List your company
ResearchFormerly WizWand

Video Depth Estimation on KITTI (Abs Rel, delta < 1.25)

0.036Abs Relpi^3
0.031480.061990.09250.12301Jul 15, 2025Sep 7, 2025Oct 31, 2025Dec 24, 2025Feb 16, 2026Apr 11, 2026Jun 4, 2026
Updated 5mo ago

Evaluation results

MethodLinks
pi^3Params=959M, Inference Speed Complexity=O(N^2), Alignment Protocol=joint scale-and-shift2026.030.03698.6
π³Align=scale & shift2026.020.03798.5
TrajVGAlign=scale & shift2026.020.03798.3
π³Align=scale2026.020.03898.6
pi^3Params=959M, FPS=57.42025.070.03898.6
π³FPS=110.22026.060.03998.5
TrajVGAlign=scale2026.020.0498.4
DA3 + Ours (GMM)Type=FA2026.060.04497.7
VGGT + Ours (GMM)Type=FA2026.060.04798
DA3 + Ours (LMM)Type=FA2026.060.04997.9
ZipMapParams=1.40B, Inference Speed Complexity=O(N), Alignment Protocol=joint scale-and-shift2026.030.0597.2
VGGTAlign=scale & shift2026.020.05296.8
OmniVGGTFPS=84.182026.060.05396.4
AetherAlign=scale & shift2026.020.05497.7
AetherAlign=scale2026.020.05697.8
AetherParams=5.57B, FPS=6.142025.070.05697.8
AETHERAlignment=Per-sequence scale, Online=false, Venue=ICCV’252026.030.05697.8
HorizonStream2026.050.05794.8
VGGTFPS=84.462026.060.05896.1
VGGTParams=1.26B, Inference Speed Complexity=O(N^2), Alignment Protocol=joint scale-and-shift2026.030.05996.1
VGGT2026.030.06197
VGGTType=Dense-view2025.070.06197
VGGT2026.050.06197
DA3Type=FA2026.060.06197.5
VGGTAlign=scale2026.020.06296.9
VGGTParams=1.26B, FPS=43.22025.070.06296.9
MoRetype=FA2026.030.06696.7
CogniMap3DCategory=Vision Foundation Model, Optim.=false, FF=true, Global Alignment (GA)=false, FPS=14.322026.010.06996.2
PPVDType=Diff2026.060.06997.6
VGGTAlignment=Per-sequence scale, Online=false, Venue=CVPR’252026.030.0796.5
VGGTCategory=Vision Foundation Model, Optim.=false, FF=true, Global Alignment (GA)=false, FPS=21.52026.010.07296.4
MoRetype=Streaming2026.030.07296.6
VGG-TAlignment=Per-sequence scale, Type=FA2025.120.07396.5
VGGTtype=FA2026.030.07396.1
VGGTType=FA2026.060.07396.5
MoE3DType=FA2026.060.07696
Stream3Rtype=Streaming2026.030.07994.9
Complet4R2026.030.07994.9
STream3R2026.050.0894.7
LiAuto-GeoXFPS=223.82026.060.08493.1
LoGeR2026.050.0993
Point3RParam.=600M2026.030.09393.5
MonST3RAlign=scale & shift2026.020.09888.3
Lingbot-map2026.050.09890.7
ZipMapStreaming=true, Alignment=scale-only2026.030.190.3
Easi3RAlignment=Per-sequence scale, Online=false, Venue=ICCV’252026.030.10291.2
CUT3RCategory=Vision Foundation Model, Optim.=false, FF=true, Global Alignment (GA)=false, FPS=16.582026.010.10688.7
MonST3RAlign=scale2026.020.10788.4
MonST3RParams=571M, FPS=1.272025.070.10788.4
SSRAlignment=Per-sequence scale, Online=true, Venue=Ours2026.030.10991.1
DVGTFPS=77.802026.060.10989.4
DepthCrafterCategory=Depth Estimation model, Optim.=false, FF=true, Global Alignment (GA)=false, FPS=0.972026.010.1188.1
TTT3RAlignment=Metric scale, Online=true, Venue=ICLR’262026.030.1189.1
CUT3RAlign=scale & shift2026.020.11188.3
TTT3RAlignment=Per-sequence scale, Online=true, Venue=ICLR’262026.030.11390.4
SSRAlignment=Metric scale, Online=true, Venue=Ours2026.030.11389
MASt3RAlign=scale2026.020.11584.8
MASt3RAlign=scale & shift2026.020.11584.9
MASt3RParams=689M, FPS=1.012025.070.11584.8
STREAM3RαAlignment=Per-sequence scale, Type=Stream2025.120.11689.6
OursAlignment=Per-sequence scale, Type=Stream2025.120.11688.3
CUT3RAlignment=Per-sequence scale, Type=Stream2025.120.11888.1
Cut3RParam.=600M2026.030.11888.1
CUT3RAlignment=Per-sequence scale, Online=true, Venue=CVPR’252026.030.11888.1
CUT3R2026.030.11888.1
CUT3RType=Streaming2025.070.11888.1
CUT3R2026.050.11888.1
TTT3RParams=793M, Inference Speed Complexity=O(N), Alignment Protocol=joint scale-and-shift2026.030.1287
LongStream2026.050.1287
CUT3RAlign=scale2026.020.12287.6
CUT3RAlignment=Metric scale, Type=Stream2025.120.12285.5
CUT3RParams=793M, FPS=6.982025.070.12287.6
CUT3RAlignment=Metric scale, Online=true, Venue=CVPR’252026.030.12285.5
CUT3RParams=793M, Inference Speed Complexity=O(N), Alignment Protocol=joint scale-and-shift2026.030.12485
CUT3Rtype=Streaming2026.030.12487.3
OursAlignment=Metric scale, Type=Stream2025.120.12585.8
OVGGTSequence Length=1002026.030.12883.9
OVGGTSequence Length=3002026.030.13384.4
DUSt3RAlign=scale & shift2026.020.13581.8
DUSt3R-GACategory=Vision Foundation Model, Optim.=true, FF=false, Global Alignment (GA)=true, FPS=0.762026.010.13581.8
OVGGTSequence Length=5002026.030.13583.9
Point3RAlignment=Per-sequence scale, Type=Stream2025.120.13684.2
OmniStreamParam.=400M2026.030.13681.8
Point3RAlignment=Per-sequence scale, Online=true, Venue=NeurIPS’252026.030.13684.2
Point3RType=Streaming2025.070.13684.2
Point3R2026.050.13684.2
MASt3R-GACategory=Vision Foundation Model, Optim.=true, FF=false, Global Alignment (GA)=true, FPS=0.312026.010.13783.6
Fast3RAlign=scale2026.020.13883.4
Fast3RParams=648M, FPS=65.82025.070.13883.4
Fast3RAlign=scale & shift2026.020.13980.8
Fast3RParams=648M, Inference Speed Complexity=O(N^2), Alignment Protocol=joint scale-and-shift2026.030.13980.8
Depth-Anything-V2Category=Depth Estimation model, Optim.=false, FF=true, Global Alignment (GA)=false, FPS=3.132026.010.1480.4
Fast3RAlignment=Per-sequence scale, Type=FA2025.120.1483.4
DUSt3RAlign=scale2026.020.14381.4
DUSt3RParams=571M, FPS=1.252025.070.14381.4
DUSt3R-GAAlignment=Per-sequence scale, Type=Optim2025.120.14481.3
DUSt3RAlignment=Per-sequence scale, Online=false, Venue=CVPR’242026.030.14481.3
DUSt3R-GAType=Pair-wise2025.070.14481.3
DUSt3R-GA2026.050.14481.3
TTT3RStreaming=true, Alignment=scale-only2026.030.14981.2