Artistic image classification method based on convolutional neural network

Abstract

The invention discloses an artistic image classification method based on a convolutional neural network. The artistic image classification method comprises the following steps: obtaining each type ofartistic image, cutting the image, and constructing a training set and a verification set; then constructing a convolutional neural network by using hole convolution, a DKSE module and depth separableconvolution; and finally, inputting the training set into the built convolutional neural network to train the convolutional neural network to obtain a network model capable of being used for art image classification. Compared with a traditional method, the artistic image classification method has the advantages that the extracted overall features and local detail features of the art images are more comprehensive; the artistic image classification method can be suitable for classification research of various art images; the model universality is high, and the defects of existing classificationretrieval research of various art images are overcome; and the convolutional neural network designed in the invention can reduce professional requirements for art image classification personnel, andavoids complex feature extraction and data annotation work in a traditional classification algorithm; compared with other convolutional neural network structures, and the convolutional neural networkstructure is simple, and the classification accuracy is high.

Description

technical field [0001] The invention belongs to the technical field of image processing and recognition, and in particular relates to an art image classification method based on a convolutional neural network. Background technique [0002] The development of Internet digital media technology has promoted the sharing and dissemination of natural art images, but with the rapid increase in the number of art images, effective classification and retrieval of them is an urgent problem to be solved. In the face of massive artistic image data, traditional manual feature extraction methods may have problems such as labeling errors and insufficient objective labeling, and the professional requirements for artistic image classification personnel are relatively high. The existing research work on artistic image management is mainly based on the classification of Chinese painting themes and expressive techniques, the classification of painters according to their creative styles, and the ...

AI Technical Summary

Problems solved by technology

[0003] In the article "Research on Chinese Painting Classification Method Based on Expression Techniques, Chinese Journal of Computers. 2017 No. 12, 2871-2882", Gao Feng et al. proposed to use scale-invariant feature transformation feature detector and edge detection to obtain the key area of ​​the image. The description of the regional characteristics and the internal differences of the neighborhood, using the cascade classifier to analyze the difference in the expression techniques of the meticulous painting and the freehand painting, but the obtained local key regional features are not enough to represent the key features of the overall artistic image

According to the texture, color, shape and other characteristics of Chinese painting, Wang Zheng et al. used in the article "Classification and Prediction of Chinese Painting with Supervised Heterogeneous Sparse Feature Selection, Journal of Computer-Aided Design & Computer Graphics. 2013 No. 12, 1848-1855". The traditional method performs supervised heterogeneous sparse feature extraction, but the feature is only 96-dimensional, which is not enough to describe the overall characteristics of Chinese painting

Yao et al proposed in the article "Characterizing elegance of curves computationally for distinguishing Morrisseau paintings and theimitations, Proceedings of the 16th IEEE International Conference on Image Processing. The difference between paintings, but this method is only for the true and false identification of Norval Morrisseau art paintings, and the model has poor versatility

Each type of artistic image has certain similarities in texture, color, line and other characteristics. Using traditional methods to extract artistic image features cannot fully distinguish the style characteristics of each type of artistic image

Images