Depth Estimation on NYUv2 1 (test) benchmark leaderboard · SOTA2 Research

Depth Estimation on NYUv2 1 (test)

0.109RMSEOurs (direct optimization approach)

Evaluation Results

Method
Ours (direct optimization approach)2026.02	0.109	0.837	99.2	99.4	99.5
Carvalho et al.2026.02	0.144	3.6	—	—	—
Song & Lee2026.02	0.154	2.8	—	—	—
Yang et al. (DFF-DFV)2026.02	0.232	44.5	96.7	98	99
Nazir et al.2026.02	0.244	2.9	91.4	97.9	99.5
Yang et al. (DFF-FV)2026.02	0.285	47	95.6	97.9	98.8
Si et al.2026.02	0.325	17	95	97.9	98.7
Anwar et al.2026.02	0.347	9.4	—	—	—
Chang & Wetzstein2026.02	0.433	8.7	93	99	99.9
Fu et al. (DORN)2026.02	0.509	11.5	82.8	96.5	99.2
Dong et al.2026.02	0.537	14.6	79.9	95.1	98.8
Tang et al.2026.02	0.579	13.2	82.6	93.6	99.2
Xu et al. (MS-CRF)2026.02	0.586	12.1	81.1	95.4	98.7
Lu et al. (Focus-Net)2026.02	0.611	17.2	74.8	89.2	94.9
Lu et al. (Defocus-Net)2026.02	0.623	17.6	73.2	88.7	95.1
Li et al.2026.02	0.635	14.3	78.8	95.8	99.1
Gur & Wolf2026.02	0.649	18.4	72	88.7	95.1
Suwajanakorn et al.2026.02	0.95	25	68.8	80.2	91.7
Moeller et al.2026.02	0.985	26.3	67	77.8	91.2