Table 2.

Results of the proposed method in the defined evaluation sequences

Seq	Cam ID	MOTA	MT	PT	ML	FRAG	FRAG ratio	P	R	F1	APE
1	1,2	1.000	8	0	0	0	0.000	1.000	1.000	1.000	0.094
1	1	1.000	7	0	0	0	0.000	1.000	1.000	1.000	0.089
1	2	1.000	8	0	0	0	0.000	1.000	1.000	1.000	0.069
2	1,2	0.994	12	0	0	0	0.000	0.994	1.000	0.997	0.151
2	1	1.000	9	0	0	0	0.000	1.000	1.000	1.000	0.071
2	2	0.993	11	0	0	0	0.000	0.993	1.000	0.997	0.063
3	1,2	0.965	5	0	0	6	0.061	0.998	0.977	0.987	0.136
3	1	0.973	9	0	0	8	0.081	0.999	0.984	0.991	0.084
3	2	0.966	10	1	0	2	0.021	0.999	0.975	0.986	0.077
4	1,2	0.966	5	0	0	6	0.061	0.983	0.989	0.986	0.155
4	1	0.997	25	0	0	1	0.010	1.000	0.999	0.999	0.085
4	2	0.983	22	1	0	6	0.061	1.000	0.986	0.993	0.095
5	1,2	0.975	10	0	1	3	0.031	0.999	0.984	0.991	0.186
5	1	0.952	7	1	0	11	0.115	0.999	0.957	0.978	0.097
5	2	0.991	12	1	0	1	0.010	1.000	0.993	0.996	0.085
6	1	0.982	6	0	0	3	0.031	0.999	0.983	0.991	0.089
7	1	0.935	13	1	0	15	0.157	0.986	0.957	0.971	0.112

The considered metrics, described in Section 4.1, measure the quality of the generated tracks and the accuracy of the positioning. For multi-camera sequences, we also run the system for each camera separately