Knowledge-based disambiguation of hand gestures

An algorithm is presented in this paper to disambiguate sequences of hand gestures by utilizing their overall meaning. This algorithm relieves the traditional reliance on a potentially complex syntactical analysis step by sharing this burden with higher level processing. A minimal set of concurrent primitives are used to represent hand gestures, improving scalability while reducing the complexity of training and recognition. These gesture primitives are synthesized into concepts and associated with a knowledge base using an approximate graph matching technique to determine their overall meaning. Initial experimental results have shown that this technique is able to successfully disambiguate one or more hand gestures in sequences of 2-5 noisy gestures, resulting in an improved understanding of the overall meaning.