CVPR2025
Spherical Manifold Guided Diffusion Model for Panoramic Image Generation
Xiancheng Sun, Mai Xu, Shengxi Li, Senmao Ma, Xin Deng, Lai Jiang, Gang Shen
Abstract
Figure 1. A subjective comparison of zero-shot text-conditioned panoramic image generation between our SMGD model and the state-ofthe-art method PanFusion [53]. Our model demonstrates superior performance in generating images with superior overall natural content, enhanced subjective quality, and better alignment with the text description, while benefiting reduced model parameters.