图像由于其具有表达内容丰富，反映事物直观等优势，已经成为了大量信息的重

Because of its rich expressive content and intuitive reflection of things, images have become an important form of expression for a large amount of information and the main way for people to obtain information. With the advent of the information age, people pay more attention to the research of image classification technology that keeps pace with the times, so as to excavate more in-depth and important content from a large number of high-dimensional images. Due to the rapid development of deep learning in recent years, image classification algorithms based on deep learning have gradually replaced traditional manual annotated image classification algorithms with their superior performance. However, such methods usually require a large amount of balanced labeled data to train the network to achieve better classification results. Insufficient or unbalanced training data greatly reduces the model performance and greatly limits the practical application of these methods. Therefore, few-shot learning and imbalanced learning have become an important research hotspot in the field of image classification.

Because of its rich expression content and intuitive reflection of things, image has become an important form of expression of a large amount of information and the main way for people to obtain information. With the advent of the information age, people pay more attention to the image classification technology that keeps pace with the times, so as to mine more in-depth and important content from a large number of high-dimensional images. Due to the rapid development of deep learning in recent years, the image classification algorithm based on deep learning has gradually replaced the traditional manually labeled image classification algorithm with its superior performance. However, such methods usually need a large number of balanced labeled data to train the network in order to achieve better classification results. The lack or imbalance of training data greatly reduces the performance of the model, and also greatly limits the practical application of these methods. Therefore, small sample learning and unbalanced learning have become an important research hotspot in the field of image classification.

Image has become an important manifestation of a large amount of information and the main way for people to obtain information because of its rich expression content and intuitive reflection of things. With the advent of the information age, people pay more attention to the research of image classification technology that keeps pace with the times, so as to dig out more in-depth and important content from a large number of high-dimensional images. Due to the rapid development of deep learning in recent years, the image classification algorithm based on deep learning has gradually replaced the traditional manually labeled image classification algorithm with its superior performance. However, this kind of method usually needs a large amount of balanced labeled data to train the network to achieve better classification effect. Insufficient or unbalanced training data greatly reduces the performance of the model, and also greatly limits the practical application of these methods. Therefore, small sample learning and unbalanced learning have become an important research hotspot in the field of image classification.