Research map / Video / Video generation
DreamShot: Personalized Storyboard Synthesis with Video Diffusion Prior
Junjia Huang, Binbin Yang, Pengxiang Yan, Jiyang Liu, Bin Xia, Zhao Wang, Yitong Wang, Liang Lin, Guanbin Li
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 |
|---|---|---|---|
| add role-attention consistency loss L_RAC (ablation) | DreamShot · DreamShot Test Set Reference-to-Shot | CIDS-Cross: 48.3 -> 51.6 | Improved |
| add L_RAC (ablation) | DreamShot · DreamShot Test Set Reference-to-Shot | Alignment score: 3.26 -> 3.39 | Improved |
| apply a phase offset to reference RoPE (ablation) | DreamShot · Reference-to-Shot | CIDS-Cross: 51.6 -> 20.0 | Hurt |
| apply a negative positional offset to reference (ablation) | DreamShot · Reference-to-Shot | CIDS-Cross: 51.6 -> 40.9 | Hurt |
| place reference image at sequence end instead of start (ablation) | DreamShot · Reference-to-Shot | CIDS-Cross: 51.6 -> 43.5 | Hurt |
| place reference image at sequence end instead of start (ablation) | DreamShot · Reference-to-Shot | AES: 5.50 -> 5.53 | Improved |
| use video-diffusion DreamShot instead of DreamO | DreamShot · Reference-to-Shot CIDS-Cross | CIDS-Cross: 40.0 -> 51.6 | Improved |
| use DreamShot instead of second-best OmniGen2 | DreamShot · VistoryBench | CIDS-Cross: 54.8 -> 60.2 | Improved |
| train on real data only instead of real plus synthetic (ablation) | DreamShot · Reference-to-Shot | CIDS-Cross: 51.6 -> 52.2 | Mixed |
| train on synthetic data only instead of mixed (ablation) | DreamShot · Reference-to-Shot | AES: 5.50 -> 5.57 | Mixed |
| raise guidance omega1 to 8 (ablation) | DreamShot · Reference-free CFG | AES: omega1=1 -> omega1=8 | Hurt |
| raise lambda of L_RAC to 0.5 (ablation) | DreamShot · Training objective | AES: lambda 0.2 -> 0.5 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- StoryDiffusion — StoryDiffusion: Consistent Self-Attention for Long-Range Image and Video Generation (NeurIPS 2024) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Storybooth: Training-Free Multi-Subject Consistency for Improved Visual Storytelling (ICLR 2025 · similarity 0.89)
- MovieDreamer: Hierarchical Generation for Coherent Long Visual Sequences (ICLR 2025 · similarity 0.89)
- SceneDecorator: Towards Scene-Oriented Story Generation with Scene Planning and Scene Consistency (NeurIPS 2025 · similarity 0.89)
- StoryDiffusion: Consistent Self-Attention for Long-Range Image and Video Generation (NeurIPS 2024 · similarity 0.89)
- Storynizor: Consistent Story Generation via Inter-Frame Synchronized and Shuffled ID Injection (AAAI 2025 · similarity 0.87)
- CineOrchestra: Unified Entity-Centric Conditioning for Cinematic Video Generation (NeurIPS 2026 · similarity 0.87)
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.
