Research map / Image generation / Text-to-image diffusion
Unsupervised Stylish Image Description Generation via Domain Layer Norm
Cheng-Kuan Chen, Zhufeng Pan, Ming-Yu Liu, Min Sun
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 |
|---|---|---|---|
| use Domain Layer Norm rather than NST | stylish image description generation · Lyrics | content similarity: 0.037 -> 0.083 | Improved |
| use Domain Layer Norm rather than NST | stylish image description generation · Romance | content similarity: 0.088 -> 0.151 | Improved |
| use Domain Layer Norm rather than NST | stylish image description generation · Humor | content similarity: 0.103 -> 0.173 | Improved |
| use Domain Layer Norm rather than NST | stylish image description generation · Fairy tale | content similarity: 0.116 -> 0.135 | Improved |
| use Domain Layer Norm rather than NST | stylish image description generation · Lyrics | transfer accuracy: 100% -> 99.2% | Hurt |
| use Domain Layer Norm rather than NST | stylish image description generation · Humor | transfer accuracy: 99.7% -> 70.0% | Hurt |
| replace DLN's LN-LSTM with a simple RNN (ablation) | DLN · VG-Para test | BLEU-4: 0.080 -> 0.062 | Hurt |
| progressively add styles to the model (DLN-Multi) | DLN · Lyrics | content similarity: 0.083 -> 0.118 | Improved |
| progressively add styles to the model (DLN-Multi) | DLN · Romance | content similarity: 0.151 -> 0.116 | Hurt |
| progressively add styles to the model (DLN-Multi) | DLN · Fairy tale | content similarity: 0.135 -> 0.120 | Hurt |
Closest papers by meaning
- StyleNet: Generating Attractive Visual Captions With Styles (CVPR 2017 · similarity 0.87)
- SemStyle: Learning to Generate Stylised Image Captions Using Unaligned Text (CVPR 2018 · similarity 0.87)
- StyleDrop: Text-to-Image Synthesis of Any Style (NeurIPS 2023 · similarity 0.85)
- FineStyle: Fine-grained Controllable Style Personalization for Text-to-image Models (NeurIPS 2024 · similarity 0.84)
- DEADiff: An Efficient Stylization Diffusion Model with Disentangled Representations (CVPR 2024 · similarity 0.84)
- Conditional Balance: Improving Multi-Conditioning Trade-Offs in Image Generation (CVPR 2025 · similarity 0.84)
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.
