HistokatFusion

HistokatFusion aggregates tumor information for
AI development and individualized therapy.

HistokatFusion

HistokatFusion aggregates tumor information for
AI development and individualized therapy.

HistokatFusion

HistokatFusion aggregates tumor information for
AI development and individualized therapy.

Improving tumor diagnostics by providing reliable ground truth automatically.

In computational pathology an expert has to annotate thousands of tissue sections for an AI algorithm to learn and repeat his pattern recognition task automatically.

Creating manual annotations is a time-consuming and therefore expensive task that also lacks accuracy considering the interobserver variability shown between experts.

HistokatFusion is able to provide the ground truth automatically. This helps AI companies to create more accurate algorithms in less time while saving the money for an expert.


Automatic ground truth generation is achieved by accurately fusing serial sections, which allows for multiple, differently stained versions of similar tissue. The additional sections contain the molecular ground truth that can be used to train an AI algorithm.

Meet the team behind HistokatFusion.

Johannes Lotz

Johannes Lotz

Contact
Nick Weiss

Nick Weiss

Contact
Daniel Budelmann

Daniel Budelmann

Contact

HistokatFusion has been validated on thousands of differently stained image pairs.

Its potential for the fast development of accurate AI solutions has been shown with our clinical research partners.1

It has been proven to be fast, robust and accurate in an international challenge.2

Combine multiple stainings

Accurate 2

Robust in routine use 2

Extremely fast 2

Run anywhere

Support

Give it a try with a Free Trial Account

1 https://www.nature.com/articles/s41598-018-37257-4
2 HistokatFusion was ranked as the final #1 in the Automatic Non-rigid Histological Image Registration (ANHIR) challenge at the IEEE International Symposium on Biomedical Imaging (ISBI) in 2019. Winning in terms of accuracy, robustness and speed.

Winner

Winner