A growing neural gas (GNG) model for hand gesture recognition

A novel growing neural gas (GNG) based the learning topology, called the flexible node insertion GNG (FNIGNG), as well as its application to hand gesture representation are introduced in this paper. Compared with the fixed node insertion policy used in the typical GNG network, our proposed network applies a flexible node insertion scheme by analyzing the network mean squared error (MSE). Experimental results show that our approach is able to shorten the training time while not impairing the training accuracy.