Pulmonary Embolism

The initial use-case for the EVA_KI platform.

A pulmonary embolism is the blockage of a blood vessel in the lung. It is potentially life-threatening, but the prognosis is very good if detected early enough.

Structured Reporting

An unambiguous way to report clinical findings.

Up to now, medical findings have mainly been recorded in the form of free text. Smart Reporting has developed an innovative system that queries relevant medical decisions in a targeted manner, thereby achieving a far-reaching standardization.

Continual Learning

Training Machine Learning models from sequentially-arriving data.

Continual learning means the training of models with sequentially-received small amounts of data from different sources. In contrast to the usual workflow of training and evaluating Machine Learning methods, continuous learning algorithms can process successively incoming data, which may be available only for a short period of time.

Health economics and Ethics

Consideration of economic and ethical aspects.

Fundamental parts of the project are the development and verification of an economic business model for the widespread use of the platform, and an analysis of the ethical aspects.

Gullibility Analysis

Avoiding Automation Bias.

Within the project, a gullibility analysis will be performed to investigate to what extent both inexperienced and experienced radiologists rely on the predictions of neural networks.