“In the field of computer vision, the research progress of visual basic models provides a new solution to solve the problems of traditional deep learning models relying on a large amount of annotated data and insufficient generalization ability.”