Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Arbitrary Style Transfer with Style-Attentional Networks

About

Arbitrary style transfer aims to synthesize a content image with the style of an image to create a third image that has never been seen before. Recent arbitrary style transfer algorithms find it challenging to balance the content structure and the style patterns. Moreover, simultaneously maintaining the global and local style patterns is difficult due to the patch-based mechanism. In this paper, we introduce a novel style-attentional network (SANet) that efficiently and flexibly integrates the local style patterns according to the semantic spatial distribution of the content image. A new identity loss function and multi-level feature embeddings enable our SANet and decoder to preserve the content structure as much as possible while enriching the style patterns. Experimental results demonstrate that our algorithm synthesizes stylized images in real-time that are higher in quality than those produced by the state-of-the-art algorithms.

Dae Young Park, Kwang Hee Lee• 2018

Related benchmarks

TaskDatasetResultRank
Style TransferMS-COCO and WikiArt
Execution Time (s)0.011
48
Image Style Transfer(test)
Average Inference Time (s)0.015
22
Style TransferMS-COCO (content) + WikiArt (style) (test)
Lcont4.72
17
Semantic Style Transferquadruple data (val)
SSL1.6583
11
Style TransferMS-COCO & WikiArt 512 x 512 images
Average Inference Time (s)0.037
11
Style TransferStyle Transfer (test)
Lc2.44
11
Artistic Style TransferCOCO and WikiArt (test)
LPIPS0.36
11
Arbitrary Image Style TransferWikiArt and Places365 (test)
Content Loss0.15
9
Disease ClassificationCamelyon17-WILDS 1.0 (test)
Test Accuracy0.9562
7
Disease ClassificationFitzpatrick17k 1.0 (test)
Test Accuracy81.33
7
Showing 10 of 18 rows

Other info

Follow for update