Stability AI has unveiled the Stable Virtual Camera, an AI model capable of converting 2D images into 3D videos by adding desired perspectives and adjusting image depth.
The standout feature of the Stable Virtual Camera is its input, which consists of 2D images that can range from 1 to 32 images. Another input is the direction of the desired video camera in the output, supporting 14 directions such as 360-degree rotation, panning, dolly zoom, lateral movement, camera rotation, and more.
In the initial version released, the resulting videos may have low resolution quality in some cases, especially if the input images are detailed, such as people, animals, water, or highly detailed scenes, which may lead to distorted results.
Stable Virtual Camera is available for download under a non-commercial agreement through Hugging Face and GitHub, with the model still in the research preview stage.
TLDR: Stability AI introduces the Stable Virtual Camera, a model that transforms 2D images into immersive 3D videos with realistic depth and perspective without the need for complex reconstruction or scene-specific optimization.
Leave a Comment