ICCV 2025 accepted papers on Image generation
439 of the 2,701 papers accepted at ICCV 2025 (16.3%) are on Image generation, compared with 12.0% at ICCV 2023. They fall into 7 topics.
See ICCV 2025 in the interactive map
Submitting to ICCV? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Text-to-image diffusion · 160 papers
- FlowEdit: Inversion-Free Text-Based Editing Using Pre-Trained Flow Models (134 citations)
- Less-to-More Generalization: Unlocking More Controllability by In-Context Generation (123 citations)
- Lumina-Image 2.0: A Unified and Efficient Image Generative Framework (82 citations)
- Golden Noise for Diffusion Models: A Learning Framework (68 citations)
- LayerTracer: Cognitive-Aligned Layered SVG Synthesis via Diffusion Transformer (51 citations)
- CreatiLayout: Siamese Multimodal Diffusion Transformer for Creative Layout-to-Image Generation (45 citations)
- DeepMesh: Auto-Regressive Artist-mesh Creation with Reinforcement Learning (44 citations)
- Democratizing Text-to-Image Masked Generative Models with Compact Text-Aware One-Dimensional Tokens (42 citations)
- LMM4LMM: Benchmarking and Evaluating Large-multimodal Image Generation with LMMs (42 citations)
- UniCombine: Unified Multi-Conditional Combination with Diffusion Transformer (40 citations)
- HPSv3: Towards Wide-Spectrum Human Preference Score (34 citations)
- VisualCloze: A Universal Image Generation Framework via Visual In-Context Learning (33 citations)
and 148 more in the interactive map.
Image restoration · 77 papers
- Frequency Domain-Based Diffusion Model for Unpaired Image Dehazing (29 citations)
- EAMamba: Efficient All-Around Vision State Space Model for Image Restoration (28 citations)
- FoundIR: Unleashing Million-scale Training Data to Advance Foundation Models for Image Restoration (28 citations)
- StableCodec: Taming One-Step Diffusion for Extreme Image Compression (22 citations)
- MP-HSIR: A Multi-Prompt Framework for Universal Hyperspectral Image Restoration (18 citations)
- Reverse Convolution and Its Applications to Image Restoration (17 citations)
- Learning to See in the Extremely Dark (15 citations)
- RS-vHeat: Heat Conduction Guided Efficient Remote Sensing Foundation Model (15 citations)
- CCMNet: Leveraging Calibrated Color Correction Matrices for Cross-Camera Color Constancy (14 citations)
- Retinex-MEF: Retinex-based Glare Effects Aware Unsupervised Multi-Exposure Image Fusion (14 citations)
- Laboring on less labors: RPCA Paradigm for Pan-sharpening (11 citations)
- LATINO-PRO: LAtent consisTency INverse sOlver with PRompt Optimization (11 citations)
and 65 more in the interactive map.
Autoregressive and latent generation · 71 papers
- OminiControl: Minimal and Universal Control for Diffusion Transformer (246 citations)
- Randomized Autoregressive Visual Generation (105 citations)
- EasyControl: Adding Efficient and Flexible Control for Diffusion Transformer (104 citations)
- Beyond Next-Token: Next-X Prediction for Autoregressive Visual Generation (65 citations)
- LEGION: Learning to Ground and Explain for Synthetic Image Detection (49 citations)
- Scalable Image Tokenization with Index Backpropagation Quantization (47 citations)
- GigaTok: Scaling Visual Tokenizers to 3 Billion Parameters for Autoregressive Image Generation (33 citations)
- FastVAR: Linear Visual Autoregressive Modeling via Cached Token Pruning (32 citations)
- FlowTok: Flowing Seamlessly Across Text and Image Tokens (30 citations)
- Neighboring Autoregressive Modeling for Efficient Visual Generation (30 citations)
- Bridging Continuous and Discrete Tokens for Autoregressive Visual Generation (27 citations)
- Flow to the Mode: Mode-Seeking Diffusion Autoencoders for State-of-the-Art Image Tokenization (27 citations)
and 59 more in the interactive map.
Fast diffusion sampling · 55 papers
- From Reusing to Forecasting: Accelerating Diffusion Models with TaylorSeers (98 citations)
- StreamDiffusion: A Pipeline-level Solution for Real-Time Interactive Generation (74 citations)
- SANA-Sprint: One-Step Diffusion with Continuous-Time Consistency Distillation (66 citations)
- FlowDPS : Flow-Driven Posterior Sampling for Inverse Problems (45 citations)
- Improved Noise Schedule for Diffusion Training (42 citations)
- Learning Few-Step Diffusion Models by Trajectory Distribution Matching (40 citations)
- Unleashing Vecset Diffusion Model for Fast Shape Generation (24 citations)
- Unraveling the Smoothness Properties of Diffusion Models: A Gaussian Mixture Perspective (21 citations)
- Adversarial Distribution Matching for Diffusion Distillation Towards Efficient Image and Video Synthesis (20 citations)
- Compression-Aware One-Step Diffusion Model for JPEG Artifact Removal (19 citations)
- Di[M]O: Distilling Masked Diffusion Models into One-step Generator (19 citations)
- Revelio: Interpreting and leveraging semantic information in diffusion models (17 citations)
and 43 more in the interactive map.
Editing and style transfer · 48 papers
- Any2AnyTryon: Leveraging Adaptive Position Embeddings for Versatile Virtual Clothing Tasks (58 citations)
- KV-Edit: Training-Free Image Editing for Precise Background Preservation (50 citations)
- UniPortrait: A Unified Framework for Identity-Preserving Single- and Multi-Human Image Personalization (45 citations)
- LBM: Latent Bridge Matching for Fast Image-to-Image Translation (33 citations)
- OmniPaint: Mastering Object-Oriented Editing via Disentangled Insertion-Removal Inpainting (31 citations)
- UnZipLoRA: Separating Content and Style from a Single Image (24 citations)
- Multi-turn Consistent Image Editing (23 citations)
- SliderSpace: Decomposing the Visual Capabilities of Diffusion Models (21 citations)
- ICE-Bench: A Unified and Comprehensive Benchmark for Image Creating and Editing (19 citations)
- SAGI: Semantically Aligned and Uncertainty Guided AI Image Inpainting (19 citations)
- Magic Insert: Style-Aware Drag-and-Drop (16 citations)
- SuperEdit: Rectifying and Facilitating Supervision for Instruction-Based Image Editing (15 citations)
and 36 more in the interactive map.
Super-resolution · 25 papers
- Emulating Self-attention with Convolution for Efficient Image Super-Resolution (42 citations)
- DiT4SR: Taming Diffusion Transformer for Real-World Image Super-Resolution (37 citations)
- Generalized and Efficient 2D Gaussian Splatting for Arbitrary-scale Super-Resolution (22 citations)
- Perceive, Understand and Restore: Real-World Image Super-Resolution with Autoregressive Multimodal Generative Models (17 citations)
- Consistency Trajectory Matching for One-Step Generative Super-Resolution (12 citations)
- Fine-structure Preserved Real-world Image Super-resolution via Transfer VAE Training (12 citations)
- DuCos: Duality Constrained Depth Super-Resolution via Foundation Model (11 citations)
- NeurOp-Diff: Continuous Remote Sensing Image Super-Resolution via Neural Operator Diffusion (8 citations)
- Diffusion Transformer meets Multi-level Wavelet Spectrum for Single Image Super-Resolution (6 citations)
- PAN-Crafter: Learning Modality-Consistent Alignment for PAN-Sharpening (5 citations)
- Ultra High-Resolution Image Inpainting with Patch-Based Content Consistency Adapter (5 citations)
- Fast Image Super-Resolution via Consistency Rectified Flow (4 citations)
and 13 more in the interactive map.
