A Single Image Classification Method Based on Generative Adversarial Network

Abstract

The invention relates to an image single classification method based on a generative confrontation network, comprising: constructing a generator in the generative confrontation network by using a densely connected block structure; constructing a discriminator in the generative confrontation network; inputting positive sample training data, and The gradient penalty algorithm is used to train the generative confrontation network; according to the classification effect of the model on the verification set during the training process, the network parameters are adjusted, and the early stopping strategy is used to find the optimal number of iterations for the classification of the model; after the model training is completed, use The discriminator in the generative confrontation network is tested on the test set data, and the classification effect of the model is judged by the classification recall rate CRI. The invention can automatically generate a negative sample set, and solves the problem that the manual construction of a negative sample data set in the existing single classification method easily causes over-fitting of the classifier.

Description

technical field [0001] The present invention relates to the technical field of image classification methods, in particular to an image single classification method based on a generative confrontation network. Background technique [0002] Image classification is one of the most basic research topics in the field of computer vision. With the development of deep learning, the supervised learning method makes the classification task of known image categories easier, more efficient and more accurate. In this process, there is sufficient training data to drive the end-to-end learning process and the semantics of the image can be clearly represented by nonlinear mapping. However, this ideal classification requires two prerequisites: one is sufficient training data, and the other is the corresponding labels of the training data. Therefore, the classification results will be limited to these known categories. [0003] The single classification problem expands the class of samples t...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a single classification method that can automatically generate a negative sample set, aiming to solve the problem that the artificially constructed negative sample data set in the existing single classification method is likely to cause overfitting of the classifier

Images