Funny you should say that - when I was working @Intel I wrote a proposal for a Bluetooth Low Power ring that provide a simple input device for headless devices with 1 rotary input and one press input, sufficient to scroll and click on items.
Another proposal I had was a bracelet that would sense capacitance changes in the hand upon fingers touching each other; this way you could have a 12-key "keyboard" on the phalanges of the non-opposable fingers (3 phalanges x 4 fingers) touch-able by the thumb.
Such technologies would require minimal power input and provide good interaction with any headset; but at the time Intel were not interested in the research needed to build a prototype.
Speaking of ring interfaces, Ring [0] was a Kickstarter project that happened. I seem to recall that it did ship but didn't actually work too well. I suspect that your simpler interface would have been a lot more successful.
I think some of it depends on which finger you're targeting, too. Wearing it on the index finger makes it easier to tap/push buttons using the thumb.
In my HCI class, we went over this paper [1] (2010). It has a wide variety of always-available input and output and many of them would work well with these glasses. My hardware crush is a Myo [2], although I don't know how well it works. The videos of it look neat, at least :).
Also, I had the exact same idea for a hand interface as a project for my Ubicomp class, but I didn't have the skills to do it. I ended up doing research for "smart jewelry" instead :|
I found this use of the Myo (to control the robotic arm of an amputee) really cool -- it seems like they found a life-changing use for what might have ended up a cool toy -- http://developerblog.myo.com/meet-the-man-with-a-myo-control... -- not sure how perfectly it works though.
Probably not particularly well as the Myo band only detects a handful of super basic gestures, and unless they have changed their stance in the past year or two they refuse to expose the raw data to developers. This makes it impossible to distinguish more than their predefined gestures.
The main problem I'm seeing is the extreme variance of the capacitance between users and the variance across the same user during usage - some smart algorithms that constantly re-calibrate and infer what's touched are needed.
Another proposal I had was a bracelet that would sense capacitance changes in the hand upon fingers touching each other; this way you could have a 12-key "keyboard" on the phalanges of the non-opposable fingers (3 phalanges x 4 fingers) touch-able by the thumb.
Such technologies would require minimal power input and provide good interaction with any headset; but at the time Intel were not interested in the research needed to build a prototype.