VideoTetris: Towards Compositional Text-to-Video Generation
This addresses the problem of handling complex video generation scenarios for users in AI and multimedia, representing an incremental improvement over existing methods.
The paper tackles the challenge of generating complex, long videos with multiple objects or dynamic changes from text by proposing VideoTetris, a framework that achieves impressive qualitative and quantitative results in compositional text-to-video generation.
Diffusion models have demonstrated great success in text-to-video (T2V) generation. However, existing methods may face challenges when handling complex (long) video generation scenarios that involve multiple objects or dynamic changes in object numbers. To address these limitations, we propose VideoTetris, a novel framework that enables compositional T2V generation. Specifically, we propose spatio-temporal compositional diffusion to precisely follow complex textual semantics by manipulating and composing the attention maps of denoising networks spatially and temporally. Moreover, we propose an enhanced video data preprocessing to enhance the training data regarding motion dynamics and prompt understanding, equipped with a new reference frame attention mechanism to improve the consistency of auto-regressive video generation. Extensive experiments demonstrate that our VideoTetris achieves impressive qualitative and quantitative results in compositional T2V generation. Code is available at: https://github.com/YangLing0818/VideoTetris