Image Retrieval Using Low Level and Local Features Contents: A Comprehensive Review

Table 4

Average image retrieval accuracy using various feature extraction and description techniques for the augmented Wang dataset.

Category	Feature extractor-feature descriptor
	BRISK-BRISK [63]	BRISK-FREAK [63, 65]	ORB-ORB [64]	MSER-SURF [48, 53]	FAST-FREAK [51, 65]	SURF-FREAK [48, 65]	SURF-SURF [48]
Tribe	1.09	7.32	2.41	14.39	15.84	24.55	22.18
Beach	1	3.35	1.27	14.81	9.18	18.37	16.33
Monuments	1.12	6.3	1.49	15.37	11.26	12.36	13.29
Bus	1	7.74	1.6	21.2	21.3	10.55	37.77
Dinosaurs	1.17	14.6	22.95	28.8	27.48	26.58	57.25
Elephant	1.25	5.35	2.25	13.63	11.88	7.79	17.14
Roses	1	3.66	1.72	15.82	7.4	10.92	21.5
Horses	1.32	3.24	1.92	15.77	8.43	5.85	13.51
Mountains	1.01	5.25	1.22	9.9	6.13	5.19	11.2
Food items	1.02	5.74	1.31	8.2	6.03	5.3	11.37
Aeroplane	1.24	4.38	1.71	14.29	13.74	6.3	21.61
Average retrieval accuracy (%)	1.11	6.08	3.62	15.65	12.61	12.16	22.10