Cultural Heritage Synthetic Depth Dataset (CHSD_depth)

The Cultural Heritage Synthetic Depth Dataset (CHSD) is a comprehensive collection of depth maps corresponding to the RGB images of 3D cultural heritage objects. This dataset was developed to support the evaluation of deep learning-based methods for monocular depth estimation. The 3D objects featured in this dataset are derived from the Virtual Hampson Museum, encompassing 81 unique cultural heritage artifacts. Using sophisticated 3D modeling software, depth maps were generated by capturing each object from various angles with five virtual cameras positioned at -90, -45, 0, 45, and 90 degrees. The cameras rotated around each object in 1-degree steps, resulting in 1800 depth maps per object and a total of 145,800 depth maps.