Image gallery
Analysis results
Download csv
Select images from the Upload tab to compare them here.
This app computes image statistics based on processing fluency theory, a framework from cognitive psychology that explains how easily information is processed affects aesthetic judgments and preferences. The metrics quantify aesthetic principles that facilitate fluent (easy, fast) cognitive processing of visual stimuli.
Visual processing fluency is influenced by multiple factors including contrast, complexity, symmetry, and typicality. Images that are easier to process (more fluent) are generally preferred and judged as more aesthetically pleasing, though context matters.
Measures overall luminance contrast using RMS (root mean square) contrast. Higher values indicate greater contrast between light and dark areas. This metric captures the overall luminance variation in the image.
Method: RMS contrast of luminanceQuantifies fractal-like patterns and repetitive structures across different scales. Higher values indicate more self-similar, fractal-like patterns. Uses quadtree decomposition to analyze similarity at different levels.
Method: Quadtree decompositionInverse measure of visual complexity based on compressed file size. Higher values indicate simpler images. Computed as 1 minus the ratio of compressed to maximum possible size. Simple images compress better than complex ones.
Method: 1 - (compressed / maximum)Measures vertical mirror symmetry by correlating the image with its horizontally flipped version. Values range from 0 (no symmetry) to 1 (perfect symmetry). Captures bilateral symmetry around the vertical axis.
Method: Pixel correlation after mirroringMeasures how typical each image is relative to the set of all uploaded images. Computed as the correlation of each image with the mean of all images. Values range from -1 (inversely typical) through 0 (not typical) to 1 (perfectly typical). Requires at least 2 images.
Method: Correlation with mean imageA: Up to 100 images at once. Note: Processing more than 10 images may take considerable time.
A: PNG, JPEG, BMP, and TIFF.
A: Typically 0.1-2 seconds per image per metric.
A: Yes! Please cite the package and relevant papers above.
© 2019-2026 Stefan Mayer (University of Tübingen)