Table 2. Performance analysis using adaptive or weighted feature fusion of SURF-FREAK descriptors on the Corel-1000 image collection (bold values indicate best performance).

Weight W of SURF-FREAK	MAP performance (in %) on the different sizes of the dictionary
	20	50	100	200	400	600	800	1000	1200
1.0–0.0	64.44	66.80	69.67	69.89	70.34	72.79	70.26	69.66	69.54
0.9–0.1	65.03	66.91	70.16	70.31	70.82	72.89	72.16	72.81	69.21
0.8–0.2	65.14	67.11	70.23	70.43	70.91	72.99	72.81	72.99	68.02
0.7–0.3	65.23	67.24	70.29	70.63	71.23	73.11	73.02	73.39	67.91
0.6–0.4	65.45	67.38	70.38	70.71	71.41	73.23	73.59	73.82	67.73
0.5–0.5	65.51	67.48	70.48	70.84	71.63	73.51	73.81	73.98	67.71
0.4–0.6	65.59	67.88	70.83	71.03	71.89	73.83	73.89	74.05	67.68
0.3–0.7	65.43	67.83	70.81	70.98	71.83	73.78	73.51	74.01	67.66
0.2–0.8	65.39	67.76	70.62	70.91	71.65	73.68	73.32	73.93	67.64
0.1–0.9	64.21	67.09	70.06	70.86	71.51	73.53	72.08	73.82	67.62
0.0–1.0	61.99	64.01	66.89	67.88	66.40	67.68	69.99	66.41	67.60