Automated Organ Segmentation in MRI

In this use case we realized a robust and accurate segmentation of the liver, the spleen, the kidneys and the heart in various Magnetic Resonance Imaging (MRI) sequences. The approach has proven to cope with even non-standardized image intensities and varying contrast in MRI.

All aspects of a typical AI service project have been realized in-house by our annotation and software experts.

Data

Input data:

120 3D MRI series of the torso

Annotation classes:

Liver
Spleen
Kidney left and right separately
Heart

A set of neighboring axial slices taken from an evaluation MRI series that was not contained within training. The automatically segmented organs are marked in color.

Results

We have achieved an average DICE coefficient across the sequences of 92%.