An art painting style explainable classifier grounded on logical and commonsense reasoning

Abstract This paper presents the art painting style explainable classifier named ANYXI. The is based on specialists’ knowledge of styles and human-understandable color traits. ANYXI overcomes principal flaws in few classifiers literature. In this way, we first propose, using studies, categorizations Baroque, Impressionism, Post-Impressionism. Second, carry out a human survey with aim validating appropriateness features used for understanding. Then, analyze discuss accuracy interpretability classifier. study ends an evaluation rationality explanations automatically generated by We enrich discussion empirical validation considering quantitative qualitative comparison versus other classifiers. reported results show how outstanding from point view while keeping high (comparable to non-explainable classifiers). Moreover, automated generations are endowed good level rationality.