NeurIPS 2025 accepted papers on Image generation
440 of the 5,812 papers accepted at NeurIPS 2025 (7.6%) are on Image generation, compared with 7.3% at NeurIPS 2024. They fall into 7 topics.
See NeurIPS 2025 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Fast diffusion sampling · 168 papers
- Mean Flows for One-step Generative Modeling (26 citations)
- Derivative-Free Guidance in Continuous and Discrete Diffusion Models with Soft Value-based Decoding (6 citations)
- Representation Entanglement for Generation: Training Diffusion Transformers Is Much Easier Than You Think (5 citations)
- Training-Free Guidance Beyond Differentiability: Scalable Path Steering with Tree Search in Diffusion and Flow Models (4 citations)
- Constrained Discrete Diffusion (3 citations)
- Inference-Time Scaling for Flow Models via Stochastic Generation and Rollover Budget Forcing (3 citations)
- Noise Matters: Optimizing Matching Noise for Diffusion Classifiers (3 citations)
- Normalized Attention Guidance: Universal Negative Guidance for Diffusion Models (3 citations)
- Test-Time Scaling of Diffusion Models via Noise Trajectory Search (3 citations)
- Adaptive Inference-Time Scaling via Cyclic Diffusion Search (2 citations)
- Blockwise Flow Matching: Improving Flow Matching Models For Efficient High-Quality Generation (2 citations)
- Diffusion Tree Sampling: Scalable inference‑time alignment of diffusion models (2 citations)
and 156 more in the interactive map.
Autoregressive and latent generation · 102 papers
- VisualQuality-R1: Reasoning-Induced Image Quality Assessment via Reinforcement Learning to Rank (8 citations)
- Dual Data Alignment Makes AI-Generated Image Detector Easier Generalizable (5 citations)
- U-REPA: Aligning Diffusion U-Nets to ViTs (4 citations)
- Boosting Generative Image Modeling via Joint Image-Feature Synthesis (3 citations)
- From Specificity to Generality: Revisiting Generalizable Artifacts in Detecting Face Deepfakes (3 citations)
- PipeFusion: Patch-level Pipeline Parallelism for Diffusion Transformers Inference (3 citations)
- Pixel-Perfect Depth with Semantics-Prompted Diffusion Transformers (3 citations)
- CAT: Content-Adaptive Image Tokenization (2 citations)
- Generative Modeling of Full-Atom Protein Conformations using Latent Diffusion on Graph Embeddings (2 citations)
- Towards Better & Faster Autoregressive Image Generation: From the Perspective of Entropy (2 citations)
- Understand Before You Generate: Self-Guided Training for Autoregressive Image Generation (2 citations)
- Adaptive Classifier-Free Guidance via Dynamic Low-Confidence Masking (1 citation)
and 90 more in the interactive map.
Text-to-image diffusion · 77 papers
- Seg4Diff: Unveiling Open-Vocabulary Semantic Segmentation in Text-to-Image Diffusion Transformers (8 citations)
- XVerse: Consistent Multi-Subject Control of Identity and Semantic Attributes via DiT Modulation (6 citations)
- DICEPTION: A Generalist Diffusion Model for Visual Perceptual Tasks (5 citations)
- UltraHR-100K: Enhancing UHR Image Synthesis with A Large-Scale High-Quality Dataset (4 citations)
- HiFlow: Training-free High-Resolution Image Generation with Flow-Aligned Guidance (3 citations)
- EraseFlow: Learning Concept Erasure Policies via GFlowNet-Driven Alignment (2 citations)
- LinEAS: End-to-end Learning of Activation Steering with a Distributional Loss (2 citations)
- OneIG-Bench: Omni-dimensional Nuanced Evaluation for Image Generation (2 citations)
- SceneDecorator: Towards Scene-Oriented Story Generation with Scene Planning and Scene Consistency (2 citations)
- SPARKE: Scalable Prompt-Aware Diversity and Novelty Guidance in Diffusion Models via RKE Score (2 citations)
- When Are Concepts Erased From Diffusion Models? (2 citations)
- Aligning Text to Image in Diffusion Models is Easier Than You Think (1 citation)
and 65 more in the interactive map.
Editing and style transfer · 41 papers
- ImgEdit: A Unified Image Editing Dataset and Benchmark (18 citations)
- Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer (16 citations)
- CSGO: Content-Style Composition in Text-to-Image Generation (7 citations)
- ThermalGen: Style-Disentangled Flow-Based Generative Models for RGB-to-Thermal Image Translation (4 citations)
- Free-Lunch Color-Texture Disentanglement for Stylized Image Generation (3 citations)
- FSI-Edit: Frequency and Stochasticity Injection for Flexible Diffusion-Based Image Editing (3 citations)
- CamEdit: Continuous Camera Parameter Control for Photorealistic Image Editing (2 citations)
- Color Conditional Generation with Sliced Wasserstein Guidance (2 citations)
- DNAEdit: Direct Noise Alignment for Text-Guided Rectified Flow Editing (2 citations)
- DreamLight: Towards Harmonious and Consistent Image Relighting (2 citations)
- OmniConsistency: Learning Style-Agnostic Consistency from Paired Stylization Data (2 citations)
- RelationAdapter: Learning and Transferring Visual Relation with Diffusion Transformers (2 citations)
and 29 more in the interactive map.
Image restoration · 36 papers
- DeblurDiff: Real-Word Image Deblurring with Generative Diffusion Models (5 citations)
- BlurDM: A Blur Diffusion Model for Image Deblurring (3 citations)
- LuxDiT: Lighting Estimation with Video Diffusion Transformer (3 citations)
- Deno-IF: Unsupervised Noisy Visible and Infrared Image Fusion Method (2 citations)
- Bio-Inspired Image Restoration (1 citation)
- Efficient RAW Image Deblurring with Adaptive Frequency Modulation (1 citation)
- FRN: Fractal-Based Recursive Spectral Reconstruction Network (1 citation)
- Image Stitching in Adverse Condition: A Bidirectional-Consistency Learning Framework and Benchmark (1 citation)
- Latent Harmony: Synergistic Unified UHD Image Restoration via Latent Space Regularization and Controllable Refinement (1 citation)
- MODEM: A Morton-Order Degradation Estimation Mechanism for Adverse Weather Image Recovery (1 citation)
- NPN: Non-Linear Projections of the Null-Space for Imaging Inverse Problems (1 citation)
- Real-World Adverse Weather Image Restoration via Dual-Level Reinforcement Learning with High-Quality Cold Start (1 citation)
and 24 more in the interactive map.
Super-resolution · 15 papers
- DOVE: Efficient One-Step Diffusion Model for Real-World Video Super-Resolution (8 citations)
- 4KAgent: Agentic Any Image to 4K Super-Resolution (5 citations)
- DP²O-SR: Direct Perceptual Preference Optimization for Real-World Image Super-Resolution (4 citations)
- Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment (2 citations)
- DIFFSSR: Stereo Image Super-resolution Using Differential Transformer (1 citation)
- FIPER: Factorized Features for Robust Image Super-Resolution and Compression (1 citation)
- Audio Super-Resolution with Latent Bridge Models
- Image Super-Resolution with Guarantees via Conformalized Generative Models
- OCTDiff: Bridged Diffusion Model for Portable OCT Super-Resolution and Enhancement
- PBR-SR: Mesh PBR Texture Super Resolution from 2D Image Priors
- PocketSR: The Super-Resolution Expert in Your Pocket Mobiles
- SRSR: Enhancing Semantic Accuracy in Real-World Image Super-Resolution with Spatially Re-Focused Text-Conditioning
and 3 more in the interactive map.