Ctrl-V: Higher Fidelity Video Generation with Bounding-Box Controlled Object Motion

Ge Ya Luo, ZhiHao Luo, Anthony Gosselin, Alexia Jolicoeur-Martineau, Christopher Pal

Keywords: video generation, video synthesis, computer vision, diffusion models, autonomous driving, controllable video generation

Abstract: Controllable video generation has attracted significant attention, largely due to advances in video diffusion models. In domains like autonomous driving in particular it can be critical to develop highly accurate predictions for object motions. This paper tackles a crucial challenge of how to exert precise control over object motion for realistic video synthesis in a safety critical setting. To achieve this, we 1) use a separate, specialized model to predict object bounding-box trajectories given the past and optionally future locations of bounding boxes, and 2) generate video conditioned on these high quality trajectory predictions. This formulation allows us to test the quality of different model components separately and together. To address the challenges of conditioning video generation on object trajectories in settings where objects may disappear and appear within a scene, we propose an approach based on rendering 2D or 3D boxes as videos. Our method, Ctrl-V, leverages modified and fine-tuned Stable Video Diffusion (SVD) models to solve both trajectory and video generation. Extensive experiments conducted on the KITTI, Virtual-KITTI 2, BDD 100k, and nuScenes datasets validate the effectiveness of our approach in producing realistic and controllable video generation.

Primary Area: applications to computer vision, audio, language, and other modalities

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 8018