Modeling, Quantifying, and Predicting Subjectivity of Image Aesthetics
This work addresses the problem of handling subjective variability in aesthetic preferences for computer vision applications, offering a method to quantify uncertainty for tasks like image recommendation.
The paper tackles the challenge of modeling and quantifying subjectivity in image aesthetics assessment by proposing a unified probabilistic framework based on subjective logic, which uses a beta distribution to derive probabilities for pleasing, unpleasing, and uncertain ratings, and shows effectiveness in improving subjectivity prediction through deep neural networks.
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.