List your company

Image-to-Text Retrieval on MS-COCO 1K (test)

82R@1VSE∞
26.04840.57455.169.626Apr 13, 2023May 26, 2023Jul 8, 2023Aug 20, 2023Oct 2, 2023Nov 14, 2023Dec 27, 2023

Evaluation results

MethodLinks
VSE∞Backbone=ViT-B/322023.098297.298.9—
PCME++Backbone=ViT-B/322023.0981.697.299—
CRCLGlobal-level method=false2023.1080.796.598.6528.2
NAAFBackbone=ResNet-1012023.0980.596.598.8—
RCLGlobal-level method=false2023.1080.496.498.7526.6
PAUBackbone=ViT-B/322023.0980.496.298.5—
CCR&CCSGlobal-level method=false2023.1080.296.898.7526.4
DAABackbone=ResNet-1012023.0980.296.498.8—
PCMEBackbone=ViT-B/322023.0980.196.698.7—
Baseline (CLIP)Backbone=ViT-B/322023.0980.195.798.2—
SGRAFGlobal-level method=false2023.1079.696.298.5524.3
CRCLNoise Rate=20%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1079.696.198.7525.6
SGRAFBackbone=ResNet-1012023.0979.696.298.5—
DECLGlobal-level method=false2023.1079.196.398.7523.1
RCLNoise=20%2023.1278.99698.4521.4
NAAFGlobal-level method=false2023.1078.99698.7524.6
BiCroNoise=20%, additional priors=true2023.1278.896.198.6523.2
BiCroNoise Rate=20%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1078.896.198.6523.2
MV-VSEGlobal-level method=true2023.1078.795.798.7521.9
NCRGlobal-level method=false2023.1078.795.898.5522.5
CMCANGlobal-level method=false2023.1078.696.598.9524.8
SREMNoise=20%2023.1278.596.898.8524.1
VSE∞Global-level method=true2023.1078.59698.7520.8
GSMNGlobal-level method=false2023.1078.496.498.6522.5
BiCroGlobal-level method=false2023.1078.395.898.5521
CRCLNoise Rate=40%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1078.295.798.3521.5
MSCNNoise Rate=20%2023.0478.197.298.8—
MSCNNoise=20%, additional priors=true2023.1278.197.298.8524.6
MSCNNoise Rate=20%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1078.197.298.8524.6
UARDAGlobal-level method=false2023.1077.89597.6500.3
NCRNoise Rate=20%2023.0477.795.598.2—
MSCNNoise Rate=50%2023.0477.596.298.7—
DECLNoise=20%2023.1277.595.998.4518.2
SREMNoise=40%2023.1277.29698.5518.4
MSCNNoise=40%, additional priors=true2023.1277.195.798.4515.7
DECLNoise Rate=20%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1077.195.998.4517.3
DECLNoise Rate=40%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1077.195.798.3517.1
BiCroNoise=40%, additional priors=true2023.127795.998.3517.1
RCLNoise=40%2023.127795.598.3515.3
NCRNoise Rate=20%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.107795.698.1516.6
BiCroNoise Rate=40%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.107795.998.3517.1
CGMNGlobal-level method=false2023.1076.895.498.3520.7
IMRAMGlobal-level method=false2023.1076.795.698.5516.6
NCRNoise=20%2023.1276.695.698.2515
NCRNoise Rate=40%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1076.59598.2513.9
CRCLNoise Rate=60%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1076.395.197.9514.2
VSRNGlobal-level method=true2023.1076.294.898.2516.8
DECLNoise=40%2023.1275.695.598.3512
SGR-CNoise Rate=20%2023.0475.495.297.9—
MSCNNoise Rate=40%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1074.894.998510.9
NCRNoise=40%2023.1274.794.698509.7
NCRNoise Rate=50%2023.0474.694.697.8—
SREMNoise=60%2023.1274.594.597.9506.9
MSCNNoise Rate=70%2023.0474.494.997.7—
MSCNNoise Rate=60%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1074.495.197.9506.5
MSCNNoise=60%, additional priors=true2023.1274.194.497.6503.4
RCLNoise=60%2023.127494.397.5503.3
BiCroNoise=60%, additional priors=true2023.1273.994.497.8505.5
BiCroNoise Rate=60%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1073.994.497.8505.5
DECLNoise Rate=60%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1073.894.797.7508.2
VSE∞Noise Rate=20%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1073.593.397500.5
DECLNoise=60%2023.127394.297.9502.5
SCANGlobal-level method=false2023.1072.794.898.4507.9
NCRNoise Rate=60%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1072.79497.6503.3
CRCLNoise Rate=80%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1072.793.597.6501.8
CAMPGlobal-level method=false2023.1072.394.898.3506.8
SGR-CNoise Rate=50%2023.0471.794.197.7—
SAFNoise=20%2023.1271.59497.5499.1
IMRAMNoise=20%2023.1269.993.697.4490.8
SGR-CNoise Rate=70%2023.0469.893.697.5—
CVSEGlobal-level method=true2023.1069.293.397.5496.4
PVSEBackbone=ResNet-1012023.0969.291.696.6—
IMRAMNoise Rate=20%2023.0468.692.897.6—
SGR*Noise Rate=20%2023.0467.891.796.2—
SAFNoise Rate=20%2023.0467.392.596.6—
MSCNNoise Rate=80%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1066.891.696.2481.2
SCANNoise Rate=20%2023.0466.29196.4—
DECLNoise Rate=80%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1065.691.696.6480.1
SCANNoise=20%2023.1262.29096.1464.5
VSRNNoise=20%2023.1261.887.392.9460.6
SGR*Noise Rate=50%2023.0460.687.493.6—
VSE∞Noise Rate=40%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1053.384.392.1399.9
IMRAMNoise=40%2023.1251.882.490.9412.7
SGRNoise Rate=20%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1049.183.892.7434
SGRNoise Rate=40%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1043.978.389.3406.4
SCANNoise=40%2023.1242.974.685.1343.2
SAFNoise Rate=20%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.104178.489.4406.5
SCANNoise Rate=50%2023.0440.873.584.9—
BiCroNoise Rate=80%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.104072.684.7340.1
SGRNoise Rate=60%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1037.673.386.3381.3
SAFNoise Rate=40%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.103674.487383
SGR*Noise Rate=70%2023.0435.771.285.4—
NCRNoise Rate=70%2023.0435.469.583.4—
VSE∞Noise Rate=60%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1033.464.879.1339.7
SAFNoise Rate=50%2023.0430.467.882.3—
SCANNoise Rate=40%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1030.165.279.2314.4
SCANNoise=60%2023.1229.960.974.8173
VSRNNoise=40%2023.1229.862.176.6292
SCANNoise Rate=20%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1028.964.579.5322.6
SAFNoise Rate=60%, Visual Feature=BUTD, Backbone=Bi-GRU, Evaluation=5-fold average2023.1028.263.979.4348.7