Tuning-Free Long Video Generation via Global-Local Collaborative Diffusion
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866916559653437440 |
|---|---|
| author | Ma, Yongjia Chen, Junlin Di, Donglin Xie, Qi Fan, Lei Chen, Wei Gou, Xiaofei Zhao, Na Yang, Xun |
| author_facet | Ma, Yongjia Chen, Junlin Di, Donglin Xie, Qi Fan, Lei Chen, Wei Gou, Xiaofei Zhao, Na Yang, Xun |
| contents | Creating high-fidelity, coherent long videos is a sought-after aspiration. While recent video diffusion models have shown promising potential, they still grapple with spatiotemporal inconsistencies and high computational resource demands. We propose GLC-Diffusion, a tuning-free method for long video generation. It models the long video denoising process by establishing denoising trajectories through Global-Local Collaborative Denoising to ensure overall content consistency and temporal coherence between frames. Additionally, we introduce a Noise Reinitialization strategy which combines local noise shuffling with frequency fusion to improve global content consistency and visual diversity. Further, we propose a Video Motion Consistency Refinement (VMCR) module that computes the gradient of pixel-wise and frequency-wise losses to enhance visual consistency and temporal smoothness. Extensive experiments, including quantitative and qualitative evaluations on videos of varying lengths (\textit{e.g.}, 3\times and 6\times longer), demonstrate that our method effectively integrates with existing video diffusion models, producing coherent, high-fidelity long videos superior to previous approaches. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_05484 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Tuning-Free Long Video Generation via Global-Local Collaborative Diffusion Ma, Yongjia Chen, Junlin Di, Donglin Xie, Qi Fan, Lei Chen, Wei Gou, Xiaofei Zhao, Na Yang, Xun Computer Vision and Pattern Recognition Creating high-fidelity, coherent long videos is a sought-after aspiration. While recent video diffusion models have shown promising potential, they still grapple with spatiotemporal inconsistencies and high computational resource demands. We propose GLC-Diffusion, a tuning-free method for long video generation. It models the long video denoising process by establishing denoising trajectories through Global-Local Collaborative Denoising to ensure overall content consistency and temporal coherence between frames. Additionally, we introduce a Noise Reinitialization strategy which combines local noise shuffling with frequency fusion to improve global content consistency and visual diversity. Further, we propose a Video Motion Consistency Refinement (VMCR) module that computes the gradient of pixel-wise and frequency-wise losses to enhance visual consistency and temporal smoothness. Extensive experiments, including quantitative and qualitative evaluations on videos of varying lengths (\textit{e.g.}, 3\times and 6\times longer), demonstrate that our method effectively integrates with existing video diffusion models, producing coherent, high-fidelity long videos superior to previous approaches. |
| title | Tuning-Free Long Video Generation via Global-Local Collaborative Diffusion |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2501.05484 |