Research on 24-Hour Dense Crowd Counting and Object Detection System Based on Multimodal Image Optimization Feature Fusion

Table 8

Comparison of the different state-of-the-art methods on MICC dataset.

Model	MAE	MSE	+	MAE	MSE
MCNN [62]	1.5	2.259	+RDNA	1.205	1.719
MCNN-adaptive [61]	1.489	2.114	N/A	—	—
CSRNet [63]	1.359	2.125	+RDNA	1.195	1.723
CSRNet-adaptive [61]	1.343	2.007	N/A	—	—
RetinaNet [59]	1.641	2.554	+RDNA	1.489	1.962
DetNet [60]	1.541	2.382	+RDNA	1.356	1.861
Idrees et al. [61]	1.396	2.642	+RDNA	1.299	1.849
RDNet [65]	1.38	2.551	+RDNA	1.289	1.787
Cascaded-DCNet [64]	0.836	1.031	N/A	—	—
Ours (RGBD-Net + AGK)	1.025	1.521	—	—	—