U-Net for Medical Image Segmentation

Alberto Raimondi
Sep 27, 2021

In this project, I helped researchers of Università Degli Studi di Milano Statale to implement a fully automatic image segmentation model working with 2D, 3D, and 4D data.
The implemented model was a U-net architecture for medical image segmentation based on the nnUNet project.
The results of this project are being used to research the possibility of automating craniofacial MRI and CT scans to improve the speed with which doctors can access meaningful patient analysis results.

Figure 1: An example of the resulting segmentation of a human abdomen.
Figure 1: An example of the resulting segmentation of a human abdomen.
DeepLearning Projects