Crowd Counting and Abnormal Behavior Detection via Multiscale GAN Network Combined with Deep Optical Flow

Table 2

Quantitative result for different benchmark datasets.

Models	Part_A		Part_B		UCF_CC_50
	MAE	MSE	MAE	MSE	MAE	MSE
TEDNet	64.85	109.25	88.25	12.85	249.25	354.62
PACNN	66.18	106.24	21.552	3.56	267.9	357.06
SANet	66.39	104.51	8.54	13.65	258.44	334.86
CSRNet	68.27	115.27	21.62	33.52	266.15	397.17
ACSCP	76.58	102.75	17.55	27.58	292.08	404.87
S_CNN	90.2	135.22	8.922	135.18	318.27	439.57
MCNN	110	173.25	8.28	12.84	377.78	509.21
Proposed	63.4	106.23	8.21	12.66	266.71	310.65