# Real-time Universal Style Transfer on High-resolution Images via Zero-channel Pruning

arXiv: https://arxiv.org/abs/2006.09029 · HTML: https://mlsanity.com/paper/2006.09029/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 Jun 2020.

## Built on by (most cited first, 9 of 9)

- [StyTr2: Image Style Transfer with Transformers](https://mlsanity.com/paper/2105.14576/) (2021)
- [ArtFlow: Unbiased Image Style Transfer via Reversible Neural Flows](https://mlsanity.com/paper/2103.16877/) (2021)
- [SGSST: Scaling Gaussian Splatting StyleTransfer](https://mlsanity.com/paper/2412.03371/) (2024)
- [Towards Ultra-Resolution Neural Style Transfer via Thumbnail Instance Normalization](https://mlsanity.com/paper/2103.11784/) (2021)
- [S2WAT: Image Style Transfer via Hierarchical Vision Transformer using Strips Window Attention](https://mlsanity.com/paper/2210.12381/) (2022)
- [Learning to Evaluate the Artness of AI-generated Images](https://mlsanity.com/paper/2305.04923/) (2023)
- [ArtFusion: Controllable Arbitrary Style Transfer using Dual Conditional Latent Diffusion Models](https://mlsanity.com/paper/2306.09330/) (2023)
- [Scaling Painting Style Transfer](https://mlsanity.com/paper/2212.13459/) (2022)
- [Edge Enhanced Image Style Transfer via Transformers](https://mlsanity.com/paper/2301.00592/) (2023)

## Abstract (arXiv)

Extracting effective deep features to represent content and style information is the key to universal style transfer. Most existing algorithms use VGG19 as the feature extractor, which incurs a high computational cost and impedes real-time style transfer on high-resolution images. In this work, we propose a lightweight alternative architecture - ArtNet, which is based on GoogLeNet, and later pruned by a novel channel pruning method named Zero-channel Pruning specially designed for style transfer approaches. Besides, we propose a theoretically sound sandwich swap transform (S2) module to transfer deep features, which can create a pleasing holistic appearance and good local textures with an improved content preservation ability. By using ArtNet and S2, our method is 2.3 to 107.4 times faster than state-of-the-art approaches. The comprehensive experiments demonstrate that ArtNet can achieve universal, real-time, and high-quality style transfer on high-resolution images simultaneously, (68.03 FPS on 512 times 512 images).