Research map / Image generation / Text-to-image diffusion
GWScore: A structural diversity metric for consistent text-to-image generation
Francis Snelgar, Stephen Gould, Liang Zheng, Akshay Asthana
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes GWScore, a metric for measuring structural diversity (pose and spatial layout variation) in sequences of consistent text-to-image generations.
Problem
Consistent text-to-image storytelling methods preserve identity across image sequences but often produce images with limited structural diversity in subject pose and spatial layout, and existing evaluation protocols mainly measure consistency rather than this diversity.
Approach
A metric named GWScore that scores structural diversity across generated image sequences, targeting pose and spatial layout rather than identity consistency alone. Details beyond the abstract are not available.
Experimental setup
Metrics: GWScore
Closest papers by meaning
- Storynizor: Consistent Story Generation via Inter-Frame Synchronized and Shuffled ID Injection (AAAI 2025 · similarity 0.89)
- Storybooth: Training-Free Multi-Subject Consistency for Improved Visual Storytelling (ICLR 2025 · similarity 0.88)
- Consistent Story Generation: Unlocking the Potential of Zigzag Sampling (NeurIPS 2025 · similarity 0.88)
- VinaBench: Benchmark for Faithful and Consistent Visual Narratives (CVPR 2025 · similarity 0.88)
- One-Prompt-One-Story: Free-Lunch Consistent Text-to-Image Generation Using a Single Prompt (ICLR 2025 · similarity 0.88)
- SceneDecorator: Towards Scene-Oriented Story Generation with Scene Planning and Scene Consistency (NeurIPS 2025 · similarity 0.88)
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.
