Learning to recognise 3D human action from a new skeleton‐based representation using deep convolutional neural networks | Publicación