文生视频论文趋势 2023–2026

主要结论

• 文生视频论文从 2023 年的 170 篇增长到 2026 年的 448 篇，三年间扩大到约 2.6 倍。
• CVPR 是发文生视频论文最多的会议，2023–2026 年共 574 篇，占四个会议总量的 48%。
• 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 2 篇增至 2026 年的 26 篇。
• 在「多模态生成」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的叙事生成是 15 倍，文生视频是 2.6 倍。

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 2026CausalCine: Real-Time Autoregressive Generation for Multi-Shot Video Narratives
• ICML 2026DFSAttn: Dynamic Fine-grained Sparse Attention for Efficient Video Generation
• ICLR 2026$PhyWorldBench$: A Comprehensive Evaluation of Physical Realism in Text-to-Video Models
• CVPR 20263D-Aware Implicit Motion Control for View-Adaptive Human Video Generation
• NeurIPS 2025Dynamic Shadow Unveils Invisible Semantics for Video Outpainting
• ICML 2025GRADEO: Towards Human-Like Evaluation for Text-to-Video Generation via Multi-Step Reasoning
• ICLR 20253DTrajMaster: Mastering 3D Trajectory for Multi-Entity Motion in Video Generation
• CVPR 20254D-Fly: Fast 4D Reconstruction from a Single Monocular Video
• NeurIPS 2024ChronoMagic-Bench: A Benchmark for Metamorphic Evaluation of Text-to-Time-lapse Video Generation
• ICML 2024Q-Align: Teaching LMMs for Visual Scoring via Discrete Text-Defined Levels
• ICLR 2024Provable Robust Watermarking for AI-Generated Text
• CVPR 2024Enhanced Motion-Text Alignment for Image-to-Video Transfer Learning
• NeurIPS 2023Any-to-Any Generation via Composable Diffusion
• ICML 2023Controlled Text Generation with Natural Language Instructions
• ICLR 2023CogVideo: Large-scale Pretraining for Text-to-Video Generation via Transformers
• CVPR 20231000 FPS HDR Video With a Spike-RGB Hybrid Camera
• NeurIPS 2026Character Mixing for Video Generation
• ICML 2026Noisy-Channel Minimum Bayes Risk Decoding
• ICLR 2026Anchor Frame Bridging for Coherent First-Last Frame Video Generation
• CVPR 2026Building a Precise Video Language with Human-AI Oversight

全部文生视频论文

共 91 篇，第 2 / 2 页

51. NeurIPS 2025Policy Optimized Text-to-Image Pipeline Design
52. NeurIPS 2025PolyVivid: Vivid Multi-Subject Video Generation with Cross-Modal Interaction and Enhancement
53. NeurIPS 2025Precise Information Control in Long-Form Text Generation
54. NeurIPS 2025Preventing Shortcuts in Adapter Training via Providing the Shortcuts
55. NeurIPS 2025Radial Attention: $\mathcal O(n \log n)$ Sparse Attention for Long Video Generation
56. NeurIPS 2025Rare Text Semantics Were Always There in Your Diffusion Transformer
57. NeurIPS 2025Rebalancing Contrastive Alignment with Bottlenecked Semantic Increments in Text-Video Retrieval
58. NeurIPS 2025REGen: Multimodal Retrieval-Embedded Generation for Long-to-Short Video Editing
59. NeurIPS 2025Restage4D: Reanimating Deformable 3D Reconstruction from a Single Video
60. NeurIPS 2025ROSE: Remove Objects with Side Effects in Videos
61. NeurIPS 2025Safe-Sora: Safe Text-to-Video Generation via Graphical Watermarking
62. NeurIPS 2025Sekai: A Video Dataset towards World Exploration
63. NeurIPS 2025Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion
64. NeurIPS 2025SoPo: Text-to-Motion Generation Using Semi-Online Preference Optimization
65. NeurIPS 2025Sparse VideoGen2: Accelerate Video Generation with Sparse Attention via Semantic-Aware Permutation
66. NeurIPS 2025Speculative Jacobi-Denoising Decoding for Accelerating Autoregressive Text-to-image Generation
67. NeurIPS 2025Storyboard-guided Alignment for Fine-grained Video Action Recognition
68. NeurIPS 2025StreamBridge: Turning Your Offline Video Large Language Model into a Proactive Streaming Assistant
69. NeurIPS 2025T2V-OptJail: Discrete Prompt Optimization for Text-to-Video Jailbreak Attacks
70. NeurIPS 2025Temporal In‑Context Fine‑Tuning for Versatile Control of Video Diffusion Models
71. NeurIPS 2025The Matrix: Infinite-Horizon World Generation with Real-Time Moving Control
72. NeurIPS 2025Towards Physical Understanding in Video Generation: A 3D Point Regularization Approach
73. NeurIPS 2025Training-Free Efficient Video Generation via Dynamic Token Carving
74. NeurIPS 2025Training-Free Safe Text Embedding Guidance for Text-to-Image Diffusion Models
75. NeurIPS 2025UltraVideo: High-Quality UHD Video Dataset with Comprehensive Captions
76. NeurIPS 2025UniTransfer: Video Concept Transfer via Progressive Spatio-Temporal Decomposition
77. NeurIPS 2025Unleashing the Potential of Multimodal LLMs for Zero-Shot Spatio-Temporal Video Grounding
78. NeurIPS 2025VaporTok: RL-Driven Adaptive Video Tokenizer with Prior & Task Awareness
79. NeurIPS 2025VFRTok: Variable Frame Rates Video Tokenizer with Duration-Proportional Information Assumption
80. NeurIPS 2025VIBE: Annotation-Free Video-to-Text Information Bottleneck Evaluation for TL;DR
81. NeurIPS 2025Video World Models with Long-term Spatial Memory
82. NeurIPS 2025VideoMAR: Autoregressive Video Generation with Continuous Tokens
83. NeurIPS 2025VideoREPA: Learning Physics for Video Generation through Relational Alignment with Foundation Models
84. NeurIPS 2025VideoUFO: A Million-Scale User-Focused Dataset for Text-to-Video Generation
85. NeurIPS 2025ViewPoint: Panoramic Video Generation with Pretrained Diffusion Models
86. NeurIPS 2025VimoRAG: Video-based Retrieval-augmented 3D Motion Generation for Motion Language Models
87. NeurIPS 2025VORTA: Efficient Video Diffusion via Routing Sparse Attention
88. NeurIPS 2025Whole-Body Conditioned Egocentric Video Prediction
89. NeurIPS 2025WISA: World simulator assistant for physics-aware text-to-video generation
90. NeurIPS 2025WorldModelBench: Judging Video Generation Models As World Models
91. NeurIPS 2025WorldWeaver: Generating Long-Horizon Video Worlds via Rich Perception

常见问题

2023–2026 年四大会议一共有多少篇文生视频论文？

共 1,185 篇，按年份分别是 2023 年 170 篇、2024 年 247 篇、2025 年 320 篇、2026 年 448 篇。

NeurIPS 2026 有多少篇文生视频论文？

NeurIPS 2026 有 101 篇文生视频论文，上一届是 91 篇。

ICLR 2026 有多少篇文生视频论文？

ICLR 2026 有 94 篇文生视频论文，上一届是 59 篇。

ICML 2026 有多少篇文生视频论文？

ICML 2026 有 55 篇文生视频论文，上一届是 14 篇。

CVPR 2026 有多少篇文生视频论文？

CVPR 2026 有 198 篇文生视频论文，上一届是 156 篇。

哪个会议的文生视频论文最多？

CVPR，2023–2026 年共 574 篇。