“…In recent years the computer graphics community has released multiple databases that have been extremely useful for practitioners in the field, raising the bar for new algorithms in terms of scalability and ability to handle a variety of inputs with different complexity, from easy ones to highly challenging. To make a practical example, the Thingi10K dataset has quickly become a popular means to empirically validate the robustness of surface mesh generation and processing algorithms Pietroni et al 2021] and some of its models are so pathological that being able to process them is an achievement by itself, with authors reporting both running times and memory consumption (see e.g. Fig.…”