Deep Reinforcement Active Learning for Human-in-the-Loop Person Re-Identification | IEEE Conference Publication | IEEE Xplore