Article Preview
TopIntroduction
The union of computational science and art history has brought a new era in both the research of art history and image analysis, providing novel discoveries concerning the subtle patterns within art and the objective lens through which it can be viewed (Doing et al., 2015). These recent advances have provided evidence that computerized image analysis can be effective in classifying and describing quantitative similarities in the work of different artists (Hurtut, 2011; Shamir & Tarakhovsky, 2012; van den Herik & Postma, 2000). Approaches to automatic analysis of visual art include searching and data organization methods such as the application of query by image content (QBIC) tools to retrieve images in art databases (Holt & Hartwick, 1994), as well as analyzing semantic links between paintings and captions (Tsai, 2007), allowing automatic assignment of captions to paintings the computer has not ‘seen’ before (Barnard et al., 2001).
Another application of computational methodology to art is the detection of fraudulent pieces of artwork (Kroner & Lattner, 1998; Taylor et al., 2007, 1999; Mureika et al., 2004; Lyu et al., 2004; Li et al., 2012), continued by the automatic identification of quantitative visual features that make the artistic style of the painter unique (Berezhnoy et al., 2007; Shamir, 2015).
Substantial research efforts have been invested in automatic classification of paintings. Computers have demonstrated ability to classify paintings by their creating artist (Kammerer et al., 2007; Keren, 2002; Shamir et al., 2010b; Johnson Jr et al., 2008; Shen, 2009; Cetinic & Grgic, 2013), emotion (Zhang et al., 2011, 2013), artistic movement (Zujovic et al., 2009; Culjak et al., 2011; Condorovici et al., 2013), or by keyˇ words associated with the paintings (Barnard et al., 2001; Lewis et al., 2004; Vrochidis et al., 2008). In addition to classification, computational methods can also estimate the similarities between different paintings (Garces et al., 2014), or similarities between painters by analyzing the visual content of their art (Shamir et al., 2010b; Shamir & Tarakhovsky, 2012; Shamir, 2012b; Kim & Kim, 2013; Wang & Takatsuka, 2012). Additional studies that revealed similarities in art detectable by computer analysis include algorithmic identification of the type of drawing tools and materials used to create a piece of art (Lake et al., 2004; Kammerer et al., 2007; Roussopoulos et al., 2010). These computational tools contribute a new perspective to the analysis of art that is less subject to human biases, and uses visual features that can be difficult to sense and quantify without using computer methods.