Researchers of the Warsaw University of Technology (WUT) are working on a project of use of artificial intelligence for preventing premature births.

The solution prepared by the Warsaw scientists may support computer diagnostics and allow for more accurate prediction of spontaneous premature births.

The scientists of the Warsaw University of Technology created two scientific publications under the project.

The results of the works described in “Estimation of preterm birth markers with U-Net segmentation network” include “Reduction of the prediction error of spontaneous premature births” from 30% (manually by doctors) to 18% by the neural network.

In the work ‘Spontaneous preterm birth prediction using convolutional neural networks’, scientists demonstrated improvements in segmentation quality over the first publication and achieved better classification results.

Scientists are currently working on the website in the form of a web application. There they want to make available prepared models of the neural network. This should help gynecologists analyze ultrasound images and thus support the diagnosis of spontaneous preterm labor.