PEEKABOO: Interactive Video Generation via Masked-Diffusion
This work addresses the problem of limited user interactivity in video generation for creative applications, offering a novel method that is incremental but impactful.
The paper tackles the lack of interactive control in video generation models by introducing Peekaboo, a masked attention module that enables spatio-temporal control without extra training or inference overhead, achieving up to a 3.8x improvement in mIoU over baselines while maintaining latency.
Modern video generation models like Sora have achieved remarkable success in producing high-quality videos. However, a significant limitation is their inability to offer interactive control to users, a feature that promises to open up unprecedented applications and creativity. In this work, we introduce the first solution to equip diffusion-based video generation models with spatio-temporal control. We present Peekaboo, a novel masked attention module, which seamlessly integrates with current video generation models offering control without the need for additional training or inference overhead. To facilitate future research, we also introduce a comprehensive benchmark for interactive video generation. This benchmark offers a standardized framework for the community to assess the efficacy of emerging interactive video generation models. Our extensive qualitative and quantitative assessments reveal that Peekaboo achieves up to a 3.8x improvement in mIoU over baseline models, all while maintaining the same latency. Code and benchmark are available on the webpage.