Talking Head Synthesis on Portrait Video Self-reconstruction (test) benchmark leaderboard · SOTA2 Research

Talking Head Synthesis on Portrait Video Self-reconstruction (test)

35.15PSNRHM-Talker

Evaluation Results

Method	Links
HM-Talker2025.08		35.15	0.0207	0.9971	2.514	0.53	0.22	7.807	0.6	120
SyncTalk2025.08		34.51	0.0221	0.9959	2.607	0.55	0.29	7.502	2	52
TalkingGaussian*2025.08		33.75	0.0273	0.9963	2.699	0.59	0.26	6.451	0.55	139
GaussianTalker2025.08		32.69	0.0442	0.9952	2.726	0.67	0.59	6.234	3.2	95
TalkingGaussian2025.08		32.48	0.0309	0.995	2.616	0.6	0.28	6.246	0.5	108
ER-NeRF2025.08		32.47	0.0395	0.9658	2.639	0.62	0.54	6.531	2.1	31.2
RAD-NeRF2025.08		31.95	0.062	0.966	2.847	0.74	0.76	5.742	5.3	28.7
AD-NeRF2025.08		30.07	0.1042	0.9689	2.998	1.01	0.97	6.053	18.7	0.11