The exploration and development of solutions in the field of Interpretable Artificial Intelligence, also known as Explainable AI (XAI), is one of the central research topics of VRVis, intending to lift the black-box character of artificial intelligence. This is particularly relevant in modern radiology, which increasingly relies on digital, AI-based tools to interpret and evaluate radiological images. After all, only innovative visualization methods such as those of VRVis that make the decision-making processes of AI systems transparent and comprehensible for medical experts allow for reliable diagnostics with human competence at its center.
For around two decades now, researchers of our Biomedical Image Informatics Group have been developing application-oriented and internationally patented solutions for direct use in everyday medical practice in close collaboration with AGFA Healthcare, a long-standing corporate partner and global provider of radiology solutions. The new VRVis visualization method, which has already received a great deal of attention in the scientific community and provides trust in the use of machine learning algorithms for accelerated, computer-aided diagnostic workflows, especially in such an important field like medicine, has now been honored with the eAward 2021 on October 18, 2021.
Statement of the eAward jury
The VRVis method, which is incidentally more reliable, faster and more trustworthy than any other method on the market, combines AI classification systems with decision visualization in a new way to make computer-aided diagnostic workflows more transparent. The method is generally applicable to all current AI algorithms for medical image classification and makes AI decisions not only visible and understandable but also more accurate and efficient.
Solutions that enable AI decisions to be reliable and transparent are of utmost importance, especially in medical diagnostics. The VRVis method makes the AI's "findings" - whether a radiological image, such as an X-ray, shows characteristics of a certain disease - such as pneumonia, a tumor or even Covid-19 - visible to the radiologist through accurate and reliable visualization that, unlike other methods, includes the medical context. For the continuous self-monitoring and optimization of the system, our researchers have also extended an existing classification network with another network that learns the important features for decision-making.
Since 2005, the Report Verlag annually presents the renowned business award eAward for the most innovative IT projects and companies. In addition to this year's award, VRVis has already received several awards and nominations in previous years, such as in 2020, when VRVis was nominated for two projects at once and was finally awarded the eAward in the category "Social Responsibility" for the Horizon2020 project ARCHES.
Vienna, October 18, 2021