文献类型：期刊文献

英文题名：PSGS: TEXT-DRIVEN PANORAMA SLIDING SCENE GENERATION VIA GAUSSIAN SPLATTING

作者：Zhang, Xin[1]; Chen, Shen[2]; Zhou, Jiale[1]; Li, Lei[3]

机构：[1] East China University of Science and Technology, Shanghai, China; [2] Zhejiang University, Hangzhou, China; [3] Beijing Institute of Technology, Beijing, China

年份：2026

外文期刊名：arXiv

收录：EI(收录号：20260095381)

语种：英文

外文关键词：Iterative methods - Three dimensional computer graphics

摘要：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. Copyright ? 2026, The Authors. All rights reserved.