The Minimum Semantic Content (MSC) Dataset: A Large, Balanced Resource for Computational Aesthetics Research.

Area of Science:

• Empirical Aesthetics
• Computational Neuroscience
• Computer Vision

Background:

• Image databases are crucial for empirical aesthetics, but often conflate visual features with semantic content.
• Existing databases are frequently imbalanced, overrepresenting highly appreciated images, which biases research.
• This limits the ability to isolate perceptual influences on aesthetic judgments.

Purpose of the Study:

• To introduce the Minimum Semantic Content (MSC) database, a novel resource for computational aesthetics.
• To address limitations of existing databases by minimizing semantic and cognitive confounds.
• To facilitate research on the relationship between visual features and aesthetic appreciation.

Main Methods:

• Developed a large database (10,426 images) of natural scenes with reduced and homogenized semantic content.
• Collected aesthetic ratings from approximately 10,000 participants via crowdsourcing (100 ratings per image).
• Generated 'beautified' and 'uglified' image versions to ensure uniform aesthetic spectrum coverage.

Main Results:

• The MSC database minimizes cognitive and emotional confounds in aesthetic judgment studies.
• The curated image set promotes uniform coverage across the aesthetic spectrum, mitigating bias.
• Validation demonstrated improved robustness and performance in computational models.

Conclusions:

• The MSC database offers a valuable, systematically curated resource for empirical aesthetics research.
• It enables researchers to investigate perceptual features' impact on aesthetic judgments with reduced confounds.
• This resource advances the development of more robust computational models of aesthetics.