Abstract:Image style transfer has become a key technique in modern photo-editing applications. Although significant progress has been made to blend content from one image with style from another image, the synthesized image may have a hallucinatory effect when the texture from the style image is rich when processing high-resolution image style transfer tasks. In this paper, we propose a novel attention mechanism, named compositional attention, to design a compositional transformer-based autoencoder (CTA) to solve this above-mentioned issue. With the support from this module, our model is capable of generating high-quality images when transferring from texture-riched style images to content images with semantics. Additionally, we embed region-based consistency terms in our loss function for ensuring internal structure semantic preservation in our synthesized image. Moreover, information theory-based CTA is discussed and Kullback–Leibler divergence loss is introduced to preserve more brightness information for photo-realistic style transfer. Extensive experimental results based on three benchmark datasets, namely Churches, Flickr Landscapes, and Flickr Faces HQ, confirmed excellent performance when compared to several state-of-the-art methods. Based on a user study assessment, the majority number of users, ranging from 61% to 66%, gave high scores on the transfer effects of our method compared to 9% users who supported the second best method. Further, for the questions of realism and style transfer quality, we achieved the best score, i.e., an average of 4.5 out of 5 compared to other style transfer methods.

Enhance Image Style Transfer with Depth Spatial Modulation.

Learning Structure-Aware Transformations for Arbitrary Image Style Transfer

Diverse Image Style Transfer Via Invertible Cross-Space Mapping

Correlation-based and Content-Enhanced Network for Video Style Transfer

Style Permutation for Diversified Arbitrary Style Transfer

Artistic Style Transfer with Internal-external Learning and Contrastive Learning

GLStyleNet: Exquisite Style Transfer Combining Global and Local Pyramid Features

Optimal Transport of Deep Feature for Image Style Transfer

Diversified Patch-based Style Transfer with Shifted Style Normalization

Preserving Structural Consistency in Arbitrary Artist and Artwork Style Transfer

Structure Guided Photorealistic Style Transfer.

Image Neural Style Transfer with Preserving the Salient Regions.

Structure-Preserving Neural Style Transfer

Improving the Latent Space of Image Style Transfer

Improved Style Transfer by Respecting Inter-layer Correlations

Deep Content Guidance Network for Arbitrary Style Transfer

Photographic style transfer

A Compositional Transformer Based Autoencoder for Image Style Transfer

A non-definitive auto-transfer mechanism for arbitrary style transfers

Multimodal Style Transfer Via Graph Cuts

Multi-Source Style Transfer Via Style Disentanglement Network