What do mozzarella and bamboo have in common? – They are both used for large-scale datasets.
Images of the VoDaSuRe dataset
Monday 22 June 2026
The 3D Imaging Center has made two X-ray computed tomography datasets available: MozzaVID a volumetric image dataset using mozzarella, and VoDaSuRe a large-scale dataset tailored for real-world volumetric super resolution. Both were made available as training data for deep learning-based object detection. Having large-scale datasets is essential in training these models.
MozzaVID is a large, clean and versatile volumetric classification dataset. It enables the classification of 25 cheese types, 149 cheese samples, and three different resolutions. Totaling in three datasets containing 591 to 37,824 images. Mozzarella is unique due to the complex and disordered nature of food structures, which is a challenge for a robust structural analysis model.
VoDaSuRe is a comprising 16 samples, spanning diverse microstructures and more than 194 giga-voxels of 3D data. To date it is the largest volumetric dataset to include both synthetically down sampled and physically acquired, low resolution data for all scans.
To learn more about these datasets click here: MozzaVID and VoDaSuRe
