Image classification method for two-way channel attention element learning

Abstract

The invention discloses an image classification method for two-way channel attention element learning. A relationship between a support set and a query set is learned in meta-learning through an attention model mechanism, so that the support set and the query set pay attention to each other. The advantages are mainly reflected in that: it is assumed that when visual features of image samples of the support set are mapped to category features, attention of a query set sample to the support set sample feature salient region is considered, meanwhile, the attention of the support set to the salient region of the query set is also considered, and the relationship between the support set and the query set is mined, so that the attention of the network to the salient region and the detail regionof the image feature is improved, the convergence speed is increased, and the performance of the meta-learning-based few-sample image classification method is further improved.

Description

technical field [0001] The invention relates to an image classification method. In particular, it concerns a method for image classification based on convolutional neural network-based bidirectional channel attention meta-learning. Background technique [0002] Deep learning technology is based on a large amount of data and large-scale training to simulate or realize human learning behavior to acquire new knowledge or skills. In reality, with the emergence of more application scenarios, we will inevitably face more data shortage problems. However, in the case of relatively few labeled data, neural networks are usually prone to overfitting, which makes the application and effect of deep learning difficult. are restricted. In contrast, humans have the ability to learn from small amounts of data. [0003] Therefore, how to enable machines to use learning experience to effectively learn from small samples like humans has become an important research direction. Few-shot learn...

AI Technical Summary

Problems solved by technology

At the same time, the relationship between the query set and the support set is ignored, and the image features are notable and the details are not enough.

Images