Research map / Image generation / Text-to-image diffusion
HiGFA: Hierarchical Guidance for Fine-grained Data Augmentation with Diffusion Models
Zhiguang Lu, Qianqian Xu, Peisong Wen, Siran Dai, Qingming Huang
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 contour guidance to text guidance (ablation) | HiGFA · Aircraft | top-1 accuracy (%): 83.8 -> 85.3 | Improved |
| add contour guidance to text guidance (ablation) | HiGFA · DTD | top-1 accuracy (%): 63.5 -> 66.6 | Improved |
| add fine-grained classifier guidance (ablation) | HiGFA · Aircraft | top-1 accuracy (%): 85.3 -> 86.1 | Improved |
| add fine-grained classifier guidance (ablation) | HiGFA · CUB | top-1 accuracy (%): 80.2 -> 82.7 | Improved |
| add fine-grained classifier guidance (ablation) | HiGFA · DTD | top-1 accuracy (%): 66.6 -> 70.8 | Improved |
| use text+contour guidance without classifier guidance vs no augmentation (ablation) | HiGFA · CUB / DTD | top-1 accuracy (%): DTD 63.5 text-only vs 68.5 NoAug | Hurt |
| use dynamic confidence-based guidance scale instead of constant (ablation) | HiGFA · Aircraft | top-1 accuracy (%) | Improved |
| start classifier guidance in later sampling stages (ablation) | HiGFA · Aircraft | top-1 accuracy (%) | Improved |
| use very large classifier guidance scale (ablation) | HiGFA · Aircraft | top-1 accuracy (%) | Hurt |
| combine HiGFA with FRAug (ablation) | HiGFA · Aircraft | top-1 accuracy (%): 83.2 -> 86.1 | Improved |
| combine HiGFA with AutoAug (ablation) | HiGFA · Dogs | top-1 accuracy (%): 84.6 -> 85.3 | Improved |
| HiGFA vs strongest generative baseline SaSPA (ablation) | HiGFA · CUB | top-1 accuracy (%): 81.7 -> 82.7 | Improved |
Closest papers by meaning
- HiFlow: Training-free High-Resolution Image Generation with Flow-Aligned Guidance (NeurIPS 2025 · similarity 0.87)
- Nested Diffusion Models Using Hierarchical Latent Priors (CVPR 2025 · similarity 0.87)
- Advancing Fine-Grained Classification by Structure and Subject Preserving Augmentation (NeurIPS 2024 · similarity 0.87)
- Elucidating the design space of classifier-guided diffusion generation (ICLR 2024 · similarity 0.87)
- DiffuseHigh: Training-Free Progressive High-Resolution Image Synthesis Through Structure Guidance (AAAI 2025 · similarity 0.86)
- Eliminating Oversaturation and Artifacts of High Guidance Scales in Diffusion Models (ICLR 2025 · similarity 0.86)
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.
