MedFusion Analytics

Try out the demo below! Input some data and the AI model will predict one of five possible outcomes (atelectasis, cardiomegaly, pleural effusion, lung opacity, or no finding). Or read through the documentation to learn more about the project.


MedFusion Analytics Demo

Example patients:

Fusion Inference

Inference results of the fusion model.

Try running the demo!

How it works

The fusion model derives its inputs from the base models below, taking its results and feeding it into a classifier to determine most likely results. Across all of our available data, the fusion model performed the best, returning the most predicted findings matching the real findings. To read more about how it works and how we developed it, you can reach the fusion model documentation.

Tabular Inference

Inference results of the tabular data only.

Try running the demo!

How it works

The tabular model takes the tabular data that has been inputed (temperature, heartrate, etc.) and attempts to classify into one of five possible options. In a proper setting, this would be tens to hundreds of inputs that could be taken it, improving the models ability to predict a class, however due to a limitation of data, we only have these inputs. To read more about the tabular model, see the tabular model documentation.