High-Quality 360-Degree Panoramic Image Generation Based on Masks and Stable Diffusion Model

360-degree panoramic images play a crucial role in multiple technological domains, particularly in virtual reality and augmented reality. However, current generation methodologies encounter significant limitations and are incapable of directly producing 360 -degree panoramic images through simple graphical elements such as masks. To address this gap, this study introduces a mask-based <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{3 6 0}$</tex>-degree panoramic image generation method. Utilizing a diffusion model that requires no additional training or fine-tuning, the proposed approach generates highquality RGB images during the inference phase. By incorporating masks and text prompts, the method efficiently produces complete, high-quality 360 -degree panoramic images, substantially improving generation efficiency. Experimental evaluations, employing qualitative metrics such as Absolute Category Rating and Perceptual Consistency Rating, systematically validated the effectiveness of the proposed method across various scenarios. The results demonstrate that the method not only generates highquality <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{3 6 0}$</tex>-degree panoramic images but also ensures semantic consistency and aligns closely with human visual perception.