Research on automatic classification method of artistic styles based on attention mechanism convolutional neural network

Abstract

Abstract In this article, we propose a convolutional neural network (CNN) based on an attention mechanism designed to automatically categorize artwork styles. The pyramid spatial attention (PSA) module is introduced into the traditional CNN, through which the weights of the feature maps are dynamically adjusted to enhance the model’s attention to the key features in the artworks. Experimental results show that the model proposed in this article significantly outperforms ResNet and DenseNet on the WikiArt dataset, reaching 91.52%, 90.49%, and 89.09% in accuracy, precision, and recall, respectively. This result validates the effectiveness of the PSA module in automatic classification of art styles. Future research directions may include incorporating more attention mechanisms to further improve the performance and generalization of the model.