← Text-to-Image

Safety, Evaluation & Applications

Evaluation frameworks, safety and bias analysis, robustness research, and downstream applications of text-to-image models.

⌘K

Evaluation, Safety, Bias & Robustness 25+ papers

TitleFocusVenueYear
Moonworks Lunara Aesthetic DatasetAesthetic Dataset arXiv2026
Moonworks Lunara Aesthetic IIImage Variation Dataset arXiv2026
YOLO-CountDifferentiable Object Counting for T2I Generation arXiv2025
Rich Human Feedback for T2I Generation (Best Paper)Human Feedback CVPR2024
PopAlignPopulation-Level Alignment for Fair Text-to-Image Generation arXiv2024
Fine-Grained FeedbackUntangling Challenges of Fine-Grained Feedback for T2I arXiv2024
OpenBiasOpen-set Bias Detection in Text-to-Image Generative Models CVPR2024
SafeGenMitigating Unsafe Content Generation in Text-to-Image Models arXiv2024
DIAGNOSISDetecting Unauthorized Data Usages in T2I Diffusion Models ICLR2024
Spatial ConsistencyGetting it Right: Improving Spatial Consistency in T2I Models arXiv2024
Learning Multi-dim Human PreferenceMulti-dimensional Human Preference for T2I CVPR2024
HEIMHolistic Evaluation of Text-To-Image Models NeurIPS2023
GenEvalAn Object-Focused Framework for Evaluating Text-to-Image Alignment arXiv2023
HPSv2Human Preference Score v2: A Solid Benchmark for Evaluating Human Preferences arXiv2023
ImageRewardLearning and Evaluating Human Preferences for Text-to-Image Generation arXiv2023
TIFAAccurate and Interpretable Text-to-Image Faithfulness Evaluation with QA arXiv2023
LLMScoreUnveiling the Power of LLMs in Text-to-Image Synthesis Evaluation arXiv2023
ConceptBedEvaluating Concept Learning Abilities of Text-to-Image Diffusion Models arXiv2023
IMMAImmunizing T2I Models against Malicious Adaptation arXiv2023
Rickrolling the ArtistInjecting Backdoors into Text Encoders for T2I Synthesis ICCV2023
RIATIGReliable and Imperceptible Adversarial Text-to-Image Generation with Natural Prompts CVPR2023
Demographic StereotypesEasily Accessible T2I Generation Amplifies Demographic Stereotypes at Large Scale FAACT2023
DE-FAKEDetection and Attribution of Fake Images Generated by T2I Diffusion Models arXiv2022
Cultural BiasExploiting Cultural Biases via Homoglyphs in Text-Guided Image Generation arXiv2022
Privacy AnalysisMembership Inference Attacks Against Text-to-image Generation Models arXiv2022

Applications & Downstream Tasks 15+ papers

ModelFull TitleDomainVenueYear
Acquire & AdaptSqueezing out T2I Model for Image Restoration RestorationCVPR2025
JarvisArtLiberating Human Artistic Creativity via an Intelligent Photo Retouching Agent RetouchingarXiv2024
TextDiffuserDiffusion Models as Text Painters Text RenderingarXiv2023
GlyphDrawLearning to Draw Chinese Characters in Image Synthesis Models Coherently CJK TextarXiv2023
SegGenSupercharging Segmentation Models with Text2Mask and Mask2Img Synthesis SegmentationarXiv2023
ODISEOpen-Vocabulary Panoptic Segmentation with Text-to-Image Diffusion Models SegmentationCVPR2023
Image Super-ResolutionImage Super-Resolution with Text Prompt Diffusion Super-ResolutionarXiv2023
HAARText-Conditioned Generative Model of 3D Strand-based Human Hairstyles 3D HairarXiv2023
DiffUTEUniversal Text Editing Diffusion Model Text EditingarXiv2023
Guiding T2I Towards Grounded GenerationGuiding Text-to-Image Diffusion Model Towards Grounded Generation GroundingarXiv2023
CLIP SegmenterCLIP is Also an Efficient Segmenter: A Text-Driven Approach for Weakly Supervised Semantic Segmentation SegmentationarXiv2022
PeekabooText to Image Diffusion Models are Zero-Shot Segmentors SegmentationarXiv2022
AvatarCLIPZero-Shot Text-Driven Generation and Animation of 3D Avatars 3D AvatarsSIGGRAPH2022
Text2LightZero-Shot Text-Driven HDR Panorama Generation HDR PanoramaSIGGRAPH Asia2022
DALL-E for DetectionLanguage-driven Context Image Synthesis for Object Detection DetectionarXiv2022