your face. None of your data is stored or sent to Scratch or any other site, making it a safe, fun, and creative way to explore the possibilities of AI. You can check out my video tutorial for Face Sensing blocks on the official Scratch YouTube channel and find a link to written resources related to this topic in the video description.
