Research map / Vision-language and multimodal / Human and embodied modeling
Let Your Photos Talk: Generating Narrative Paragraph for Photo Stream via Bidirectional Attention Recurrent Neural Networks
Read the paperCheck 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 |
|---|---|---|---|
| add attention and sGRU (vs BARNN-sGRU) (ablation) | BARNN · SIND retrieval | Recall@10: 46.96 -> 53.06 | Helped |
| add semantic embedding (vs BARNN-EMB) (ablation) | BARNN · SIND retrieval | Recall@10: 47.01 -> 53.06 | Helped |
| nonlinear sGRU instead of linear CLSTM (ablation) | BARNN · SIND retrieval | Recall@10: 47.39 -> 53.06 | Helped |
| add attention and sGRU (vs BARNN-sGRU) (ablation) | BARNN · NYC retrieval | Recall@10: 39.53 -> 52.10 | Helped |
| add semantic embedding (vs BARNN-EMB) (ablation) | BARNN · Disney retrieval | Recall@10: 46.50 -> 57.83 | Helped |
| BARNN visual stream model instead of CRCN | BARNN · NYC retrieval | Recall@10: 43.57 -> 52.10 | Helped |
| BARNN-sGRU/EMB variants vs CRCN | BARNN variants · NYC retrieval | Recall@10: 41.07 vs 43.57 (BCLSTM vs CRCN) | Mixed |
| BARNN instead of Beam=10 seq2seq baseline | BARNN · SIND generation | METEOR: 23.13 -> 33.32 | Helped |
| BARNN instead of Grounded baseline | BARNN · SIND generation | METEOR: 31.42 -> 33.32 | Helped |
| BARNN instead of CRCN | BARNN · NYC generation | CIDEr: 30.9 -> 41.6 | Helped |
| BARNN instead of CRCN | BARNN · Disney generation | CIDEr: 51.3 -> 54.1 | Helped |
| BARNN instead of CRCN in user study | BARNN · SIND user study | preference: 13.5% vs 68.3% | Helped |
Closest papers by meaning
- Show, Reward and Tell: Automatic Generation of Narrative Paragraph From Photo Stream by Adversarial Training (AAAI 2018 · similarity 0.92)
- A Hierarchical Approach for Generating Descriptive Image Paragraphs (CVPR 2017 · similarity 0.91)
- Hide-and-Tell: Learning to Bridge Photo Streams for Visual Storytelling (AAAI 2020 · similarity 0.89)
- Tell as You Want: Customizing Image Narrative with Knowledge and Thoughts (AAAI 2026 · similarity 0.88)
- Intelligent Grimm - Open-ended Visual Storytelling via Latent Diffusion Models (CVPR 2024 · similarity 0.86)
- One-Prompt-One-Story: Free-Lunch Consistent Text-to-Image Generation Using a Single Prompt (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.