SphericalDreamer: Generating Navigable Immersive 3D Worlds with Panorama Fusion

The generation of immersive and navigable 3D environments is increasingly prevalent with the growing adoption of virtual reality and 3D content. However, recent methods face a fundamental limitation: they cannot produce 3D worlds that simultaneously (i) are navigable over long-range spatial extents and (ii) cover the complete omnidirectional field of view ($360^\circ$ horizontally and $180^\circ$ vertically). To address this challenge, we introduce SphericalDreamer, a method for generating fully immersive and long-range 3D outdoor environments from textual prompts. Our approach is built on the generation of multiple panoramic images, which are subsequently lifted into 3D and fused together while maintaining visual and geometric consistency. SphericalDreamer produces highly detailed, fully immersive 3D environments, while substantially improving scale and navigability compared to prior approaches.

Imagine putting on a VR headset and exploring a rich, detailed outdoor world (forests, deserts, alien landscapes) generated automatically from a text description like "a misty bamboo forest". Creating such worlds by hand takes enormous time and expertise. AI can help, but existing methods have a frustrating trade-off: approaches that create fully surrounding, $360^\circ$ immersive environments can only be explored from a fixed point, while approaches that allow the user to walk long distances sacrifice the surrounding view, leaving blind spots above, below, and behind. SphericalDreamer resolves this trade-off. Our method generates multiple fully-surrounding panoramic "bubbles" from text, converts each into a 3D point cloud, and stitches them together into a single coherent world using a novel technique called Harmonic Blending, which ensures smooth visual and geometric transitions at the seams between bubbles. The result is 3D outdoor environments that are both navigable over long distances and fully immersive from every direction, generated automatically from a text prompt, at a fraction of the cost of manual creation.