AI Paper Atlas

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

ICLR 2026 · Image generation › Text-to-image diffusion · approach: Compositional text-to-image guidance · 18 citations (OpenAlex)

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 changedWhereMeasuredEffect
use all previous-round frames via GRCA instead of a 10-frame sliding window (ablation)Story-Iter · Long story benchmarkaCCS: 0.774 -> 0.802Improved
use full reference set instead of a random reference set (ablation)Story-Iter · Long story benchmarkaFID: 119.52 -> 94.30Improved
keep per-frame global tokens instead of mean-pooling global embeddings (ablation)Story-Iter · Long story benchmarkaCCS: 0.741 -> 0.802Improved
add external iterations (initialization to 10th) (ablation)Story-Iter · Long story benchmarkaCCS: 0.502 -> 0.840Improved
add external iterations (initialization to 10th) (ablation)Story-Iter · Long story benchmarkCLIP-T: 0.330 -> 0.306Hurt
increase iterations from 10 to 15 (ablation)Story-Iter · Long story benchmarkaCCS: 0.840 -> 0.848No clear effect
remove initialization (use character images as references) (ablation)Story-Iter · Long story benchmarkCLIP-TImproved
linear weighting (0.3 to 0.5) instead of fixed weight factor (ablation)Story-Iter · Long story benchmarkCLIP-T / aCCSImproved
use 4-step SDXL-LCM sampling instead of 50 stepsStory-IterXL · Long story benchmark, 100 framestime: 250 min -> 20 min; aFID 94.30 -> 109.13Mixed
replace CSA reference with GRCA global embeddingsStoryDiffusion · reference scaling, 100 framesVRAM: 40GB -> 19GBImproved
GRCA vs CSA in first iteration (ablation)Story-Iter · first iterationaCCS / aFIDHurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy