Enhancing Perceptual Attributes with Bayesian Style Generation
This provides a general framework for enhancing perceptual attributes in images, which could benefit applications in advertising, entertainment, or psychology, though it builds incrementally on existing style transfer and GAN techniques.
The paper tackles the problem of modifying images to enhance specific perceptual attributes like memorability or scariness, introducing a deep learning framework that combines style transfer and GANs to achieve state-of-the-art results on public benchmarks.
Deep learning has brought an unprecedented progress in computer vision and significant advances have been made in predicting subjective properties inherent to visual data (e.g., memorability, aesthetic quality, evoked emotions, etc.). Recently, some research works have even proposed deep learning approaches to modify images such as to appropriately alter these properties. Following this research line, this paper introduces a novel deep learning framework for synthesizing images in order to enhance a predefined perceptual attribute. Our approach takes as input a natural image and exploits recent models for deep style transfer and generative adversarial networks to change its style in order to modify a specific high-level attribute. Differently from previous works focusing on enhancing a specific property of a visual content, we propose a general framework and demonstrate its effectiveness in two use cases, i.e. increasing image memorability and generating scary pictures. We evaluate the proposed approach on publicly available benchmarks, demonstrating its advantages over state of the art methods.