recently,Stability AIThe company announced the launch of a revolutionary video processing technology, Stable Video4D, which can transform a single-view video into eight new-view videos at different angles, providing creators with unprecedented flexibility and creativity.
Stable Video4D builds on the company's previously launched Stable Video Diffusion model. Instead of converting images into videos, the new model can take in video input and generate video outputs from multiple new perspectives, making a major leap from image-based video generation to full 3D dynamic video synthesis.
When using it, users only need to upload a video and specify the desired 3D camera position, and Stable Video4D can generate videos with 8 new perspectives, providing users with a full range of multi-angle perspectives. Currently, the model can generate 5 frames of video with 8 perspectives in about 40 seconds, and the entire 4D optimization process takes about 20-25 minutes.
Compared with previous methods, Stable Video4D can generate multiple new perspective videos at the same time, greatly improving the consistency in space and time. This not only ensures the consistency of objects in multiple perspectives and timestamps, but also realizes a lighter 4D optimization framework.
Stability AI said that Stable Video4D is currently in the research stage and is expected to be widely used in game development, video editing, virtual reality and other fields in the future. The company is actively optimizing the model to process a wider range of real-world videos.
Stable Video4D is now available on the Hugging Face platform. Stability AI looks forward to further improving the technology's potential to create realistic multi-angle videos through continued research and development. The company will continue to work with researchers, experts, and the community to drive technological innovation and continuously improve model performance.
Model address: https://huggingface.co/stabilityai/sv4d