Often, binder and fibers have the same gray values in images obtained from computer tomography or FIB/SEM. Therefore, an automatic separation based on the gray value is not possible. However, it is possible to separate the different materials based on their shape in the image.

The FiberFind module can reconstruct 3D models obtained from segmented computer tomography or FIB/SEM images of fibers with binder. For this, the FiberFind module provides the possibility of using Artificial Intelligence (AI) separating binder from the fibers and identifying individual fibers. Since GeoDict 2021, the capability to create and parameterize these neural networks is available in the module GeoDict-AI. For more details see the GeoDict-AI User Guide.

It is the starting point to analyze physical properties on the segmented 3D-scans considering the different materials.

In this tutorial a carbon paper gas diffusion layer will be used as an example on how to separate the binder from the fibers and how to identify individual fibers and post-process the results.