Research map / Image generation / Text-to-image diffusion
Leveraging Panoptic Scene Graph for Evaluating Fine-Grained Text-to-Image Generation
Xueqing Deng, Linjie Yang, Qihang Yu, Chenglin Yang, Liang-Chieh Chen
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 PSG-Score instead of GenEval for ranking models on complex prompts | T2I evaluation · PSG-Bench | ranking vs human Likert | Improved |
| use PSG-Score instead of GenEval as prompt difficulty rises to Hard | T2I evaluation · PSG-Bench (Hard prompts) | Spearman rho with human: 0.16 -> 0.54 | Improved |
| use PSG-Score instead of TIFA on Easy prompts | T2I evaluation · PSG-Bench (Easy prompts) | Spearman rho with human: 0.52 -> 0.64 | Improved |
| use PSG-Score instead of TIFA on Hard prompts | T2I evaluation · PSG-Bench (Hard prompts) | Spearman rho with human: 0.24 -> 0.54 | Improved |
| use simple GenEval prompts instead of PSG-Bench prompts | T2I evaluation · GenEval prompts | score spread across models: 0.57 to 0.68 | Hurt |
| use yes/no QA (TIFA) for evaluation | TIFA · PSG-Bench | human agreement (%): 64 | Hurt |
| use yes/no QA (DSG) for evaluation | DSG · PSG-Bench | human agreement (%): 67 | Hurt |
| penalize extra foreground objects via saliency detector | PSG-Score · PSG-Bench | FP counting | Improved |
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.
- GenEval — GenEval: An object-focused framework for evaluating text-to-image alignment (NeurIPS 2023) · 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
- OneIG-Bench: Omni-dimensional Nuanced Evaluation for Image Generation (NeurIPS 2025 · similarity 0.89)
- Revisiting text-to-image evaluation with Gecko: on metrics, prompts, and human rating (ICLR 2025 · similarity 0.87)
- Interleaved Scene Graphs for Interleaved Text-and-Image Generation Assessment (ICLR 2025 · similarity 0.87)
- DynT2I-Eval: A Dynamic Evaluation Framework for Text-to-Image Models (NeurIPS 2026 · similarity 0.87)
- Panoptic Scene Graph Generation with Semantics-Prototype Learning (AAAI 2024 · similarity 0.87)
- ConceptMix: A Compositional Image Generation Benchmark with Controllable Difficulty (NeurIPS 2024 · 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.
