Design and analyze the structure based on deep belief network for gesture recognition
Miao Ma, Xidan Xu, Jie Wu, Min Guo · 2018
As an essential component for the usage of neural network, the structure of a neural network (e.g. the number of hidden layer and the number of units in each layer) plays an important role on its performance. So, the guidance for the designation of neural network is very necessary, especially for the deep neural network. In this paper, using American Sign Language dataset as an example, a greedy method was adopted for the designation of the structure of deep belief network for gesture recognition. The experiment results show that, compared with some classical gesture recognition methods, the designed DBN shown a better performance.