Stability AI Launches Model That Transforms 2D Photos Into 3D Video Scenes
Stability AI introduces Stable Virtual Camera, a new generative AI tool that converts static images into immersive videos with realistic depth. The research preview is currently available for noncommercial use on Hugging Face.
Stability AI releases Stable Virtual Camera, a new generative AI model that transforms 2D images into immersive videos with realistic depth and perspective. Users input up to 32 images and specify desired camera angles to generate novel views of a scene. The model offers dynamic camera paths and presets such as Spiral, Dolly Zoom, Move, and Pan to give creators greater control over their digital filmmaking.
This research preview supports square, portrait, and landscape aspect ratios and produces videos up to 1,000 frames in length. However, Stability AI warns that the tool struggles with certain subjects and environments. Images featuring humans, animals, dynamic textures like water, and highly ambiguous scenes frequently cause flickering artifacts, especially when the target viewpoint differs significantly from the input photos.
Stable Virtual Camera is available for noncommercial research use and can be downloaded from the Hugging Face platform. This release is part of a broader turnaround effort for the beleaguered company, which recently appoints a new CEO, adds James Cameron to its board of directors, and partners with Arm on mobile audio generation. Investors including Eric Schmidt and Sean Parker back these new initiatives as Stability AI attempts to recover from past financial mismanagement.