Title: Research on the classification method of artistic painting image style based on naive Bayesian

Authors: Shuo Liu

Addresses: Department of Design, Hubei Institute of Fine Arts, Wuhan City, Hubei Province, China

Abstract: Art painting image style classification is greatly affected by image noise, which leads to the low accuracy of art painting image style classification, and long classification time. This paper proposes an art painting image style classification method based on naive Bayes. Through Fourier transform, the style characteristics of art painting image are obtained. Sparse decomposition method was used to divide the art painting image into sparse components and other components to complete the art painting image denoising process. Using naive Bayes classification algorithm, the similarity coefficient of art painting image style samples was calculated, and the art painting image according to the art painting image style similarity rules was constructed. The similarity matrix of style samples realises the classification of art painting image styles. Through comparison, we can see that the accuracy rate of the proposed method is the highest, about 98.63%, and the shortest classification time is about 2 s.

Keywords: naive Bayesian algorithm; artistic painting; image style; classification algorithm.

DOI: 10.1504/IJICT.2022.126491

International Journal of Information and Communication Technology, 2022 Vol.21 No.4, pp.398 - 411

Received: 02 Nov 2020
Accepted: 23 Dec 2020
Published online: 26 Oct 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article