In daily life and communication, people usually recognize the emotions of people around them through facial expressions, voice dialogues, body language and other external behavioral signals, so as to ...

Abstract: This article tackles a new problem setting: reinforcement learning with pixel-wise rewards (pixelRL) for image processing. After the introduction of the deep Q-network, deep RL has been ...