Zircon classification from cathodoluminescence images using deep learning

<p>The dataset contains a total of 4282 original and processed zircon CL images collected from published papers and unpublished personal collections. Igneous, metamorphic, and hydrothermal zircon were classified and labeled manually. The dataset was used for a deep learning classification task which was submitted to &quot;Geoscience Frontiers&quot;, and was entitled as &quot;Zircon classification from cathodoluminescence images using deep learning&quot;.</p>