Related Experiment Video
Updated: Dec 6, 2025

Applying Hyperspectral Reflectance Imaging to Investigate the Palettes and the Techniques of Painters
Published on: June 18, 2021
Dissecting landscape art history with information theory
Byunghwee Lee1, Min Kyung Seo2, Daniel Kim3
1Department of Physics, Korea Advanced Institute of Science and Technology, Daejeon 34141, Korea.
Quantitative analysis of 14,912 landscape paintings reveals that dominant compositional dissection ratios reveal artist and style evolution. This data-driven approach offers a new lens for art historical metanarratives.
Area of Science:
- Computational Art History
- Digital Humanities
- Quantitative Art Analysis
Background:
- Art history traditionally relies on qualitative methods, limiting large-scale analysis of creative evolution.
- A lack of quantitative data and methods hinders objective evaluation of art historical metanarratives.
- Understanding the macroscopic and microscopic dynamics of artistic creation requires new analytical approaches.
Purpose of the Study:
- To quantitatively analyze creative processes in landscape painting to verify and question existing art historical metanarratives.
- To systematically investigate the evolution of compositional proportions in landscape art.
- To explore the potential of data-driven methods in art history.
Main Methods:
- Utilized a benchmark dataset of 14,912 landscape paintings from the Western Renaissance to contemporary art.
- Applied an information-theoretic dissection method to analyze dominant horizontal and vertical compositional partition directions.
- Conducted frequency distribution analyses of compositions and network analyses of artists and style periods.
Main Results:
- Dominant dissection ratios serve as meaningful signatures for individual artist styles, creation dates, and art historical periods.
- Compositional evolution can be systematically tracked using these quantitative measures.
- Network analyses revealed three distinct supergroups of artists and styles, clustered meaningfully by time, while artist nationality proved problematic.
Conclusions:
- Quantitative analysis provides a systematic verification for art historical narratives and reveals underlying patterns in artistic evolution.
- Dominant compositional ratios offer a novel way to characterize and trace the development of artistic styles and individual oeuvres.
- Data-driven approaches, like this information-theoretic dissection, can complement traditional art historical scholarship by uncovering nonintuitive patterns and relationships.
Related Concept Videos
Gestalt Principles of Perception
Manipulation and Analysis
Thematic Layering in GIS
Methods of Obtaining Topography
Plotting of Topographic Maps
Levels of Use of a GIS

