When and Where do Events Switch in Multi-Event Video Generation?

2025/10/03 by Liao, Ruotong, Huang, Guowen, Cheng, Qing +3
#Artificial Intelligence (cs.AI) #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.2510.03049

Abstract

Text-to-video (T2V) generation has surged in response to challenging questions, especially when a long video must depict multiple sequential events with temporal coherence and controllable content. Existing methods that extend to multi-event generation omit an inspection of the intrinsic factor in event shifting. The paper aims to answer the central question: When and where multi-event prompts control event transition during T2V generation. This work introduces MEve, a self-curated prompt suite for evaluating multi-event text-to-video (T2V) generation, and conducts a systematic study of two representative model families, i.e., OpenSora and CogVideoX. Extensive experiments demonstrate the importance of early intervention in denoising steps and block-wise model layers, revealing the essential factor for multi-event video generation and highlighting the possibilities for multi-event conditioning in future models.

Citations

• Demystifying Flux Architecture
• Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
• FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space
• ShotAdapter: Text-to-Multi-Shot Video Generation with Diffusion Models
• Open-Sora 2.0: Training a Commercial-Level Video Generation Model in 200k
• Open-Sora: Democratizing Efficient Video Production for All
• Evaluation Agent: Efficient and Promptable Evaluation Framework for Visual Generative Models
• Mind the Time: Temporally-Controlled Multi-Event Video Generation
• HunyuanVideo: A Systematic Framework For Large Video Generative Models
• Deep Compression Autoencoder for Efficient High-Resolution Diffusion Models
• StoryMaker: Towards Holistic Consistent Characters in Text-to-image Generation
• CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer
• MiraData: A Large-Scale Video Dataset with Long Durations and Structured Captions
• StoryDiffusion: Consistent Self-Attention for Long-Range Image and Video Generation
• Training-Free Consistent Text-to-Image Generation
• Lumiere: A Space-Time Diffusion Model for Video Generation
• Latte: Latent Diffusion Transformer for Video Generation
• Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasets
• PixArt-α: Fast Training of Diffusion Transformer for Photorealistic Text-to-Image Synthesis
• InternVid: A Large-scale Video-Text Dataset for Multimodal Understanding and Generation
• DINOv2: Learning Robust Visual Features without Supervision
• Tune-A-Video: One-Shot Tuning of Image Diffusion Models for Text-to-Video Generation
• Scalable Diffusion Models with Transformers
• DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation
• DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation
• An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion
• CogVideo: Large-scale Pretraining for Text-to-Video Generation via Transformers
• Colossal-AI: A Unified Deep Learning System For Large-Scale Parallel Training
• Learning Transferable Visual Models From Natural Language Supervision
• Denoising Diffusion Probabilistic Models
• Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer

Related

• Mind the Time: Temporally-Controlled Multi-Event Video Generation
• MEVG: Multi-event Video Generation with Text-to-Video Models
• Enhancing Scene Transition Awareness in Video Generation via Post-Training
• What Happens When: Learning Temporal Orders of Events in Videos
• TC-Bench: Benchmarking Temporal Compositionality in Text-to-Video and Image-to-Video Generation