Table 3.

Benchmark using 1-NN (266 classes).

Features	Acc.	Pr	Re	F1
Hist L	21.62	13.69	14.79	14.22
Hist RGB	47.58	40.33	39.22	39.77
3 marg. hist.	42.86	33.44	35.22	34.30
qHist RGB	47.89	40.51	39.82	40.16
sHist RGB	31.61	31.82	23.60	27.10
Chrom. Mom.	21.75	16.39	16.85	16.62
Cooc. Matr.	9.31	4.80	4.76	4.78
SFTA	15.52	10.12	10.12	10.12
Granulometry	32.53	27.37	28.24	27.80
GIST	34.70	29.89	29.24	29.56
DT-CWT	31.71	26.47	27.56	27.00
CEDD	28.62	21.96	22.07	22.02
HOG	17.36	13.63	12.71	13.15
Gabor	36.39	31.33	31.15	31.24
LBP	33.87	29.82	29.98	29.90
LBP-nri	38.11	32.04	31.68	31.86
LBP-LCC	36.92	30.30	30.21	30.25
vgg16	70.63	68.01	66.96	67.48
vgg19	70.81	67.06	67.07	67.06
squeezenet	68.09	62.47	63.62	63.04
Inception V3	76.73	71.56	70.65	71.10
Google Net	64.23	59.70	59.39	59.55
Resnet50 L	72.32	65.57	65.40	65.49
Resnet50	85.77	82.78	82.13	82.45