On the learning of complex movement sequences

We introduce a rule-based approach for the learning and recognition of complex movement sequences in terms of spatio-temporal attributes of primitive event sequences. During learning, spatio-temporal decision trees are generated that satisfy relational constraints of the training data. The resulting rules are used to classify new movement sequences, and general heuristic rules are used to combine classification evidences of different movement fragments. We show that this approach can successfully learn how people construct objects, and can be used to classify and diagnose unseen movement sequences.