Multiobject Tracking in Videos Based on LSTM and Deep Reinforcement Learning

Figure 3

The typical framing of a reinforcement learning.