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

Table 4

Comparison of the different state-of-the-art methods on RGBT-CC dataset.

Model	+	MAE	MSE
MCNN [33]	N/A	21.89	37.44
SANet [69]	N/A	21.99	41.6
CSRNet [33]	N/A	20.4	35.26
Bayesian Loss [70]	N/A	18.7	32.67
MCNN + IADM [33]	N/A	19.77	30.34
MCNN [33]	+RTNA	18.04	29.16
SANet + IADM [33]	N/A	18.18	33.72
SANet [69]	+RTNA	17.98	32.04
CSRNet + IADM [33]	N/A	17.94	30.91
CSRNet [33]	+RTNA	17.82	30.14
Bayesian Loss + IADM [33]	N/A	15.61	28.18
Bayesian Loss [70]	+RTNA	15.48	27.96
Ours (RGBT-Net + AGK)	—	18.16	32.14