Semi-supervised image classification method based on generative adversarial network

Abstract

The invention discloses a semi-supervised image classification method based on a generative adversarial network, which is used for solving the technical problem of low classification precision causedby lack of identification degree and diversity of features extracted by the network in the prior art, and comprises the following steps of: obtaining a training sample set and a test sample set; building a generative adversarial network model; performing iterative training on the generative adversarial network model; and obtaining a semi-supervised image classification result. According to the method, the features extracted at different levels are fused through a feature pyramid network, the classification capacity of the model is improved through the gaming process of the generative adversarial network, the recognition degree and diversity of the features are improved, the inter-class features of the classified images can be represented more richly, the image classification precision is improved, and a very good image classification effect can be obtained from a sample set only containing a small number of accurately labeled samples, and the method can be used for target classification in an actual scene.

Description

technical field [0001] The invention belongs to the technical field of image processing and relates to an image classification method, in particular to a semi-supervised image classification method based on a generative confrontation network, which can be used in the fields of target detection, target classification and the like. Background technique [0002] Image classification is one of the core problems in the field of computer vision. The task is to assign a label to an image from a given set of categories. The label is always from a predefined set of possible categories. At present, image classification includes methods based on distance measures, methods based on texture features, and methods based on deep learning. In recent years, due to the rapid development of deep learning, significant progress has been made in the field of image classification. At present, better classification accuracy can be obtained in some real-world image classification problems. Image clas...

AI Technical Summary

Problems solved by technology

However, this method only uses the high-level semantic features extracted by the deep network to perform classification tasks, and the image features are relatively simple. In order to obtain high classification accuracy, a large number of labeled samples are required to train the model, and it is difficult to perform in the absence of labeled training samples. play a role in the problem, greatly limiting its scope of application

Images