CVMay 10, 2018

Arbitrary Style Transfer with Deep Feature Reshuffle

arXiv:1805.04103v4201 citations
Originality Incremental advance
AI Analysis

This addresses style transfer issues for image processing applications, offering an incremental improvement over prior neural methods.

The paper tackles arbitrary style transfer by reshuffling deep features of the style image, proving that this connects global and local style losses and avoids distortions and artifacts compared to existing methods, producing better quality results.

This paper introduces a novel method by reshuffling deep features (i.e., permuting the spacial locations of a feature map) of the style image for arbitrary style transfer. We theoretically prove that our new style loss based on reshuffle connects both global and local style losses respectively used by most parametric and non-parametric neural style transfer methods. This simple idea can effectively address the challenging issues in existing style transfer methods. On one hand, it can avoid distortions in local style patterns, and allow semantic-level transfer, compared with neural parametric methods. On the other hand, it can preserve globally similar appearance to the style image, and avoid wash-out artifacts, compared with neural non-parametric methods. Based on the proposed loss, we also present a progressive feature-domain optimization approach. The experiments show that our method is widely applicable to various styles, and produces better quality than existing methods.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes