LUCES-MV
收藏资源简介:
LUCES-MV是由东芝欧洲剑桥团队创建的多视角近场点光源光度立体数据集,旨在解决现有光度立体数据集在复杂物体和近场光照条件下的不足。该数据集包含15个具有多样材质和形状的对象,每个对象在15个LED光源的不同光照条件下进行拍摄,提供了高分辨率图像(2080×1552像素)和详细的相机与光源校准信息。数据集的创建过程包括精确的3D扫描和校准,确保了数据的高质量和可重复性。LUCES-MV主要应用于光度立体技术中的3D形状重建,特别是在复杂材质和形状的物体上,推动了该领域的研究进展。
LUCES-MV is a multi-view near-field point-light photometric stereo dataset developed by the Toshiba Europe Cambridge team, aiming to address the limitations of existing photometric stereo datasets in scenarios involving complex objects and near-field illumination. This dataset includes 15 objects with diverse materials and shapes. Each object was captured under varying lighting conditions using 15 LED light sources, providing high-resolution images (2080×1552 pixels) as well as detailed camera and light source calibration information. The dataset was constructed via precise 3D scanning and calibration, ensuring high data quality and reproducibility. LUCES-MV is primarily applied to 3D shape reconstruction in photometric stereo technology, especially for objects with complex materials and shapes, thus advancing research progress in this field.

- 1LUCES-MV: A Multi-View Dataset for Near-Field Point Light Source Photometric Stereo东芝欧洲剑桥 · 2024年



