6DoF Face Pose Estimation on ARKitFace (test) benchmark leaderboard · SOTA2 Research

6DoF Face Pose Estimation on ARKitFace (test)

0.72Yaw ErrorPerspNet

Evaluation Results

Method	Links
PerspNet2022.05		0.72	1.1	0.54	0.79	0.92	1.47	9.59	3.99	9.99	—
PerspNet w/o LM2022.05		0.78	1.15	0.54	0.82	0.99	1.58	9.88	4.15	10.34	—
PerspNet2024.07		0.85	1.31	0.49	0.88	0.91	2.06	8.86	3.95	9.42	1.81
TRGv22024.07		0.85	1.26	0.52	0.87	0.8	1.9	8.23	3.64	8.76	1.76
PerspNet w/o PE2022.05		0.94	1.38	0.59	0.97	0.94	1.97	10.64	4.52	11.07	—
MPC2024.07		0.96	1.38	0.59	0.98	0.97	2.06	10.02	4.35	10.21	—
PerspNet2022.05		0.98	1.43	0.55	0.99	1	2.41	9.73	4.38	10.3	—
PerspNet2024.07		0.98	1.43	0.55	0.99	1	2.41	9.73	4.38	10.3	—
PerspNet2022.05		0.99	1.43	0.55	0.99	0.97	2.12	9.45	4.18	10.01	—
PerspNet w/o PE2022.05		1	1.52	0.62	1.05	1.18	2.39	12.14	5.24	12.71	—
PerspNet w/o PE2022.05		1.01	1.53	0.61	1.05	1.17	2.39	11.77	5.11	12.34	—
PerspNet w/o LM2022.05		1.04	1.45	0.59	1.03	1.14	2.21	10.84	4.73	11.45	—
PerspNet w/o LM2022.05		1.04	1.45	0.6	1.03	1.09	2.13	10.28	4.5	10.89	—
JMLR2024.07		1.13	1.75	0.61	1.16	0.98	2.48	11.13	4.86	11.87	2.39
Direct 6DoF Regress2022.05		1.86	2.72	1.03	1.87	2.8	5.23	19.16	9.06	21.39	—
Direct 6DoF Regress2024.07		1.86	2.72	1.03	1.87	2.8	5.23	19.16	9.06	21.39	—
Refined Pix2Pose2022.05		1.95	2.62	2.48	2.35	2.43	4.23	35.33	14	36.44	—
Refined Pix2Pose2024.07		1.95	2.62	2.48	2.35	2.43	4.23	35.33	14	36.44	—
img2pose (retrain)2022.05		5.07	7.32	4.25	5.55	1.39	3.72	15.95	7.02	20.54	—
img2pose2024.07		5.07	7.32	4.25	5.55	1.39	3.72	15.95	7.02	20.54	—