An Adaptive Smoothness Parameter Strategy for Variational Optical Flow Model

Table 2

The AEE/AAE of DFOVOFM with different smoothness parameters in Middlebury.

	Algorithm	Army	Mequn	Scheflera	Woden	Grove	Urban	Yosemite	Teddy
AEE		0.08	0.19	0.27	0.16	0.52	0.40	0.16	0.40
		0.10	0.17	0.25	0.18	0.61	0.36	0.20	0.37
		0.10	0.19	0.23	0.20	0.64	0.30	0.22	0.43
		0.11	0.21	0.24	0.25	0.69	0.34	0.30	0.47
		0.13	0.23	0.28	0.26	0.73	0.42	0.35	0.53
		0.13	0.25	0.30	0.32	0.78	0.51	0.40	0.58
	OSP	0.08	0.16	0.19	0.15	0.45	0.29	0.15	0.35
AAE		2.67	2.13	2.28	2.75	2.34	2.68	2.25	1.46
		2.69	2.03	2.16	2.93	2.39	2.56	2.36	1.33
		2.73	2.15	2.04	3.14	2.43	2.45	2.48	1.52
		2.85	2.24	2.20	3.22	2.54	2.57	2.73	1.73
		2.92	2.36	2.34	3.45	2.65	2.72	2.92	1.95
		3.04	2.53	2.46	3.62	2.77	2.86	3.05	2.16
	OSP	2.63	1.98	1.96	2.71	2.26	2.42	2.13	1.25

OSP is the result of our own algorithm, and other bold values are the lowest one (using the fixed smoothness parameter) for each sequence.