Multi-Scale Guided Attention Network for Crowd Counting

Figure 6

ShanghaiTech dataset visualization results show that the first line is the original picture, the second line is the ground truth, the third line is the prediction result of the model, and the fourth line is the true value/predicted value.