People use gestures to quickly express a concise response or intention. More and more devices incorporate sensors like accelerometers and gyroscopes that enable us to measure acceleration and respectively orientation. We should be excited as we could add cool new features like auto screen orientation or more… we could understand how people work and use their phone: we could detect and react to certain motion gestures and ease their lives. As great as it sounds, recognizing gestures is not a trivial task and requires theoretical and applied knowledge with probability and statistics.
Here’s a simple scenario that you’ve probably performed today. Your phone is sitting on the table and someone calls you – your phone is ringing. You pick it up half the way to look who’s calling you. You push the green answer button and take the phone to your ear to talk. Don’t you feel you’re doing something extra that you shouldn’t be doing?
Why do you push the green answer button? The phone should know when to automatically answer the call because you’ve taken the phone to your ear. You basically do two things and one is extra. When you push the answer button, you let the phone know that you intend to talk with the other person. But wait! You’re already making the gesture that you intend to answer the call by moving your phone to your ear. So why do you still need to push the answer button? You won’t need to!
I have already taken an in depth practical look over continuous density hidden Markov models applied on data measured by a 3-Axis ±3G accelerometer (Nintendo Wii Remote). I have implemented in C++ a prototype for my bachelor of computer science thesis. Luckily, OpenMoko FreeRunner will have 2 accelerometers.
The solution is based on continuous density hidden Markov Models. These statistical models have been successfully applied in speech recognition since 1989.