Adaptive feature descriptor selection based on a multi-table reinforcement learning strategy | Publicación