Modeling, Quantifying, and Predicting Subjectivity of Image Aesthetics

TL;DR

This paper introduces a probabilistic framework for modeling and quantifying the subjectivity in image aesthetic preferences, utilizing deep learning to improve prediction accuracy and applying it to image recommendation systems.

Contribution

It proposes a novel unified probabilistic model based on subjective logic to quantify aesthetic subjectivity and develops a deep neural network approach for aesthetic prediction.

Findings

The framework effectively models aesthetic subjectivity using beta distributions.

Deep neural networks improve the accuracy of aesthetic and subjectivity prediction.

Application to image recommendation demonstrates practical benefits.

Abstract

Assessing image aesthetics is a challenging computer vision task. One reason is that aesthetic preference is highly subjective and may vary significantly among people for certain images. Thus, it is important to properly model and quantify such \textit{subjectivity}, but there has not been much effort to resolve this issue. In this paper, we propose a novel unified probabilistic framework that can model and quantify subjective aesthetic preference based on the subjective logic. In this framework, the rating distribution is modeled as a beta distribution, from which the probabilities of being definitely pleasing, being definitely unpleasing, and being uncertain can be obtained. We use the probability of being uncertain to define an intuitive metric of subjectivity. Furthermore, we present a method to learn deep neural networks for prediction of image aesthetics, which is shown to be effective in improving the performance of subjectivity prediction via experiments. We also present an application scenario where the framework is beneficial for aesthetics-based image recommendation.