Stanford's Fei-Fei Li team has released ZeroNVS, achieving scene-level full scene generation. They introduced a 3D perception diffusion model, ZeroNVS, to synthesize natural and diverse images. The researchers used metrics like LPIPS in their evaluations, achieving remarkable zero-shot performance. This achievement signifies that AI can generate 360-degree full scene images from a single image. The technology holds broad application prospects and will have a profound impact on fields such as virtual reality and game development.