Texture Classification on GTOS
87.2AccuracyRADAM
Evaluation results
| Method | Links | |
|---|---|---|
| RADAMBackbone=ConvNeXt-XL, Input size=224x224, Pre-training dataset=ImageNet-21K2023.03 | 87.2 | |
| RADAMBackbone=ConvNeXt-B, Input size=224x224, Pre-training dataset=ImageNet-21K2023.03 | 86.6 | |
| RADAMBackbone=ConvNeXt-L, Input size=224x224, Pre-training dataset=ImageNet-21K2023.03 | 85.9 | |
| FENet2026.05 | 85.7 | |
| CLASSNetBackbone=ResNet50, Input size=224x2242023.03 | 85.6 | |
| CLASSNet2026.05 | 85.6 | |
| RADAMBackbone=ConvNeXt-T, Input size=224x224, Pre-training dataset=ImageNet-21K2023.03 | 85.4 | |
| DSRNetBackbone=ResNet50, Input size=224x2242023.03 | 85.3 | |
| DSRNet2026.05 | 85.3 | |
| MAPNetBackbone=ResNet50, Input size=224x2242023.03 | 84.7 | |
| DeepTENBackbone=ResNet50, Input size=352x3522023.03 | 84.5 | |
| CLASSNetBackbone=ResNet18, Input size=224x2242023.03 | 84.3 | |
| DeepTEN2026.05 | 84.3 | |
| RADAMBackbone=ConvNeXt-T, Input size=224x2242023.03 | 84.2 | |
| RADAM2026.05 | 84.2 | |
| RADAMBackbone=ConvNeXt-B, Input size=224x2242023.03 | 84.1 | |
| RADAMBackbone=ConvNeXt-L, Input size=224x2242023.03 | 84 | |
| RADAMBackbone=ConvNeXt-nano, Input size=224x2242023.03 | 83.7 | |
| RADAMBackbone=ResNet50, Input size=224x2242023.03 | 81.8 | |
| RADAM lightBackbone=MobileNet V2 1.4, Input size=224x2242023.03 | 81.7 | |
| RADAM lightBackbone=MobileNet V2, Input size=224x2242023.03 | 81 | |
| DSRNetBackbone=ResNet18, Input size=224x2242023.03 | 81 | |
| RADAMBackbone=ResNet18, Input size=224x2242023.03 | 80.6 | |
| MAPNetBackbone=ResNet18, Input size=224x2242023.03 | 80.3 | |
| EfficientNet-B5training_mode=fine-tuning2026.05 | 78.7 |