Draw a line to answer the big questions of art
The algorithms showed that the compositional structure of Western landscape images changed “suspiciously” smoothly between MS 1500 and 2000, indicating a potential bias chosen by art curators or physicists from the Korea Institute of Advanced Science and Technology in art history literature (KAIST) and colleagues Data from the National Academy of Sciences (PNAS).
KAIST statistical physicist Hawoong Jeong, horizon placement of computer algorithms, etc. He has worked with statisticians, digital analysts, and art historians in Korea, Estonia, and the United States to determine whether they can help resolve long-standing questions about the design principles used in landscape paintings. key features.
“A key question among art historians is whether works of art have principles that take precedence over culture and time, and yes, how these principles have evolved over time,” Jeong explains. “We have developed an information-theoretical approach that can capture the composition ratio in landscape images, and we have seen that the preferred composition ratio develops systematically over time.”
From the Western Renaissance in the 1500s to the last era of modern art, digital versions of about 15,000 canonical landscapes were developed using a computer algorithm. The algorithm gradually divides the works of art into horizontal and vertical lines, depending on the amount of information in each subsequent section. It allows scientists to evaluate how artists and different art styles design landscape art, as well as how high or low the horizon of a landscape is, in terms of the placement of the most important components of a piece.
The scientists began to analyze the first two section lines defined by the algorithm in the drawings and found that they would be divided into four groups: first a horizontal line, then a second horizontal line (HH); first a horizontal line, then a second vertical line (HV); vertical, then horizontal line (VH); or a vertical and then a vertical line (VV) (see Figures 1 and 2). Then they looked at the categories over time.
Before the middle of the nineteenth century, HV was found to be the dominant compositional type, followed by HH, VH, and VV. In the middle of the nineteenth century, the rise of the HV composition style brought changes, along with a decline in popularity. The other two styles remained relatively stable.
Scientists have also studied how the horizon separating the sky from the land changes over time. At 16in In the 19th century, the dominant horizon of the painting was above the middle of the canvas, but gradually descended to the lower middle of the canvas until the 17th century, where it remained until the middle of the 19th century. After that, the horizon gradually began to rise again.
Interestingly, the algorithm showed that these findings were similar in cultures and art periods, and even in periods of diversity in art styles. This similarity may be a function of the bias in the database.
“In recent decades, art historians have preferred to argue that there is a great difference in the evolution of artistic expression, rather than presenting a relatively smoother consensus story in Western art,” he says. “This research is a reminder that existing large-scale data sets are seriously biased.”
Scholars then aim to expand their analysis to include more diverse works of art, as this special database was ultimately biased by the West and men. Future analyzes say that diagonal compositions in paintings should also be taken into account.
Reference: Byunghwee Lee, Min Kyung Seo, Daniel Kim, In-seob Shin, Maximilian Schich, Hawoong Jeong and Seung Kee Han, “Separating the History of Landscape Art with Information Theory”, October 27, 2020, Materials of the National Academy of Sciences.
DOI: 10.1073 / pnas.2011927117
This work was supported by the National Research Foundation of Korea (NRF).