Article: Intelligent judgement of calligraphy and painting image categories based on integrated classifier learning Journal: International Journal of Information and Communication Technology (IJICT) 2024 Vol.25 No.11 pp.1 - 20 Abstract: Traditional methods for categorising calligraphic painting images are often difficult to deal with diverse artistic styles and category imbalance. In order to solve these problems, this paper proposes an intelligent judgement method for calligraphy painting image categories. First, by comparing four base classifiers, Fisher, pseudo inverse, plain Bayes and C4.5 decision tree, the generalisation ability of the model in the face of diverse art styles is improved. Secondly, a dynamic training subset construction strategy (DWSCS) based on sample weights is introduced and MCACSAF is designed for calligraphy painting images. Experimental results show that compared with the traditional AdaBoost algorithm, MCACSAF improves the classification accuracy from 0.878 to 0.912, which is a 3.4% improvement, when using C4.5 decision tree as the base classifier. When dealing with minority class samples, the F1 score improves from 0.815 to 0.857, an improvement of 5.2%. Inderscience Publishers - linking academia, business and industry through research

Title: Intelligent judgement of calligraphy and painting image categories based on integrated classifier learning

Authors: Nannan Xu

Addresses: School of Fine Arts and Design, Suzhou University, Suzhou, 234000, China

Abstract: Traditional methods for categorising calligraphic painting images are often difficult to deal with diverse artistic styles and category imbalance. In order to solve these problems, this paper proposes an intelligent judgement method for calligraphy painting image categories. First, by comparing four base classifiers, Fisher, pseudo inverse, plain Bayes and C4.5 decision tree, the generalisation ability of the model in the face of diverse art styles is improved. Secondly, a dynamic training subset construction strategy (DWSCS) based on sample weights is introduced and MCACSAF is designed for calligraphy painting images. Experimental results show that compared with the traditional AdaBoost algorithm, MCACSAF improves the classification accuracy from 0.878 to 0.912, which is a 3.4% improvement, when using C4.5 decision tree as the base classifier. When dealing with minority class samples, the F1 score improves from 0.815 to 0.857, an improvement of 5.2%.

Keywords: image classification; AdaBoost; dynamic weights; sample construction; multiple classifier comparison; category imbalance.

DOI: 10.1504/IJICT.2024.143414

International Journal of Information and Communication Technology, 2024 Vol.25 No.11, pp.1 - 20

Received: 27 Oct 2024
Accepted: 22 Nov 2024
Published online: 18 Dec 2024 *

Title: Intelligent judgement of calligraphy and painting image categories based on integrated classifier learning

Keep up-to-date