Stanford-ORB
收藏资源简介:
Stanford-ORB是一个真实世界的3D物体逆渲染基准数据集,由斯坦福大学创建。该数据集包含14个真实世界物体,在7个不同的自然场景中捕获,每个物体配有地面实况3D扫描、多视角图像和环境光照。数据集旨在通过一系列指标评估不同逆渲染方法的性能,包括形状估计、自然场景中的重光照和新视角合成。Stanford-ORB的应用领域包括3D内容生成和机器人操作,旨在解决复杂真实世界环境中的逆渲染问题。
Stanford-ORB is a real-world 3D object inverse rendering benchmark dataset created by Stanford University. This dataset includes 14 real-world objects captured across 7 distinct natural scenes, with each object accompanied by ground-truth 3D scans, multi-view images, and environmental lighting data. The dataset is designed to evaluate the performance of diverse inverse rendering methods via a range of metrics, encompassing shape estimation, relighting in natural scenes, and novel view synthesis. Application domains of Stanford-ORB cover 3D content generation and robotic manipulation, aiming to address inverse rendering problems in complex real-world environments.

- 1Stanford-ORB: A Real-World 3D Object Inverse Rendering Benchmark斯坦福大学 · 2024年



