Research map / Image generation / Text-to-image diffusion
Story-Iter: A Training-free Iterative Paradigm for Long Story Visualization
Jiawei Mao, Xiaoke Huang, Yunfei Xie, Yuanqi Chang, Mude Hui, Bingjie Xu, Zeyu Zheng, Zirui Wang, Cihang Xie, Yuyin Zhou
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 all previous-round frames via GRCA instead of a 10-frame sliding window (ablation) | Story-Iter · Long story benchmark | aCCS: 0.774 -> 0.802 | Improved |
| use full reference set instead of a random reference set (ablation) | Story-Iter · Long story benchmark | aFID: 119.52 -> 94.30 | Improved |
| keep per-frame global tokens instead of mean-pooling global embeddings (ablation) | Story-Iter · Long story benchmark | aCCS: 0.741 -> 0.802 | Improved |
| add external iterations (initialization to 10th) (ablation) | Story-Iter · Long story benchmark | aCCS: 0.502 -> 0.840 | Improved |
| add external iterations (initialization to 10th) (ablation) | Story-Iter · Long story benchmark | CLIP-T: 0.330 -> 0.306 | Hurt |
| increase iterations from 10 to 15 (ablation) | Story-Iter · Long story benchmark | aCCS: 0.840 -> 0.848 | No clear effect |
| remove initialization (use character images as references) (ablation) | Story-Iter · Long story benchmark | CLIP-T | Improved |
| linear weighting (0.3 to 0.5) instead of fixed weight factor (ablation) | Story-Iter · Long story benchmark | CLIP-T / aCCS | Improved |
| use 4-step SDXL-LCM sampling instead of 50 steps | Story-IterXL · Long story benchmark, 100 frames | time: 250 min -> 20 min; aFID 94.30 -> 109.13 | Mixed |
| replace CSA reference with GRCA global embeddings | StoryDiffusion · reference scaling, 100 frames | VRAM: 40GB -> 19GB | Improved |
| GRCA vs CSA in first iteration (ablation) | Story-Iter · first iteration | aCCS / aFID | Hurt |
Closest papers by meaning
- StoryGPT-V: Large Language Models as Consistent Story Visualizers (CVPR 2025 · similarity 0.90)
- Boosting Consistency in Story Visualization with Rich-Contextual Conditional Diffusion Models (AAAI 2025 · similarity 0.88)
- Intelligent Grimm - Open-ended Visual Storytelling via Latent Diffusion Models (CVPR 2024 · similarity 0.88)
- StoryDiffusion: Consistent Self-Attention for Long-Range Image and Video Generation (NeurIPS 2024 · similarity 0.87)
- Storybooth: Training-Free Multi-Subject Consistency for Improved Visual Storytelling (ICLR 2025 · similarity 0.87)
- ViStoryBench: Comprehensive Benchmark Suite for Story Visualization (CVPR 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.
