PanoWorld: A Generative Spatial World Model for Consistent Whole-House Panorama Synthesis

TL;DR

PanoWorld is a novel generative model that synthesizes consistent, photorealistic whole-house VR panoramas by combining geometry-aware generation with spatial memory, enabling multi-room coherence.

Contribution

It introduces a new autoregressive spatial model that integrates floorplan-based geometry with a dynamic 3D memory for high-quality, multi-room panorama synthesis.

Findings

Achieves consistent multi-room panoramas with high-frequency detail.

Maintains cross-view spatial coherence and material consistency.

Outperforms existing methods in realism and layout accuracy.

Abstract

Generating a consistent whole-house VR tour from a floorplan and style reference requires both photorealistic panoramas and cross-view spatial coherence. Pure 2D generators produce appealing single panoramas but re-imagine geometry and materials when the viewpoint changes, whereas monolithic 3D generation becomes expensive and loses fine texture at multi-room scale. We introduce PanoWorld, a generative spatial world model that treats whole-house synthesis as autoregressive generation of node-based 360-degree panoramas, matching the discrete navigation used by real VR tour products. PanoWorld uses a floorplan-derived 3D shell as a global geometric proxy and a dynamic 3D Gaussian Splatting cache as renderable spatial memory. A feed-forward panoramic LRM designed for metric-scale multi-room 360-degree inputs lifts generated panoramas into local 3DGS updates, while Room-aware Group Attention suppresses cross-room feature interference. A topology-aware progressive caching strategy fuses these local updates without repeatedly reconstructing the full history. By decoupling shell-based geometry guidance from cache-rendered visual memory, PanoWorld preserves high-frequency 2D synthesis quality while improving cross-node layout and material consistency. The project link is https://jjrcn.github.io/PanoWorld-project-home/