Researchers have developed a foundation model for analyzing rock thin-section images, which may improve classification and understanding of rock structures. The model's accuracy and generalization are being studied, with various techniques such as data space augmentation and hybrid feature attention networks being explored. Additionally, related research is being conducted on the properties of rocks, including seismic anisotropy and pore structure, although some studies appear to be unrelated to the foundation model.