Enhance-A-Video: Better Generated Video for Free
This work addresses video generation enhancement for AI researchers and practitioners, but it is incremental as it builds on existing DiT-based frameworks without retraining.
The paper tackles the problem of improving coherence and quality in DiT-based generated videos by introducing Enhance-A-Video, a training-free approach that enhances cross-frame correlations using non-diagonal temporal attention distributions, resulting in promising improvements in temporal consistency and visual quality across various models.
DiT-based video generation has achieved remarkable results, but research into enhancing existing models remains relatively unexplored. In this work, we introduce a training-free approach to enhance the coherence and quality of DiT-based generated videos, named Enhance-A-Video. The core idea is enhancing the cross-frame correlations based on non-diagonal temporal attention distributions. Thanks to its simple design, our approach can be easily applied to most DiT-based video generation frameworks without any retraining or fine-tuning. Across various DiT-based video generation models, our approach demonstrates promising improvements in both temporal consistency and visual quality. We hope this research can inspire future explorations in video generation enhancement.