QuantArt: Quantizing Image Style Transfer Towards High Visual Fidelity
Preprint
Abstract
The mechanism of existing style transfer algorithms is by minimizing a hybrid loss function to push the generated image toward high similarities in both content and style. However, this type of approach cannot guarantee visual fidelity, i.e., the generated artworks should be indistinguishable from real ones. In this paper, we devise a new style transfer framework called QuantArt for high visual-fidelity stylization. QuantArt pushes the latent representation of the generated artwork toward the centroids of the real artwork distribution with vector quantization. By fusing the quantized and continuous latent representations, QuantArt allows flexible control over the generated artworks in terms of content preservation, style similarity, and visual fidelity. Experiments on various style transfer settings show that our QuantArt framework achieves significantly higher visual fidelity compared with the existing style transfer methods.Keywords
All Related Versions
This publication has 46 references indexed in Scilit:
- ArtFID: Quantitative Evaluation of Neural Style Transfer2022
- Joint Bilateral Learning for Real-Time Universal Photorealistic Style Transfer2020
- Domain-Specific Mappings for Generative Adversarial Style Transfer2020
- Contrastive Learning for Unpaired Image-to-Image Translation2020
- A Closed-Form Solution to Photorealistic Image Stylization2018
- Diverse Image-to-Image Translation via Disentangled Representations2018
- Image Style Transfer Using Convolutional Neural Networks2016
- Perceptual Losses for Real-Time Style Transfer and Super-Resolution2016
- Split and Match: Example-Based Adaptive Patch Sampling for Unsupervised Style Transfer2016
- Microsoft COCO: Common Objects in Context2014