List your company
ResearchFormerly WizWand

Camera Pose Estimation on 7Scenes (AUC & Performance Metrics)

24.86AUC@3VGGT* + DA-VGGT
-0.8675.81212512.4912519.170375Jul 2, 2026Jul 3, 2026
Updated 3mo ago

Evaluation results

MethodLinks
VGGT* + DA-VGGTFrames=5002026.0724.8670.1382.8616.9949.09
VGGT* + DA-VGGTFrames=3002026.0724.8170.282.9212.6827.65
VGGT* + DA-VGGTFrames=10002026.0724.6769.9882.7418.3287.32
VGGT*Frames=3002026.0724.1269.882.6918.7654.16
VGGT*Frames=5002026.0723.5969.4282.4428.41149.68
LiteVGGTFrames=3002026.0722.3468.1681.6516.4315.35
VGGT*Frames=10002026.0722.0168.5381.8369.56584.72
LiteVGGT + DA-VGGTFrames=5002026.0721.1767.3681.0811.9115.97
LiteVGGTFrames=5002026.0720.8966.9180.7125.9433.51
LiteVGGT + DA-VGGTFrames=10002026.0720.8367.0780.8713.9831.68
LiteVGGT + DA-VGGTFrames=3002026.0720.6866.9880.8610.049.84
LiteVGGTFrames=10002026.0718.7761.9876.148.6295.11
FasterVGGT + DA-VGGTFrames=3002026.0717.665.5980.2412.9825.99
FasterVGGT + DA-VGGTFrames=5002026.0717.5865.4680.0915.6542.6
FasterVGGT + DA-VGGTFrames=10002026.0717.565.3479.9818.3183.84
FasterVGGTFrames=3002026.0716.6265.0479.9225.1838.87
FasterVGGTFrames=5002026.0716.2864.6279.640.993.94
SAF3RAttention=Sparse, Base Architecture=DA32026.070.2884—0.868——
DA3Attention=Full2026.070.2859—0.8671——
FastDA3Attention=Sparse2026.070.2853—0.859——
ADA3Attention=Sparse2026.070.2505—0.8515——
Aπ³Attention=Sparse2026.070.2422—0.8607——
π³Attention=Full2026.070.2409—0.8615——
Fastπ³Attention=Sparse2026.070.238—0.853——
VGGTAttention=Full2026.070.2366—0.8464——
SAF3RAttention=Sparse, Base Architecture=VGGT2026.070.2357—0.8449——
SAF3RAttention=Sparse, Base Architecture=π³2026.070.2348—0.8585——
AVGGTAttention=Sparse2026.070.2225—0.8451——
FastVGGTAttention=Sparse2026.070.1867—0.8301——
Co-MeAttention=Sparse2026.070.1619—0.807——
SparseVGGTAttention=Sparse2026.070.1472—0.8117——
Sparseπ³Attention=Sparse2026.070.1225—0.8115——