Search AI Products and News

For a long time, efficiently generating high-quality, wide-angle 3D scenes from a single image has been a challenge faced by researchers. Traditional methods often rely on multi-view data or require time-consuming per-scene optimization, and they struggle with background quality and reconstruction of unseen areas. Existing technologies, when dealing with single-view 3D scene generation, often produce errors or distortions in occluded areas due to insufficient information, along with blurry backgrounds and difficulties in inferring the geometric structure of unseen regions. While regression-based models can synthesize new views in a feedforward manner, they struggle with complex...

最近，来自北京大学和鹏城实验室的研究团队推出了一款名为 HoloDreamer 的新型3D 场景生成框架，旨在通过用户的文本描述生成身临其境、完整的3D 场景。

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.