Classification on Caltech101 (test)
93.79AccuracyZS_En
Evaluation Results
| Method | Links | |
|---|---|---|
| ZS_EnEvaluation Protocol=Zero-shot2023.11 | 93.79 | |
| SPP-net2014.06 | 93.42 | |
| CLIPBackbone=ViT-B/16, Evaluation Protocol=Zero-shot2023.11 | 92.94 | |
| MNPsnoise levels=102023.04 | 92.83 | |
| MGPnoise levels=102023.04 | 92.68 | |
| ReLICv2Backbone=ResNet-50, Pre-training dataset=JFT, Pre-training epochs=1000 ImageNet-equivalent, Evaluation protocol=Linear transfer2022.01 | 92.2 | |
| ReLICv2Backbone=ResNet-50, Pre-training dataset=JFT, Pre-training epochs=5000 ImageNet-equivalent, Evaluation protocol=Linear transfer2022.01 | 91.4 | |
| CLIPBackbone=ViT-B/32, Evaluation Protocol=Zero-shot2023.11 | 91.12 | |
| Proposed FrameworkBackbone=ResNet-502018.05 | 91 | |
| DnCBackbone=ResNet-50, Pre-training dataset=JFT, Pre-training epochs=4500 ImageNet-equivalent, Evaluation protocol=Linear transfer2022.01 | 90.2 | |
| BYOLBackbone=ResNet-50, Pre-training dataset=JFT, Pre-training epochs=5000 ImageNet-equivalent, Evaluation protocol=Linear transfer2022.01 | 90.1 | |
| CLIPBackbone=RN101, Evaluation Protocol=Zero-shot2023.11 | 89.66 | |
| Chatfield et al.2014.06 | 88.54 | |
| DeCAF2014.06 | 86.91 | |
| Zeiler & Fergus2014.06 | 86.5 | |
| Zeiler-Fergus2018.05 | 86.5 | |
| CLIPBackbone=RN50, Evaluation Protocol=Zero-shot2023.11 | 85.96 | |
| SVMClassifier=SVM, Backbone=seven layers fixed model, Pre-training=ImageNet2018.05 | 85.5 | |
| SoftmaxClassifier=Softmax, Backbone=seven layers fixed model, Pre-training=ImageNet2018.05 | 85.4 | |
| FL+EN2018.05 | 83.2 | |
| DE (LF)noise levels=102023.04 | 81.85 | |
| SHDL2018.05 | 81.5 | |
| Proposed FrameworkBackbone=VGG-162018.05 | 80.8 | |
| FKimplementation=as in Table 6 (a)2014.06 | 77.78 | |
| LLCimplementation=as in Table 6 (a)2014.06 | 76.95 | |
| Shaban2018.05 | 75.1 | |
| VQreported_by=[27]2014.06 | 74.41 | |
| DE (EF)noise levels=102023.04 | 73.99 | |
| MCDnoise levels=102023.04 | 73.45 | |
| TMCnoise levels=102023.04 | 71.77 | |
| Image Codes2018.05 | 71.4 | |
| ETPnoise levels=102023.04 | 66.62 | |
| SNGPnoise levels=102023.04 | 56.57 |