遇见数据集

cyberagent/mvscps

收藏
Hugging Face2025-09-29 更新2025-09-13 收录
官方服务:

资源简介:

MVSCPS数据集包含用于多视角三维重建和反向渲染任务的多视角One-Light-at-a-Time (OLAT)图像。它由6个场景组成,每个场景都包含了原始图像、JPEG图像、前景掩码、相机参数和场景概览图等文件和文件夹。该数据集适用于评估不需要光照立体线索的神经多视角自校准光度立体方法。

The MVSCPS dataset consists of multi-view One-Light-at-a-Time (OLAT) images for multi-view 3D reconstruction and inverse rendering tasks. It includes 6 scenes with directories containing raw images, JPEG images, foreground masks, camera parameters, and an overview image of the scene. This dataset is suitable for evaluating the neural multi-view self-calibrated photometric stereo method without photometric stereo cues.

提供机构:
cyberagent
二维码
社区交流群
二维码
科研交流群
商业服务