Vision algorithms automate analysis of artistic painting styles

Researchers from the University of Girona and the Max Planck Institute have been able to categorize a painting’s artistic style using vision algorithms. The paintings are analyzed for many aspects at different levels:

• Low level: brush thickness and color composition
• Medium level: objects and scene composition
• High Level: historical context, including knowledge of artist and trends

According to the AlphaGalileo article, results from this research can be used in museums, including development of tools for viewing and analysis.