Towards Ultra-Resolution Neural Style Transfer via Thumbnail Instance Normalization
Abstract
We present an extremely simple Ultra-Resolution Style Transfer framework, termed URST, to flexibly process arbitrary high-resolution images (e.g., 10000x10000 pixels) style transfer for the first time. Most of the existing state-of-the-art methods would fall short due to massive memory cost and small stroke size when processing ultra-high resolution images. URST completely avoids the memory problem caused by ultra-high resolution images by (1) dividing the image into small patches and (2) performing patch-wise style transfer with a novel Thumbnail Instance Normalization (TIN). Specifically, TIN can extract thumbnail features' normalization statistics and apply them to small patches, ensuring the style consistency among different patches. Overall, the URST framework has three merits compared to prior arts. (1) We divide input image into small patches and adopt TIN, successfully transferring image style with arbitrary high-resolution. (2) Experiments show that our URST surpasses existing SOTA methods on ultra-high resolution images benefiting from the effectiveness of the proposed stroke perceptual loss in enlarging the stroke size. (3) Our URST can be easily plugged into most existing style transfer methods and directly improve their performance even without training. Code is available at https://git.io/URST.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
5 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- MicroAST: Towards Super-fast Ultra-Resolution Arbitrary Style Transfer cites Arbitrary Style Transfer via Multi-Adaptation Network
- MicroAST: Towards Super-fast Ultra-Resolution Arbitrary Style Transfer cites Dynamic Instance Normalization for Arbitrary Style Transfer
- MicroAST: Towards Super-fast Ultra-Resolution Arbitrary Style Transfer cites A Closed-Form Solution to Universal Style Transfer
- Progressive Painterly Image Harmonization from Low-Level Styles to High-Level Styles cites Dynamic Instance Normalization for Arbitrary Style Transfer
- Arbitrary Style Transfer via Multi-Adaptation Network cites Dynamic Instance Normalization for Arbitrary Style Transfer
- Learning Dynamic Style Kernels for Artistic Style Transfer cites Dynamic Instance Normalization for Arbitrary Style Transfer
- Painterly Image Harmonization by Learning from Painterly Objects cites Dynamic Instance Normalization for Arbitrary Style Transfer
Cited by top-tier papers5
- MicroAST: Towards Super-fast Ultra-Resolution Arbitrary Style TransferZhizhong Wang, Lei Zhao, Zhiwen Zuo, Ailin Li et al.AAAI 2023 · 68 citations
- Progressive Painterly Image Harmonization from Low-Level Styles to High-Level StylesLi Niu, Yan Hong, Junyan Cao, Liqing ZhangAAAI 2024 · 3 citations
- Painterly Image Harmonization by Learning from Painterly ObjectsLi Niu, Junyan Cao, Yan Hong, Liqing ZhangAAAI 2024 · 2 citations
- SGSST: Scaling Gaussian Splatting Style TransferBruno Galerne, Jianling Wang, Lara Raad, Jean-Michel MorelCVPR 2025
- Learning Dynamic Style Kernels for Artistic Style TransferWenju Xu, Chengjiang Long, Yongwei NieCVPR 2023
Builds on6
- Photorealistic Style Transfer via Wavelet TransformsJaejun Yoo, Youngjung Uh, Sanghyuk Chun, Byeongkyu Kang et al.ICCV 2019 · 412 citations
- Dynamic Instance Normalization for Arbitrary Style TransferYongcheng Jing, Xiao Liu, Yukang Ding, Xinchao Wang et al.AAAI 2020 · 212 citations
- Switchable Whitening for Deep Representation LearningXingang Pan, Xiaohang Zhan, Jianping Shi, Xiaoou Tang et al.ICCV 2019 · 204 citations
- Arbitrary Style Transfer via Multi-Adaptation NetworkYingying Deng, Fan Tang, Weiming Dong, Wen Sun et al.ACM MM 2020 · 194 citations
- A Closed-Form Solution to Universal Style TransferMing Lu, Hao Zhao, Anbang Yao, Yurong Chen et al.ICCV 2019 · 91 citations
Related papers
- Collaborative Distillation for Ultra-Resolution Universal Style TransferHuan Wang, Yijun Li, Yuehai Wang, Haoji Hu et al.CVPR 2020
- AesUST: Towards Aesthetic-Enhanced Universal Style TransferZhizhong Wang, Zhanjie Zhang, Lei Zhao, Zhiwen Zuo et al.ACM MM 2022 · 70 citations
- InfinityGAN: Towards Infinite-Pixel Image SynthesisChieh Hubert Lin, Hsin-Ying Lee, Yen-Chi Cheng, Sergey Tulyakov et al.ICLR 2022 · 84 citations
- Dual-head Genre-instance Transformer Network for Arbitrary Style TransferMeichen Liu, Shuting He, Songnan Lin, Bihan WenACM MM 2024 · 3 citations
- Puff-Net: Efficient Style Transfer with Pure Content and Style Feature Fusion NetworkSizhe Zheng, Pan Gao, Peng Zhou, Jie QinCVPR 2024 · 20 citations
