图像生成论文趋势 2023–2026
文生图与图像合成
统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 5,184 篇被归入该主题的论文。
主要结论
- 图像生成论文从 2023 年的 780 篇增长到 2026 年的 1,721 篇，三年间扩大到约 2.2 倍。
- CVPR 是发图像生成论文最多的会议，2023–2026 年共 2,262 篇，占四个会议总量的 44%。
- 交叉方向里涨得最快的是深度学习 · generative models and autoencoders：同时属于两个主题的论文从 2023 年的 35 篇增至 2026 年的 150 篇。
- 在「计算机视觉」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的视频理解是 2.6 倍，图像生成是 2.2 倍。
各会议论文数
按年份和会议统计
| 年份 | NeurIPS | ICLR | ICML | CVPR | 合计 |
|---|---|---|---|---|---|
| 2023 | 170 | 118 | 74 | 418 | 780 |
| 2024 | 318 | 199 | 156 | 543 | 1,216 |
| 2025 | 430 | 278 | 147 | 612 | 1,467 |
| 2026 | 453 | 349 | 230 | 689 | 1,721 |
代表论文
按年份和会议均匀挑选，方便快速看到每一届的样貌
- NeurIPS 2026$\texttt{bispectrum}$: Selective $G$-Bispectra Made Practical
- ICML 2026Klein Hyperbolic Metric Learning
- ICLR 2026A Bayesian Nonparametric Framework for Private, Fair, and Balanced Tabular Data Synthesis
- CVPR 2026Adaptive Learned Image Compression with Graph Neural Networks
- NeurIPS 2025Alchemist: Turning Public Text-to-Image Data into Generative Gold
- ICML 2025BiMark: Unbiased Multilayer Watermarking for Large Language Models
- ICLR 2025Accelerating Auto-regressive Text-to-Image Generation with Training-free Speculative Jacobi Decoding
- CVPR 2025A Polarization-Aided Transformer for Image Deblurring via Motion Vector Decomposition
- NeurIPS 2024Einsum Benchmark: Enabling the Development of Next-Generation Tensor Execution Engines
- ICML 2024Adaptive Feature Selection for No-Reference Image Quality Assessment by Mitigating Semantic Noise Sensitivity
- ICLR 2024AutomaTikZ: Text-Guided Synthesis of Scientific Vector Graphics with TikZ
- CVPR 2024DemoCaricature: Democratising Caricature Generation with a Rough Sketch
- NeurIPS 20234D Panoptic Scene Graph Generation
- ICML 2023Architecture-Agnostic Masked Image Modeling -- From ViT back to CNN
- ICLR 2023Accurate Image Restoration with Attention Retractable Transformer
- CVPR 20233D GAN Inversion With Facial Symmetry Prior
- NeurIPS 2026A Theory of Time-Sensitive Language Generation: Sparse Hallucination Beats Mode Collapse
- ICML 2026Noisy-Channel Minimum Bayes Risk Decoding
- ICLR 2026All Patches Matter, More Patches Better: Enhance AI-Generated Image Detection via Panoptic Patch Learning
- CVPR 2026AIMDepth Asymmetric Image-Event Mamba for Monocular Depth Estimation
全部图像生成论文
共 944 篇，第 7 / 19 页
- ICLR 2026Text2Interact: High-Fidelity and Diverse Text-to-Two-Person Interaction Generation
- ICLR 2026The Intricate Dance of Prompt Complexity, Quality, Diversity and Consistency in T2I Models
- ICLR 2026There and Back Again: On the relation between Noise and Image Inversions in Diffusion Models
- ICLR 2026TIGaussian: Disentangle Gaussians for Spatial-Awared Text-Image-3D Alignment
- ICLR 2026Time-to-Move: Training-Free Motion-Controlled Video Generation via Dual-Clock Denoising
- ICLR 2026TIPO: Text to Image with Text Pre-sampling for Prompt Optimization
- ICLR 2026ToolACE-MT: Non-Autoregressive Generation for Agentic Multi-Turn Interaction
- ICLR 2026Topology-Preserved Auto-regressive Mesh Generation in the Manner of Weaving Silk
- ICLR 2026ToProVAR: Efficient Visual Autoregressive Modeling via Tri-Dimensional Entropy-Aware Semantic Analysis and Sparsity Optimization
- ICLR 2026TOUCH: Text-guided Controllable Generation of Free-Form Hand-Object Interactions
- ICLR 2026Towards One-step Causal Video Generation via Adversarial Self-Distillation
- ICLR 2026Towards Sequence Modeling Alignment between Tokenizer and Autoregressive Model
- ICLR 2026Towards Text-Mask Consistency in Medical Image Segmentation
- ICLR 2026Training-Free Reward-Guided Image Editing via Trajectory Optimal Control
- ICLR 2026Trajectory Generation with Conservative Value Guidance for Offline Reinforcement Learning
- ICLR 2026TrajTok: What makes for a good trajectory tokenizer in behavior generation?
- ICLR 2026TriC-Motion: Tri-Domain Causal Modeling Grounded Text-to-Motion Generation
- ICLR 2026Trust The Typical
- ICLR 2026TS-Attn: Temporal-wise Separable Attention for Multi-Event Video Generation
- ICLR 2026TTOM: Test-Time Optimization and Memorization for Compositional Video Generation
- ICLR 2026Turning Internal Gap into Self-Improvement: Promoting the Generation-Understanding Unification in MLLMs
- ICLR 2026TwinFlow: Realizing One-step Generation on Large Models with Self-adversarial Flows
- ICLR 2026Ultra-Fast Language Generation via Discrete Diffusion Divergence Instruct
- ICLR 2026Uncovering Conceptual Blindspots in Generative Image Models Using Sparse Autoencoders
- ICLR 2026Understanding vs. Generation: Navigating Optimization Dilemma in Multimodal Models
- ICLR 2026Uni-CoT: Towards Unified Chain-of-Thought Reasoning Across Text and Vision
- ICLR 2026UniCalli: A Unified Diffusion Framework for Column-Level Generation and Recognition of Chinese Calligraphy
- ICLR 2026UniEdit-Flow: Unleashing Inversion and Editing in the Era of Flow Models
- ICLR 2026UniF$^2$ace: A $\underline{Uni}$fied $\underline{F}$ine-grained $\underline{Face}$ Understanding and Generation Model
- ICLR 2026Uniform Discrete Diffusion with Metric Path for Video Generation
- ICLR 2026UniLiP: Adapting CLIP for Unified Multimodal Understanding, Generation and Editing
- ICLR 2026Using maximal information auxiliary variables to improve synthetic data generation based on TabPFN foundation models
- ICLR 2026Value Matching: Scalable and Gradient-Free Reward-Guided Flow Adaptation
- ICLR 2026VARestorer: One-Step VAR Distillation for Real-World Image Super-Resolution
- ICLR 2026VINCIE: Unlocking In-context Image Editing from Video
- ICLR 2026VIRTUE: Visual-Interactive Text-Image Universal Embedder
- ICLR 2026VisualPrompter: Semantic-Aware Prompt Optimization with Visual Feedback for Text-to-Image Synthesis
- ICLR 2026VLM-Guided Adaptive Negative Prompting for Creative Generation
- ICLR 2026VMDiff: Visual Mixing Diffusion for Limitless Cross-Object Synthesis
- ICLR 2026VOGUE: Unified Understanding, Generation, and Editing for Videos
- ICLR 2026VSF: Simple, Efficient, and Effective Negative Guidance in Few-Step Image Generation Models By Value Sign Flip
- ICLR 2026W-EDIT: A Wavelet-Based Frequency-Aware Framework for Text-Driven Image Editing
- ICLR 2026Watermark-based Attribution of AI-Generated Content
- ICLR 2026Weak-to-Strong Diffusion
- ICLR 2026WeTok: Powerful Discrete Tokenization for High-Fidelity Visual Reconstruction
- ICLR 2026What Exactly Does Guidance Do in Masked Discrete Diffusion Models
- ICLR 2026WithAnyone: Toward Controllable and ID Consistent Image Generation
- ICLR 2026WorldSplat: Gaussian-Centric Feed-Forward 4D Scene Generation for Autonomous Driving
- ICLR 2026XIL: Cross-Expanding Incremental Learning
- ICLR 2025$\text{I}^2\text{AM}$: Interpreting Image-to-Image Latent Diffusion Models via Bi-Attribution Maps
在 Claude、Cursor 里直接问这个趋势
接上 Paperlist MCP 后，写综述、挑方向或准备投稿时，不用离开对话框就能拿到这页的数字：每年各会议的论文数、增长倍数、涨得最快的子方向和代表论文，AI 会按你的问题整理。
可以直接复制这句话问 AI
用 paperlist 查图像生成在 CVPR 2023–2026 的论文趋势，哪一年涨得最快？
AI 会调用的工具：get_topic_stats {"topic": "image-generation"}
常见问题
- 2023–2026 年四大会议一共有多少篇图像生成论文？
- 共 5,184 篇，按年份分别是 2023 年 780 篇、2024 年 1,216 篇、2025 年 1,467 篇、2026 年 1,721 篇。
- NeurIPS 2026 有多少篇图像生成论文？
- NeurIPS 2026 有 453 篇图像生成论文，上一届是 430 篇。
- ICLR 2026 有多少篇图像生成论文？
- ICLR 2026 有 349 篇图像生成论文，上一届是 278 篇。
- ICML 2026 有多少篇图像生成论文？
- ICML 2026 有 230 篇图像生成论文，上一届是 147 篇。
- CVPR 2026 有多少篇图像生成论文？
- CVPR 2026 有 689 篇图像生成论文，上一届是 612 篇。
- 哪个会议的图像生成论文最多？
- CVPR，2023–2026 年共 2,262 篇。