2 ms·

I had the opportunity to get some hands on time with the Leap before Christmas. Leap Motion sponsored a hackathon at my school and brought in dev boards for ev

I had the opportunity to get some hands on time with the Leap before Christmas. Leap Motion sponsored a hackathon at my school and brought in dev boards for everyone to borrow, although they were not the completed product like is pictured in this article. I cannot tell you how incredible this product is. I'm a first year CS student, and I've never done anything even remotely close to gesture tracking before. But at the end of the night, I was able to play rock paper scissors with my computer. The API is that simple to use. Yet, as mentioned, it's so incredibly accurate. One of the biggest bug we faced was even when we thought our hands were still, the device still registered imperceptible movements which were translated into false moves. Overall it's a great product, especially for the price.

defrost 14y ago
There's a paper [1] that looks back at a decade of gesture tracking research and talks about the end system: "an automatic Australian sign language (Auslan) recognition system, which tracks multiple target objects (the face and hands) throughout an image sequence and extracts features for the recognition of sign phrases." which achieves "Using a known grammar, the system achieved over 97% recognition rate on a sentence level and 99% success rate at a word level." While this is about the extraction of finger gestures from 2D video, many of the techniques translate forward and are applicable to 3D point cloud data (if that is what the raw leap motion device exposes). [1] http://link.springer.com/article/10.1007%2Fs00138-005-0003-1 http://link.springer.com/article/10.1007%2Fs00138-005-0003-1 Should be a fun read, other papers by the same authors are also relevant to the problem domain. EDIT: Some form of running debouncing filter should kill the jitter- the Savitzky Golay filter family is good for that.

defrost 14y ago

There's a paper [1] that looks back at a decade of gesture tracking research and talks about the end system: "an automatic Australian sign language (Auslan) recognition system, which tracks multiple target objects (the face and hands) throughout an image sequence and extracts features for the recognition of sign phrases." which achieves "Using a known grammar, the system achieved over 97% recognition rate on a sentence level and 99% success rate at a word level." While this is about the extraction of finger gestures from 2D video, many of the techniques translate forward and are applicable to 3D point cloud data (if that is what the raw leap motion device exposes). [1] http://link.springer.com/article/10.1007%2Fs00138-005-0003-1 http://link.springer.com/article/10.1007%2Fs00138-005-0003-1 Should be a fun read, other papers by the same authors are also relevant to the problem domain. EDIT: Some form of running debouncing filter should kill the jitter- the Savitzky Golay filter family is good for that.