Open AccessImage classification model based on GAT
Author(s) -
Bin Xu,
Sizhe Ding,
Yan Zhang
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1570/1/012082
Subject(s) - computer science , regularization (linguistics) , graph , data mining , pattern recognition (psychology) , artificial intelligence , artificial neural network , network model , image (mathematics) , machine learning , theoretical computer science
In the field of image classification, graph neural network (GNN) is a kind of structured data modeling architecture with larger functions. However, there are still some problems, such as low efficiency of updating nodes, fixed network parameters and the inability to effectively model the information features of some edges in the graph. In order to solve these problems, this paper introduces attention mechanism on the basis of GNN to improve it, proposes a graph attention network (GAT), establishes a double-layer GAT model, and uses regularization method in model iterative training to achieve image classification. The model is applied to three datasets for experiments. The experimental results show that the average classification accuracy of the proposed model is high and it has good application performance.