Exploiting semantic segmentation to boost reinforcement learning in video game environments | Publicación