Closed-loop object recognition using reinforcement learning | IEEE Conference Publication | IEEE Xplore