Our DICOM dataset - now open source!
High-quality data remains one of the biggest bottlenecks in developing reliable AI models. This is especially true in medical environments, where accuracy, robustness, and trust are essential because clinical decisions directly affect human lives. To help address this challenge, we built a large, carefully annotated micro-ultrasound imaging dataset.
Working closely with clinicians, our team has labeled gigabytes of imaging data to create a high-quality foundation for training and validating advanced AI models. This enables us to further improve our own segmentation, detection, and decision-support systems for prostate imaging and biopsy planning.
We are committed to open-sourcing more data as soon as we are able to expand our internal collection. We want to help others by giving them the opportunity to use a high-quality dataset This way, other researchers can use it to develop, compare, and validate their own methods. By making high-quality medical imaging data more accessible, we hope to accelerate innovation, encourage collaboration, and support the development of safer and more effective AI solutions in healthcare.
You can explore our dataset and use it freely under the specified license. DATASET
