Multiobject Tracking in Videos Based on LSTM and Deep Reinforcement Learning

Figure 9

Qualitative tracking results of our tracker on TUD-Crossing.

(a)

(b)

(c)

(d)