Single-Object Tracking Algorithm Based on Two-Step Spatiotemporal Deep Feature Fusion in a Complex Surveillance Scenario

Table 6

Quantitative comparison for different sequences.

Sequences	Success rate							Center location error
	KCF	SAMF	DSST	DKCF	CFNet	SiamRPN	Our	KCF	SAMF	DSST	DKCF	CFNet	SiamRPN	Our
Video 8	0.78	0.56	0.62	0.70	0.75	0.79	0.86	25.3	42.3	24.4	24.9	17.0	14.4	9.1
Video 7	0.60	0.48	0.65	0.61	0.78	0.81	0.81	21.7	28.6	9.2	37.4	21.2	10.8	7.5
Video 6	0.79	0.67	0.58	0.82	0.67	0.79	0.79	12.7	7.1	6.3	12.6	16.8	15.6	9.3
Video 5	0.61	0.52	0.54	0.77	0.47	0.40	0.42	14.6	25.0	11.4	17.1	15.2	2.9	8.1
Video 4	0.73	0.68	0.71	0.75	0.75	0.88	0.82	27.3	22.3	14.4	24.9	17.0	14.9	10.1
Video 3	0.68	0.70	0.72	0.76	0.79	0.83	0.87	31.7	28.3	9.2	37.4	21.2	10.8	8.4
Video 2	0.59	0.57	0.62	0.65	0.75	0.81	0.85	12.5	9.1	6.3	12.6	16.8	15.6	8.9
Video 1	0.68	0.62	0.66	0.72	0.61	0.79	0.79	74.6	23.0	21.4	27.1	15.2	19.9	18.2