Low-Rank and Sparse Based Deep-Fusion Convolutional Neural Network for Crowd Counting

Table 3

Performance comparison of crowd counting methods for Shanghaitech Part B dataset.

Network	Part B dataset
	MAE	MSE	MRE
ACF [3]	69.7	108.0	70.4
LBP + RR	59.1	81.7	69.2
LBP + LSSVM	48.3	67.8	57.6
Patch-CNN [11]	32.0	49.8	37.6
MCNN [12]	26.4	41.3	24.2
Patch-count CNN [25]	26.1	37.7	25.9
Patch-multitask CNN [26]	20.3	31.0	22.6
TSCCM [27]	21.7	32.4	20.26
Long-short CNN [13]	19.8	33.1	18.1
Hydra-CNN [14]	17.1	26.3	15.5
Deep-fusion network	17.3	28.9	16.4
Fusion + RR	20.5	28.7	18.9
Fusion + LSSVM	17.6	30.1	15.3
LFCNN	14.7	25.4	11.8