Research map / Image generation / Editing and style transfer
StyleNet: Generating Attractive Visual Captions With Styles
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 the romantic style-specific factor matrices in place of the standard multi-task LSTM | FlickrStyle10K image captioning · romantic references | BLEU-4: 9.5 -> 10.4 | Helped |
| Use the romantic style-specific factor matrices in place of the standard multi-task LSTM | FlickrStyle10K image captioning · romantic references | CIDEr: 0.29 -> 0.31 | Helped |
| Use the romantic style-specific factor matrices in place of the standard multi-task LSTM | FlickrStyle10K image captioning · romantic references | METEOR: 0.145 -> 0.154 | Helped |
| Use the humorous style-specific factor matrices in place of the standard multi-task LSTM | FlickrStyle10K image captioning · humorous references | BLEU-4: 8.8 -> 10.1 | Helped |
| Use the humorous style-specific factor matrices in place of the standard multi-task LSTM | FlickrStyle10K image captioning · humorous references | CIDEr: 0.25 -> 0.27 | Helped |
| Use the humorous style-specific factor matrices in place of the standard multi-task LSTM | FlickrStyle10K image captioning · humorous references | METEOR: 0.148 -> 0.152 | Helped |
| Generate captions with the romantic or humorous style using StyleNet rather than factual-caption baselines | FlickrStyle10K image captioning · image caption attractiveness | human preference: 45.2%; 40.6%; 7.8%; 6.4% | Helped |
| Generate captions with romantic or humorous style using StyleNet rather than the video-caption baseline | Youtube2Text video captioning · video caption attractiveness | human preference: 43.7%; 39.1%; 17.2% | Helped |
| Use factored LSTM rather than traditional LSTM in multi-task training | FlickrStyle10K image captioning · romantic and humorous references | automatic caption metrics | Helped |
| Train romantic and humorous style factors together rather than separately (ablation) | StyleNet · FlickrStyle10K image captioning | caption performance | No clear effect |
Closest papers by meaning
- CAPability: A Comprehensive Visual Caption Benchmark for Evaluating Both Correctness and Thoroughness (NeurIPS 2025 · similarity 0.86)
- SCSA: A Plug-and-Play Semantic Continuous-Sparse Attention for Arbitrary Semantic Style Transfer (CVPR 2025 · similarity 0.85)
- BACON: Improving Clarity of Image Captions via Bag-of-Concept Graphs (CVPR 2025 · similarity 0.85)
- FineStyle: Fine-grained Controllable Style Personalization for Text-to-image Models (NeurIPS 2024 · similarity 0.85)
- Revisit Large-Scale Image-Caption Data in Pre-training Multimodal Foundation Models (ICLR 2025 · similarity 0.85)
- Harmonizing Visual Text Comprehension and Generation (NeurIPS 2024 · similarity 0.85)
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.