3D human action recognition and style transformation using resilient backpropagation neural networks

This paper addresses the problem of 3D human action class and style class recognition as well as style transformations using Artificial Neural Networks. The training process is selected uniquely to suit the problem and a quantitative evaluation method is proposed for the results. Few other intelligent methods have also been applied for recognition and compared to our original approach. The results demonstrate the high classification and transformation precision of our method, while both tasks are performed using the same system.