We present a method for the unsupervised segmentation of electron microscopy images, which are powerful descriptors of materials and chemical systems. Images are oversegmented into overlapping chips, and similarity graphs are generated from embeddings extracted from a domainx2010pretrained convolutional neural network (CNN). The Louvain method for community detection is then applied to perform segmentation. The graph representation provides an intuitive way of presenting the relationship between chips and communities. We demonstrate our method to track irradiationx2010induced amorphous fronts in thin films used for catalysis and electronics. This method has potential for "onx2010thex2010fly" segmentation to guide emerging automated electron microscopes.