Abstract: In a dynamic open environment, intelligent visual perception systems need to constantly learn new categories of visual concepts. However, due to the high cost of data acquisition and labeling, new categories often only have a small number of labeling samples available. Few-shot class-incremental learning (FSCIL) aims to solve the problem by enabling models to learn efficiently with limited samples of new categories while avoiding forgetting old ones. Firstly, an overview of FSCIL technology is given, highlighting its application potential in the field of intelligent perception. Secondly, according to different visual tasks, the mainstream methods and the performance are systematically analyzed. Finally, the future development trend of the field is prospected, including the development of theory and technology and the expansion of problem setting.