Table 3 Test accuracy comparison of... | Emerald Publishing

Table 3

Test accuracy comparison of different detectors for JPEG compressed images.

Method	ap2or	hor2zeb	win2sum	citysc.	facades	map2sat	Ukiyoe	Van Gogh	Cezanne	Monet	Average
Steganalysis feat.	79.39	90.02	56.66	92.17	73.62	69.39	65.83	95.30	94.73	80.89	81.09
GAN discr.	63.29	91.08	51.90	53.14	88.75	79.35	76.56	80.32	96.41	81.83	73.33
Cozzolino2017	79.57	89.82	53.74	86.81	62.88	89.64	67.67	98.80	99.93	87.33	82.62
Bayar2016	54.64	95.34	50.27	54.00	90.63	52.69	58.90	74.27	99.77	78.60	69.17
Rahmouni2017	84.96	98.35	54.30	57.60	91.88	54.93	96.83	99.63	99.77	89.72	80.97
DenseNet	78.27	93.44	66.94	97.83	98.19	80.45	97.54	98.53	99.57	83.95	88.51
InceptionNet v3	78.60	95.23	64.54	96.09	90.14	63.84	99.53	96.31	100.00	86.21	87.37
XceptionNet	93.52	93.77	67.07	95.11	99.22	67.97	99.66	95.18	99.97	84.02	89.03
RGGID	69.04	91.52	94.42	95.38	93.71	62.83	95.18	98.36	99.55	84.54	88.45

Method	ap2or	hor2zeb	win2sum	citysc.	facades	map2sat	Ukiyoe	Van Gogh	Cezanne	Monet	Average
Steganalysis feat.	79.39	90.02	56.66	92.17	73.62	69.39	65.83	95.30	94.73	80.89	81.09
GAN discr.	63.29	91.08	51.90	53.14	88.75	79.35	76.56	80.32	96.41	81.83	73.33
Cozzolino2017	79.57	89.82	53.74	86.81	62.88	89.64	67.67	98.80	99.93	87.33	82.62
Bayar2016	54.64	95.34	50.27	54.00	90.63	52.69	58.90	74.27	99.77	78.60	69.17
Rahmouni2017	84.96	98.35	54.30	57.60	91.88	54.93	96.83	99.63	99.77	89.72	80.97
DenseNet	78.27	93.44	66.94	97.83	98.19	80.45	97.54	98.53	99.57	83.95	88.51
InceptionNet v3	78.60	95.23	64.54	96.09	90.14	63.84	99.53	96.31	100.00	86.21	87.37
XceptionNet	93.52	93.77	67.07	95.11	99.22	67.97	99.66	95.18	99.97	84.02	89.03
RGGID	69.04	91.52	94.42	95.38	93.71	62.83	95.18	98.36	99.55	84.54	88.45