Few-sample image classification method and system based on mutual neighbors

Abstract

The invention discloses a few-sample image classification method and system based on mutual neighbors, and the method comprises the steps: (1) carrying out forward derivation of the visual feature representation of a query image and a support image through a neural network model; (2) screening local feature descriptors related to tasks in the few-sample classification by using a mutual neighbor algorithm; (3) calculating the similarity between the query image and each category in the support image set by using the screened descriptors; (4) carrying out few-sample task division by using the image data set with the label and then training a neural network model; and (5) sorting according to the similarity between the classes in the support image set, and selecting the class with the maximum similarity as the class prediction of the image. Interference of a large number of local feature descriptors from a background on similarity calculation can be eliminated in a training process of few-sample image classification, so that a classification result is more robust.

Description

technical field [0001] The invention belongs to the technical field of image classification, and in particular relates to a small-sample image classification method and system based on mutual proximity. Background technique [0002] In recent years, object classification, as an important branch in the field of computer vision, has attracted the attention of researchers from industry and academia. Benefiting from the rapid development of deep learning technology, the supervised object classification task has made great progress, but at the same time, there are some limitations in this supervised training method, that is, in supervised classification, each class needs enough labeled training samples. However, in practical applications, there may not be enough training samples for each class, and labeling image data requires certain expertise and often takes a lot of manpower. [0003] The goal of few-sample image classification is to learn a machine learning model for image ...

AI Technical Summary

Problems solved by technology

Benefiting from the rapid development of deep learning technology, the supervised target classification task has made great progress, but at the same time, there are some limitations in the training method under this supervised condition, that is, in supervised classification, each class needs enough The labeled training samples of

However, in practical applications, there may not be enough training samples for each class, and labeling image data requires certain expertise and often takes a lot of manpower

Images