PPA Model Inversion Attack on FaceScrub (224x224, private) + FFHQ (public) (test) benchmark leaderboard · SOTA2 Research

PPA Model Inversion Attack on FaceScrub (224x224, private) + FFHQ (public) (test)

98.02AccuracyMIDRE(0.1,0.4)

Evaluation Results

Method	Links
MIDRE(0.1,0.4)2024.09		98.02	43.58	139.01	0.81	—
MIDRE(0.1,0.4)2024.09		97.9	42.44	139.66	0.82	—
MIDRE(0.1, 0.4)2024.09		97.28	48.16	131.72	0.8	—
MIDRE(0.1,0.8)2024.09		95.47	15.97	155.61	0.95	—
NoDef2024.09		95.43	86.51	113.03	0.73	—
MIDRE(0.1,0.8)2024.09		95.15	15.47	155.8	0.96	—
NoDef2024.09		94.86	83	128.6	0.76	—
NLS(-0.05)2024.09		94.79	33.14	130.94	0.9	712.29
NoDef2024.09		94.22	88.67	123.85	0.74	—
TTS2024.09		94.16	77.26	—	—	8.2
TTS2024.09		93.97	73.59	—	—	8.85
SSF2024.09		93.79	70.71	—	—	9.63
SSF2024.09		93.79	71.06	—	—	11.16
MIDRE(0.1,0.8)2024.09		93.33	13.89	154.79	0.97	84.02
RoLSS2024.09		93	64.98	—	—	8.86
MID2024.09		92.7	82.08	122.96	0.76	0.43
RoLSS2024.09		92.4	58.68	—	—	9.89
BiDO2024.09		91.8	58.14	147.28	0.87	7.82
MID2024.09		91.56	66.18	137.18	0.86	5.25
NLS2024.09		91.5	14.34	—	1.23	18.36
DP2024.09		91.36	74.88	131.38	0.82	2.32
BiDO2024.09		91.33	76.56	127.86	0.75	4.54
MID2024.09		91.15	65.47	137.75	0.87	7.56
TL-DMI2024.09		91.12	22.36	—	—	21.39
BiDO2024.09		90.31	67.07	139.15	0.84	3.5
TL-DMI2024.09		90.1	31.82	—	—	10.75
DP2024.09		89.8	75.26	130.41	0.82	3.03