# Assessing aesthetics of generated abstract images using correlation structure

arXiv: https://arxiv.org/abs/2105.08635 · HTML: https://mlsanity.com/paper/2105.08635/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 7 of its 30 references that are in the map.
- First posted on arXiv 18 May 2021.

## Builds on (most cited first, 7 of 7)

- [Deep Image Prior](https://mlsanity.com/paper/1711.10925/) (2017)
- [Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Images](https://mlsanity.com/paper/1412.1897/) (2014)
- [Weight Agnostic Neural Networks](https://mlsanity.com/paper/1906.04358/) (2019)
- [Image Aesthetic Assessment: An Experimental Survey](https://mlsanity.com/paper/1610.00838/) (2016)
- [Will People Like Your Image? Learning the Aesthetic Space](https://mlsanity.com/paper/1611.05203/) (2016)
- [Neural Aesthetic Image Reviewer](https://mlsanity.com/paper/1802.10240/) (2018)
- [Aesthetics of Neural Network Art](https://mlsanity.com/paper/1903.05696/) (2019)

## Built on by (most cited first, 2 of 2)

- [Exploring to establish an appropriate model for image aesthetic assessment via CNN-based RSRL: An empirical study](https://mlsanity.com/paper/2106.03316/) (2021)
- [Exploring CNN-based models for image's aesthetic score prediction with using ensemble](https://mlsanity.com/paper/2210.05119/) (2022)

## Abstract (arXiv)

Can we generate abstract aesthetic images without bias from natural or human selected image corpi? Are aesthetic images singled out in their correlation functions? In this paper we give answers to these and more questions. We generate images using compositional pattern-producing networks with random weights and varying architecture. We demonstrate that even with the randomly selected weights the correlation functions remain largely determined by the network architecture. In a controlled experiment, human subjects picked aesthetic images out of a large dataset of all generated images. Statistical analysis reveals that the correlation function is indeed different for aesthetic images.