Generative adversarial network prototype correction-based few-sample image classification method and system

The invention discloses a few-sample image classification method and system based on generative adversarial network prototype correction, and the method comprises the steps: dividing a data set into a training set, a verification set and a test set, and obtaining a feature embedded network and an adversarial network through the training of the training set; extracting sample features by using the feature embedded network, correcting the sample features to obtain a pre-corrected category prototype, inputting noise and the pre-corrected category prototype into the adversarial network, generating pseudo sample features, and obtaining a pseudo category prototype; fusing the pseudo category prototype and the pre-corrected category prototype to obtain a secondarily-corrected category prototype, and performing similarity measurement on training set samples to obtain classification loss of the samples for fine tuning of the feature embedded network; verifying the performance of the feature embedding network through t