Stanford Light Field

“概述斯坦福大学的计算机图形学实验室已经获得了几个用于计算机图形学和视觉研究的光场。我们正在向研究人员公开这些数据。光场可以用于学术和研究目的，但不能用于用于商业目的，未经我们许可，不得出现在待售产品中。如果您在出版物中使用这些模型，请注明斯坦福计算机图形实验室。数据格式单击上面的 Light Fields 链接以访问数据集. 每个光场都可以作为 PNG 格式的图像压缩存档。我们提供原始相机图像以及梯形失真校正图像。对于某些场景，校准信息（例如梯形失真校正的单应性和相机位置）作为文本文件；对于其他人，简单的校准目标在原始相机图像中可见。单击校准链接了解更多详细信息。

Overview: The Computer Graphics Lab at Stanford University has acquired several light fields for computer graphics and vision research. We are making these datasets publicly available to researchers.

These light fields may only be used for academic and research purposes, and commercial use is strictly prohibited. Without our prior written permission, they may not be included in any products for sale. If you use these datasets in your publications, please cite the Stanford Computer Graphics Lab.

Data Format: Click the Light Fields link above to access the datasets. Each light field is distributed as a compressed archive of images in PNG format. We provide both the raw camera images and the keystone distortion corrected images. For some scenes, calibration information (such as the homography for keystone distortion correction and camera positions) is provided as text files; for others, simple calibration targets are visible within the raw camera images. Click the Calibration link for more detailed information.

搜集汇总

背景与挑战

背景概述

Stanford Light Field是由斯坦福大学计算机图形学实验室于2008年发布的光场数据集，主要用于计算机图形学和视觉研究，如超分辨率和深度估计。数据集包含15.5GB的PNG格式图像，提供原始相机图像和校正图像，以及部分校准信息，但仅限于学术和非商业用途，使用时需引用相关论文。