This project will primarily focus on the development of deep learning classification models involving geological and archaeological data at the Florida Museum of Natural History. In geology and archaeology, the classification of ancient rocks and ceramics regularly takes place through a process called petrographic thin-section analysis. In this process, small samples are taken out of artifacts and sliced into microscopically thin sheets. Cross-polarized light is then passed through the sheets to reveal unique physical structures. Differing mineral compositions will yield differing outputs depending on their specific composition, the angle, and the composition of the light passing through the sheets. These tests allow researchers to classify ceramic forms, giving access to a plethora of useful identifying characteristics within the molecular structures. While effective, researchers are required to manually go about the process, often taking days, even weeks, to reach a conclusion. This project will work to automate and optimize this process using deep learning for pattern-based recognition. This will specifically be done through the process of feature extraction and classification. The model will be able to take in an image, segment differing portions and characteristics within, then classify each segmentation by its defining qualities. The differing mineral compositions within our geological dataset will be used to train and test the deep learning algorithms.