Registration of infrared and visible light image based on visual saliency and scale invariant feature transform

EURASIP Journal on Image and Video Processing

Table 2 The registration result comparison with some algorithms

	Registration algorithm	Matching pairs	Accurate pairs	Accuracy rate (%)	Time (s)
Fig. 4a	SIFT	18	13	72.22	0.715
	GLOH	19	15	78.95	0.652
	Harris	33	22	66.67	0.578
	The proposed algorithm	9	8	88.89	0.101
Fig. 4b	SIFT	32	27	84.38	1.312
	GLOH	30	25	83.33	1.208
	Harris	48	33	68.75	1.157
	The proposed algorithm	21	20	95.24	0.172
Fig. 4c	SIFT	24	19	79.17	1.138
	GLOH	26	21	80.77	1.087
	Harris	43	29	67.44	0.958
	The proposed algorithm	12	10	83.33	0.112
Fig. 4d	SIFT	38	31	81.58	1.498
	GLOH	41	34	82.93	1.511
	Harris	62	44	70.97	1.482
	The proposed algorithm	26	23	88.46	0.187
Fig. 4e	SIFT	21	15	71.43	0.907
	GLOH	19	15	78.95	0.843
	Harris	30	19	63.33	0.829
	The proposed algorithm	10	8	80	0.105
Fig. 4f	SIFT	27	20	74.07	1.226
	GLOH	28	22	78.57	1.138
	Harris	48	34	70.83	1.082
	The proposed algorithm	18	16	88.89	0.167