We use technology based on diffusion probabilistic models to generate synthetic medical images. We train our models on real data. We can generate images from scratch or modify existing images.

Absolutely! We can train our models on your data. We can also fine-tune our models on your data. We can work with different data formats and data modalities. Reach out and we can discuss!

We can work with different data protection standards. There are different techniques depending on the privacy constraints. We can bring compute to the data with federated learning. Or, we can filter out images with high risk of re-identification.

We measure a series of image quality metrics during training. We tend to measure along four axes: realism, diversity, autenticity, and benchmarking.

We are working hard to make the API available as soon as possible. Reach out in our contact form if you want early access.