LOW LATENCY RECOGNITION AND REPRODUCTION OF NATURAL GESTURE TRAJECTORIES

摘要: In human-machine interaction scenarios, low latency recognition and reproduction is crucial for successful communication. For of general gesture classes it important to realize a representation that insensitive with respect the variation performer specific speed development along trajectories. Here, we present an approach learning speed-invariant models provide fast convenient We evaluate our model data set comprising 520 examples 48 classes. The results indicate able learn gestures from few observations high accuracy.