DiLiGenT-MV
收藏资源简介:
DiLiGenT-MV数据集是由浙江大学计算机科学与技术学院创建,用于评估多视角光度立体方法。该数据集包含5个具有复杂BRDF的物体,从20个不同视角拍摄,每个视角使用96个校准点光源照明。数据集提供了所有20个视角的法线图以及扫描的3D形状,旨在解决复杂材料和光照条件下的3D重建问题。创建过程中,使用了简单的数据捕捉设置,包括数字相机和LED灯,通过精确的校准和重建算法来处理困难。该数据集适用于评估多视角立体方法,特别是在处理复杂材料和光照变化方面。
The DiLiGenT-MV dataset was created by the College of Computer Science and Technology, Zhejiang University, to evaluate multi-view photometric stereo methods. It contains five objects with complex BRDFs, captured from 20 distinct viewpoints, where each viewpoint is illuminated by 96 calibrated point light sources. The dataset provides normal maps and scanned 3D shapes for all 20 viewpoints, aiming to address 3D reconstruction problems under complex material and lighting conditions. During its creation, a simple data capture setup was employed, including a digital camera and LED lights, with challenging tasks handled via precise calibration and reconstruction algorithms. This dataset is suitable for evaluating multi-view stereo methods, especially in handling complex materials and lighting variations.

- 1Multi-View Photometric Stereo: A Robust Solution and Benchmark Dataset for Spatially Varying Isotropic Materials浙江大学计算机科学与技术学院 · 2020年



