PSGS: Text-driven Panorama Sliding Scene Generation via Gaussian Splatting

TL;DR

PSGS is a novel two-stage framework that generates high-fidelity panoramic 3D scenes from text, combining semantic reasoning and Gaussian Splatting to improve scene realism and consistency.

Contribution

It introduces a new two-layer optimization architecture and a panorama sliding mechanism for coherent 3D scene generation from text.

Findings

Outperforms existing panorama generation methods.

Produces more detailed and visually appealing 3D scenes.

Enhances scene coherence with depth and semantic losses.

Abstract

Generating realistic 3D scenes from text is crucial for immersive applications like VR, AR, and gaming. While text-driven approaches promise efficiency, existing methods suffer from limited 3D-text data and inconsistent multi-view stitching, resulting in overly simplistic scenes. To address this, we propose PSGS, a two-stage framework for high-fidelity panoramic scene generation. First, a novel two-layer optimization architecture generates semantically coherent panoramas: a layout reasoning layer parses text into structured spatial relationships, while a self-optimization layer refines visual details via iterative MLLM feedback. Second, our panorama sliding mechanism initializes globally consistent 3D Gaussian Splatting point clouds by strategically sampling overlapping perspectives. By incorporating depth and semantic coherence losses during training, we greatly improve the quality and detail fidelity of rendered scenes. Our experiments demonstrate that PSGS outperforms existing methods in panorama generation and produces more appealing 3D scenes, offering a robust solution for scalable immersive content creation.