Abstract
Clustering artworks based on style has many potential real-world applicationslike art recommendations, style-based search and retrieval, and the study ofartistic style evolution in an artwork corpus. However, clustering artworksbased on style is largely an unaddressed problem. A few present methods forclustering artworks principally rely on generic image feature representationsderived from deep neural networks and do not specifically deal with theartistic style. In this paper, we introduce and deliberate over the notion ofstyle-based clustering of visual artworks. Our main objective is to exploreneural feature representations and architectures that can be used forstyle-based clustering and observe their impact and effectiveness. We developdifferent methods and assess their relative efficacy for style-based clusteringthrough qualitative and quantitative analysis by applying them to four artworkcorpora and four curated synthetically styled datasets. Our analysis providessome key novel insights on architectures, feature representations, andevaluation methods suitable for style-based clustering.