Deepfake Detection on GANs Out-of-domain (test) benchmark leaderboard · SOTA2 Research

Deepfake Detection on GANs Out-of-domain (test)

96.8AccuracyμFlow

Evaluation Results

Method	Links
μFlow2026.06		96.8	96.8	96
μFlow2026.06		94.7	94.7	95.3
μFlow2026.06		94.7	94.7	95.3
μFlow2026.06		93.6	93.6	95
μFlow2026.06		93.5	93.6	94.9
ODDN2026.06		92.7	91.5	93.5
D32026.06		92.6	94.5	94.2
ODDN2026.06		91.8	85.3	83.5
ODDN2026.06		91.4	87	88.5
ODDN2026.06		91	88.8	83.8
D32026.06		90.4	92.5	93.7
μFlow2026.06		90.3	90.4	93.9
μFlow2026.06		90.3	90.4	93.9
μFlow2026.06		90.3	90.4	93.9
μFlow2026.06		90.3	90.4	93.9
D32026.06		86.8	90.7	92.5
LAA-Net2026.06		86	88.2	94.2
D32026.06		85.4	87.3	87.5
NPR2026.06		85.3	94	91.7
ODDN2026.06		85.2	71.2	70.3
AUNet2026.06		84.8	87	93.3
ODDN2026.06		84.5	79.2	75.9
SBI2026.06		84.3	86.8	93
PCL+I2G2026.06		84.1	86	92.2
NPR2026.06		83.6	83.2	85.8
NPR2026.06		83	89.8	89.5
LAA-Net2026.06		83	84.9	86.2
ODDN2026.06		81.8	83.6	85
D32026.06		80.9	82	87.9
FreqNet2026.06		79.9	86	87.5
FreqNet2026.06		79.8	81.2	82.6
AUNet2026.06		78.1	80.2	77.9
DFX-SN2026.06		77.8	80.1	80.4
SBI2026.06		77.8	79.8	77.1
FreqNet2026.06		77.6	80	79
PCL+I2G2026.06		77.6	79.2	76.6
NPR2026.06		77.4	78.2	79.7
NPR2026.06		77.2	81	80.9
FreqNet2026.06		77.1	81.3	83
FreqNet2026.06		76.7	74.3	75.4
DFX-SN2026.06		75.6	75.6	75.8
DFX-SN2026.06		75.2	77.9	77.1
D32026.06		75.1	83.4	85
NPR2026.06		75.1	68.7	73.5
DFX-SN2026.06		74.1	75.8	76
DFX-SN2026.06		73.3	70.1	68
FreqNet2026.06		71.8	67.6	70.7
D32026.06		71.1	76.6	77.9
DFX-SN2026.06		69.9	70.4	68.6
NPR2026.06		68.5	68.5	69.1
FreqNet2026.06		68.3	66	65.3
DFX-SN2026.06		67.9	67	66.1
AFSD2026.06		67.3	69	69.3
AFSD2026.06		65	64.4	68.8
OC-FakeDect2026.06		57.4	58.1	56.5
OC-FakeDect2026.06		48.9	51.1	50