Research map / Label-efficient and robust learning / Recognition and long-tail
Learning Deep Bilinear Transformation for Fine-grained Image Representation
Heliang Zheng, Jianlong Fu, Zheng-Jun Zha, Jiebo Luo
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add DBT blocks to ResNet-50 | ResNet-50 · Stanford-Car | accuracy: 92.8 -> 94.1 | Improved |
| Add DBT blocks to ResNet-50 | ResNet-50 · FGVC-Aircraft | accuracy: 90.0 -> 91.2 | Improved |
| Add DBT blocks to ResNet-50 | ResNet-50 · CUB-200-2011 | accuracy: 88.1 -> 87.5 | Hurt |
| Add DBT blocks to ResNet-50 | ResNet-50 · iNaturalist-2017 | accuracy: 59.9% -> 62.0% | Improved |
| Add semantic grouping constraint during pre-training (ablation) | DBTNet-50 · CUB-200-2011, 224 × 224 | accuracy: 79.8 -> 84.6 | Improved |
| Increase pre-training semantic grouping weight from 3e-4 to 3e-3 (ablation) | DBTNet-50 · CUB-200-2011, 224 × 224 | accuracy: 84.6 -> 83.1 | Hurt |
| Use fine-tuning semantic grouping weight 1e-5 along with pre-training weight 3e-4 (ablation) | DBTNet-50 · CUB-200-2011, 224 × 224 | accuracy: 84.6 -> 85.1 | Improved |
| Add group index encoding with t=1.5 (ablation) | DBTNet-50 · CUB-200-2011, 224 × 224 | accuracy: 84.5 -> 85.1 | Improved |
| Use group index encoding with t=10 instead of t=1.5 (ablation) | DBTNet-50 · CUB-200-2011, 224 × 224 | accuracy: 85.1 -> 84.5 | Hurt |
| Add shortcut connection in Stage V (ablation) | DBTNet-50 · CUB-200-2011, 224 × 224, Stage V | accuracy: 84.6 -> 84.8 | Improved |
| Add shortcut connection in Stages V+IV (ablation) | DBTNet-50 · CUB-200-2011, 224 × 224, Stages V+IV | accuracy: 83.1 -> 85.1 | Improved |
| Add DBT in Stages V+IV in addition to the last layer (ablation) | DBTNet-50 · CUB-200-2011, 224 × 224 | accuracy: 84.4 -> 85.1 | Improved |
Closest papers by meaning
- Factorized Bilinear Models for Image Recognition (ICCV 2017 · similarity 0.86)
- Hierarchical Bilinear Pooling for Fine-Grained Visual Recognition (ECCV 2018 · similarity 0.86)
- Bilinear Convolutional Neural Networks for Fine-Grained Visual Recognition (TPAMI 2017 · similarity 0.85)
- AVT: Unsupervised Learning of Transformation Equivariant Representations by Autoencoding Variational Transformations (ICCV 2019 · similarity 0.84)
- Bi-directional Feature Reconstruction Network for Fine-Grained Few-Shot Image Classification (AAAI 2023 · similarity 0.84)
- Multilinear Operator Networks (ICLR 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
