CCM-DiT: Camera-pose Controllable Method for DiT-based Video Generation
Keywords: Video Generation, Diffusion Models
Abstract: Despite the significant advancements made by Diffusion Transformer (DiT)-based methods in video generation, there remains a notable gap with camera-pose perspectives. Existing works such as OpenSora do not adhere precisely to anticipated trajectories, thereby limiting the utility in downstream applications such as content creation.
Therefore, we introduce a novelty approach that achieves fine-grained control by embedding sparse camera-pose information into the temporal self-attention layers. We employ LoRA to minimize the impact on the original attention layer parameters during fine-tuning and enhance the supervision of camera-pose in the loss function.
After fine-tuning the OpenSora’s ST-DiT framework on the RealEstate10K dataset, experiments demonstrate that our method outperforms LDM-based methods for long video generation, while maintaining optimal performance in trajectory consistency and object consistency.
Primary Area: generative models
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: 13388
Loading