List your company
ResearchFormerly WizWand

Image Retrieval on Oxford105k (test)

95.2mAPProposed diffusion w/ late fusion
32.90449.07765.2581.423Nov 18, 2015May 20, 2016Nov 20, 2016May 23, 2017Nov 23, 2017May 26, 2018Nov 27, 2018
Updated 8mo ago

Evaluation results

MethodLinks
Proposed diffusion w/ late fusionFeature=R-MAC (ResNet), Global=true, Regional=true2018.1195.2
Proposed diffusionFeature=R-MAC (ResNet), Global=false, Regional=true2018.1194.8
Iscen's diffusionFeature=R-MAC (ResNet), Global=true, Regional=true2018.1194.2
Proposed diffusionFeature=R-MAC (ResNet), Global=true, Regional=false2018.1191.8
Proposed diffusion w/ late fusionFeature=R-MAC (VGG), Global=true, Regional=true2018.1191.2
Iscen's diffusionFeature=R-MAC (VGG), Global=true, Regional=true2018.1190.3
GeM+alphaQENet=VGG, F-tuned=yes, Dim=5122017.1189.6
R-match + AQEFeature=R-MAC (ResNet), Global=false, Regional=true2018.1189.6
GeM+alphaQENet=Res, F-tuned=yes, Dim=20482017.1189.5
R-MAC+QENet=Res, F-tuned=yes, Dim=20482017.1189.4
Proposed diffusionFeature=R-MAC (VGG), Global=false, Regional=true2018.1188.6
k-NN + AQEFeature=R-MAC (ResNet), Global=true, Regional=false2018.1188.3
Iscen's diffusionFeature=R-MAC (ResNet), Global=true, Regional=false2018.1187.4
R-MAC+QENet=VGG, F-tuned=yes, Dim=5122017.1187.3
Proposed diffusionFeature=R-MAC (VGG), Global=true, Regional=false2018.1186.8
R-matchFeature=R-MAC (ResNet), Global=false, Regional=true2018.1185.7
GeMNet=Res, F-tuned=yes, Dim=20482017.1184.6
HQENet=n/a, F-tuned=n/a, Dim=n/a2017.1184
GeMNet=VGG, F-tuned=yes, Dim=5122017.1183.3
R-MACNet=Res, F-tuned=yes, Dim=20482017.1182.8
Iscen's diffusionFeature=R-MAC (VGG), Global=true, Regional=false2018.1182.7
k-NN searchFeature=R-MAC (ResNet), Global=true, Regional=false2018.1180.8
k-NN + AQEFeature=R-MAC (VGG), Global=true, Regional=false2018.1179.7
Mikulik et al. (2013)2015.1179.5
BoW-fVocab+R+QENet=n/a, F-tuned=n/a, Dim=n/a2017.1179.5
R-MACNet=VGG, F-tuned=yes, Dim=5122017.1178.6
R-match + AQEFeature=R-MAC (VGG), Global=false, Regional=true2018.1178.6
Chum et al. (2011)2015.1176.7
Danfeng et al. (2011)2015.1176.7
BoW+R+QENet=n/a, F-tuned=n/a, Dim=n/a2017.1176.7
Fisher VectorNet=VGG, F-tuned=yes, Dim=5122017.1176.6
R-matchFeature=R-MAC (VGG), Global=false, Regional=true2018.1176.5
R-MAC+QENet=Res, F-tuned=no, Dim=20482017.1175.5
MACDimensionality=512, Backbone=VGG, Training Protocol=Fine-tuned, Query Cropping=Crop X2016.0475.1
Tolias et al. (2015)2015.1175
R-MAC+AML+QE2015.1173.2
R-MAC+R+QENet=VGG, F-tuned=no, Dim=5122017.1173.2
Shen et al. (2014)2015.1172.9
k-NN searchFeature=R-MAC (VGG), Global=true, Regional=false2018.1172.1
CroW+QENet=VGG, F-tuned=no, Dim=5122017.1170.6
R-MACDimensionality=128, Backbone=VGG, Training Protocol=Fine-tuned, Query Cropping=Crop X2016.0469.7
MACDimensionality=128, Backbone=VGG, Training Protocol=Fine-tuned, Query Cropping=Crop2016.0468.6
CroWNet=VGG, F-tuned=no, Dim=5122017.1165.3
BOW-CNN+R+QENet=VGG, F-tuned=no, Dim=n/a2017.1165.1
R-MACNet=Res, F-tuned=no, Dim=20482017.1163.7
R-MACD=5122015.1161.6
R-MACNet=VGG, F-tuned=no, Dim=5122017.1161.6
SPOCNet=VGG, F-tuned=no, Dim=5122017.1161.1
BoW-CNNNet=VGG, F-tuned=no, Dim=n/a2017.1159.3
Babenko et al. (2014)D=1282015.1152.3
Jégou & Zisserman (2014)D=10242015.1150.2
Babenko & Lempitsky (2015)D=2562015.1150.1
Razavian et al. (2014b)D=2562015.1148.9
MACNet=VGG, F-tuned=no, Dim=5122017.1147.8
R-MACD=2562015.1147
Jégou & Zisserman (2014)D=1282015.1135.3