TesticulUS

A multicenter testicular ultrasound resource for segmentation and parenchymal-inhomogeneity classification research.

Collection	Annotations	Availability
Two Italian clinical centers	Segmentation mask for every image	Via Ditto only
First 860 images	Parenchymal-inhomogeneity classification label	Via Ditto only

Overview

TesticulUS combines testicular ultrasound images acquired at two Italian clinical centers. Every image is paired with a pixel-wise segmentation mask. The first 860 images additionally provide labels for parenchymal inhomogeneity, supporting both segmentation and classification research.

The preview panel shows representative examples with translucent teal mask overlays. It is provided solely to illustrate the annotation format.

Access

The complete dataset is not hosted on Hugging Face. Request and download it through the Ditto platform.

License

TesticulUS is released under CC BY-NC-SA 4.0.

Citation

For segmentation use, cite:

@inproceedings{morelli2026new,
  title={A New Multicenter Testicular US Dataset and a Lightweight Cond-UNet for Generalization in US Segmentation},
  author={Morelli, Nicola and Marchesini, Kevin and Santi, Daniele and Grana, Costantino and Bolelli, Federico and others},
  booktitle={Proceedings of the British Machine Vision Conference},
  year={2026}
}

For classification use, cite:

@inproceedings{morelli2025enhancing,
  title={Enhancing testicular ultrasound image classification through synthetic data and pretraining strategies},
  author={Morelli, Nicola and Marchesini, Kevin and Lumetti, Luca and Santi, Daniele and Grana, Costantino and Bolelli, Federico},
  booktitle={International Conference on Image Analysis and Processing},
  pages={26--39},
  year={2025},
  organization={Springer}
}