Table 2.

Performance comparisons in terms of center location error (CLE) (in pixels) and VOC overlap rate (VOR) (%). The best results are shown in bold and red fonts. MIL, multiple instance tracker; CT, compressive tracker; Frag, fragments-based tracker; APG, accelerated proximal gradient.

Sequence	CT		MIL		K-SVM		Frag		L1		L1APG		Proposed
	CLE	VOR	CLE	VOR	CLE	VOR	CLE	VOR	CLE	VOR	CLE	VOR	CLE	VOR
Car4	63.5	0.21	60.1	0.34	17.8	0.48	179	0.22	4.11	0.84	6.63	0.82	3.54	0.89
Car11	16.7	0.44	43.5	0.17	17.2	0.43	63.9	0.09	33.3	0.43	1.89	0.79	1.46	0.85
Singer1	16.1	0.34	15.2	0.34	31.6	0.38	22.1	0.34	4.57	0.71	4.43	0.69	4.11	0.83
Caviar1	16.7	0.52	48.2	0.26	3.79	0.69	5.53	0.68	119	0.28	56.4	0.29	1.95	0.82
Caviar2	66.3	0.29	69.8	0.25	5.62	0.71	5.64	0.56	3.34	0.79	62.1	0.35	2.34	0.85
Occlusion1	16.7	0.75	32.3	0.59	9.63	0.81	5.62	0.89	6.51	0.87	10.7	0.76	5.95	0.85
Occlusion2	17.3	0.58	14.1	0.61	7.94	0.65	15.5	0.61	11.1	0.67	10.4	0.72	4.32	0.82
Mhyang	13.3	0.61	20.3	0.51	5.12	0.81	12.5	0.65	3.45	0.77	5.27	0.72	3.25	0.79
Girl	18.8	0.31	13.7	0.41	3.23	0.71	20.6	0.46	3.34	0.68	3.54	0.66	3.26	0.71
DavidIndoor	15.7	0.45	18.3	0.43	6.25	0.54	82.1	0.18	18.9	0.45	21.5	0.36	5.41	0.61
Dudek	26.6	0.65	17.8	0.71	28.9	0.66	82.1	0.54	23.2	0.68	64.1	0.52	12.5	0.69
Dog1	6.96	0.54	7.82	0.54	6.26	0.66	12.1	0.55	3.75	0.66	3.41	0.74	3.18	0.83
CarScale	26.1	0.44	33.1	0.42	44.9	0.49	19.5	0.44	81.5	0.44	68.5	0.53	12.3	0.79
boy	9.03	0.51	12.8	0.51	2.64	0.72	40.5	0.39	6.98	0.74	7.21	0.77	2.33	0.84
Average	23.6	0.47	29.1	0.44	13.6	0.62	40.5	0.47	23.1	0.65	23.3	0.62	4.71	0.80