Research map / Label-efficient and robust learning / Recognition and long-tail
Img2Tab: Automatic Class Relevant Concept Discovery from StyleGAN Features for Explainable Image Classification
Young-Jae Song, Sung Kuk Shyn, Kwangsu Kim
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 |
|---|---|---|---|
| mask Makeup concept at feature level (ablation) | Img2Tab · CelebA Gender | top-1 accuracy: 97.80 -> 97.90 | No clear effect |
| mask Hair Length concept (ablation) | Img2Tab · CelebA Gender | top-1 accuracy: 97.80 -> 97.84 | No clear effect |
| mask Smiling concept (ablation) | Img2Tab · CelebA Gender | top-1 accuracy: 97.80 -> 97.83 | No clear effect |
| mask Facial Hair concept (ablation) | Img2Tab · CelebA Gender | top-1 accuracy: 97.80 -> 97.95 | No clear effect |
| choose XGBoost as classifier on StyleGAN features rather than other tabular models (ablation) | Img2Tab · class-relevancy of top-5 features | Wasserstein-1 W_k | Improved |
| use EfficientNet-b4 instead of Img2Tab | Img2Tab · CelebA Gender | top-1 accuracy: 97.85 -> 98.10 | Improved |
| use Img2Tab instead of ResNet-50 | ResNet-50 · CelebA Gender | top-1 accuracy: 97.04 -> 97.85 | Improved |
| use Img2Tab instead of ResNet-50 | ResNet-50 · CelebA Bangs | top-1 accuracy: 95.39 -> 94.88 | Hurt |
| use GH-Feat instead of Img2Tab | Img2Tab · MNIST | top-1 accuracy: 98.24 -> 99.06 | Improved |
| use Img2Tab instead of ALAE | ALAE · MNIST | top-1 accuracy: 97.61 -> 98.24 | Improved |
Closest papers by meaning
- DEPICT: Diffusion-Enabled Permutation Importance for Image Classification Tasks (ECCV 2024 · similarity 0.86)
- Explaining in Diffusion: Explaining a Classifier with Diffusion Semantics (CVPR 2025 · similarity 0.85)
- CE-FAM: Concept-Based Explanation via Fusion of Activation Maps (ICCV 2025 · similarity 0.85)
- Explanation-based Data Augmentation for Image Classification (NeurIPS 2021 · similarity 0.85)
- ConceptBed: Evaluating Concept Learning Abilities of Text-to-Image Diffusion Models (AAAI 2024 · similarity 0.84)
- Finer-CAM: Spotting the Difference Reveals Finer Details for Visual Explanation (CVPR 2025 · 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.
