These 3D models were created by Sasha Arden, the Rachel and Jonathan Wilf & Andrew W. Mellon Fellow in Time-Based Media Conservation at The Conservation Center, Institute of Fine Arts, New York University, using a process called photogrammetry.
This process begins by taking a series of digital photographs of each object with a high-quality camera and a fixed lens. The camera must capture at least three views of each point on the object in order for it to be included in the 3D model, and the resulting image set for each object can contain hundreds of photographs. Specialized software called Metashape then interprets this image set to generate the first iteration of the 3D model. First, the program analyzes and compares the pixels in every image in the set to identify those points that belong to the object and those that are part of the background. This process yields a “sparse cloud” of the most likely points to comprise the final model, which is then refined over several additional steps to create a “dense cloud.” Based on those points (or based on depth maps derived from an analysis of the image set), a fine mesh of geometric shapes is then created to generate a “solid” form. The mesh often needs to be manually edited in Blender, an open-source 3D-modeling software, to correct the areas where Metashape was unable to accurately interpret the image set. Once those corrections are made, the edited mesh is imported back into Metashape where the model’s “texture” is created using the original image set. This final texture is a composite image drawn from colors of the original object mapped onto the mesh. The final 3D model is then uploaded to a viewing platform where environmental variables like background color and lighting can be adjusted so as to create a faithful representation of the original object.