• 文献检索
  • 文档翻译
  • 深度研究
  • 学术资讯
  • Suppr Zotero 插件Zotero 插件
  • 邀请有礼
  • 套餐&价格
  • 历史记录
应用&插件
Suppr Zotero 插件Zotero 插件浏览器插件Mac 客户端Windows 客户端微信小程序
定价
高级版会员购买积分包购买API积分包
服务
文献检索文档翻译深度研究API 文档MCP 服务
关于我们
关于 Suppr公司介绍联系我们用户协议隐私条款
关注我们

Suppr 超能文献

核心技术专利：CN118964589B侵权必究
粤ICP备2023148730 号-1Suppr @ 2026

文献检索

告别复杂PubMed语法，用中文像聊天一样搜索，搜遍4000万医学文献。AI智能推荐，让科研检索更轻松。

立即免费搜索

文件翻译

保留排版，准确专业，支持PDF/Word/PPT等文件格式，支持 12+语言互译。

免费翻译文档

深度研究

AI帮你快速写综述，25分钟生成高质量综述，智能提取关键信息，辅助科研写作。

立即免费体验

具有曝光校正的稳健非参数分布迁移在图像神经风格转换中的应用。

Robust Nonparametric Distribution Transfer with Exposure Correction for Image Neural Style Transfer.

机构信息

School of Software Engineering, Xi'an Jiaotong University, Xi'an 710049, China.

出版信息

Sensors (Basel). 2020 Sep 14;20(18):5232. doi: 10.3390/s20185232.

DOI:10.3390/s20185232
PMID:32937788
原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC7571219/
Abstract

Image neural style transfer is a process of utilizing convolutional neural networks to render a content image based on a style image. The algorithm can compute a stylized image with original content from the given content image but a new style from the given style image. Style transfer has become a hot topic both in academic literature and industrial applications. The stylized results of current existing models are not ideal because of the color difference between two input images and the inconspicuous details of content image. To solve the problems, we propose two style transfer models based on robust nonparametric distribution transfer. The first model converts the color probability density function of the content image into that of the style image before style transfer. When the color dynamic range of the content image is smaller than that of style image, this model renders more reasonable spatial structure than the existing models. Then, an adaptive detail-enhanced exposure correction algorithm is proposed for underexposed images. Based this, the second model is proposed for the style transfer of underexposed content images. It can further improve the stylized results of underexposed images. Compared with popular methods, the proposed methods achieve the satisfactory qualitative and quantitative results.

摘要

图像神经风格迁移是利用卷积神经网络根据风格图像生成内容图像的过程。该算法可以从给定的内容图像中计算出具有原始内容的风格化图像，而从给定的风格图像中计算出具有新风格的图像。风格迁移在学术文献和工业应用中都是一个热门话题。由于两幅输入图像之间的颜色差异和内容图像中不明显的细节，当前现有模型的风格化结果并不理想。为了解决这些问题，我们提出了两种基于鲁棒非参数分布迁移的风格迁移模型。第一个模型在进行风格迁移之前，将内容图像的颜色概率密度函数转换为风格图像的颜色概率密度函数。当内容图像的颜色动态范围小于风格图像时，该模型比现有模型渲染出更合理的空间结构。然后，提出了一种针对欠曝光图像的自适应细节增强曝光校正算法。在此基础上，提出了第二个模型用于欠曝光内容图像的风格迁移。它可以进一步提高欠曝光图像的风格化效果。与流行的方法相比，所提出的方法在定性和定量方面都取得了令人满意的结果。

https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/ae57a9cd8c52/sensors-20-05232-g019.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/f96f2ac8e184/sensors-20-05232-g001.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/45d4fd9984fe/sensors-20-05232-g002.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/20c501b08caf/sensors-20-05232-g003.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/4f95a48bbb9f/sensors-20-05232-g004.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/096b7d9feb8c/sensors-20-05232-g005.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/b90d2d513a7c/sensors-20-05232-g006.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/f89c89cc1844/sensors-20-05232-g007.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/bdb21e683f15/sensors-20-05232-g008.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/7c1cfaecd5f6/sensors-20-05232-g009.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/d4203f4b5f39/sensors-20-05232-g010.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/aa0d1089130b/sensors-20-05232-g011.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/29ae076f8bc8/sensors-20-05232-g012.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/b67481518e50/sensors-20-05232-g013.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/aca22ae29c88/sensors-20-05232-g014.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/b897ad8f4a1f/sensors-20-05232-g015.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/48b2607e041b/sensors-20-05232-g016.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/8241cf8649a6/sensors-20-05232-g017.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/a1686142f1c0/sensors-20-05232-g018.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/ae57a9cd8c52/sensors-20-05232-g019.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/f96f2ac8e184/sensors-20-05232-g001.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/45d4fd9984fe/sensors-20-05232-g002.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/20c501b08caf/sensors-20-05232-g003.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/4f95a48bbb9f/sensors-20-05232-g004.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/096b7d9feb8c/sensors-20-05232-g005.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/b90d2d513a7c/sensors-20-05232-g006.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/f89c89cc1844/sensors-20-05232-g007.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/bdb21e683f15/sensors-20-05232-g008.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/7c1cfaecd5f6/sensors-20-05232-g009.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/d4203f4b5f39/sensors-20-05232-g010.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/aa0d1089130b/sensors-20-05232-g011.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/29ae076f8bc8/sensors-20-05232-g012.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/b67481518e50/sensors-20-05232-g013.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/aca22ae29c88/sensors-20-05232-g014.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/b897ad8f4a1f/sensors-20-05232-g015.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/48b2607e041b/sensors-20-05232-g016.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/8241cf8649a6/sensors-20-05232-g017.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/a1686142f1c0/sensors-20-05232-g018.jpg
https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83cd/7571219/ae57a9cd8c52/sensors-20-05232-g019.jpg

相似文献

1
Robust Nonparametric Distribution Transfer with Exposure Correction for Image Neural Style Transfer.具有曝光校正的稳健非参数分布迁移在图像神经风格转换中的应用。
Sensors (Basel). 2020 Sep 14;20(18):5232. doi: 10.3390/s20185232.
2
Sand Painting Generation Based on Convolutional Neural Networks.基于卷积神经网络的沙画生成
J Imaging. 2024 Feb 7;10(2):44. doi: 10.3390/jimaging10020044.
3
Style Transfer Via Texture Synthesis.纹理合成的样式转换。
IEEE Trans Image Process. 2017 May;26(5):2338-2351. doi: 10.1109/TIP.2017.2678168. Epub 2017 Mar 8.
4
Neural Style Transfer: A Review.神经风格迁移：综述。
IEEE Trans Vis Comput Graph. 2020 Nov;26(11):3365-3385. doi: 10.1109/TVCG.2019.2921336. Epub 2019 Jun 6.
5
Semantic Context-Aware Image Style Transfer.语义上下文感知图像风格迁移。
IEEE Trans Image Process. 2022;31:1911-1923. doi: 10.1109/TIP.2022.3149237. Epub 2022 Feb 16.
6
Automatic Color Sketch Generation Using Deep Style Transfer.基于深度风格迁移的自动色彩草图生成
IEEE Comput Graph Appl. 2019 Mar-Apr;39(2):26-37. doi: 10.1109/MCG.2019.2899089. Epub 2019 Feb 12.
7
KBStyle: Fast Style Transfer Using a 200 KB Network With Symmetric Knowledge Distillation.KBStyle：使用具有对称知识蒸馏的200KB网络进行快速风格迁移。
IEEE Trans Image Process. 2024;33:82-94. doi: 10.1109/TIP.2023.3335828. Epub 2023 Dec 8.
8
IFFMStyle: High-Quality Image Style Transfer Using Invalid Feature Filter Modules.IFFM风格：使用无效特征过滤模块的高质量图像风格迁移
Sensors (Basel). 2022 Aug 16;22(16):6134. doi: 10.3390/s22166134.
9
Image style transfer with collection representation space and semantic-guided reconstruction.基于集合表示空间和语义引导重建的图像风格迁移。
Neural Netw. 2020 Sep;129:123-137. doi: 10.1016/j.neunet.2020.05.028. Epub 2020 Jun 2.
10
CLAST: Contrastive Learning for Arbitrary Style Transfer.
IEEE Trans Image Process. 2022;31:6761-6772. doi: 10.1109/TIP.2022.3215899. Epub 2022 Oct 28.

本文引用的文献

1
Neural Style Transfer: A Review.神经风格迁移：综述。
IEEE Trans Vis Comput Graph. 2020 Nov;26(11):3365-3385. doi: 10.1109/TVCG.2019.2921336. Epub 2019 Jun 6.
2
Gated-GAN: Adversarial Gated Networks for Multi-Collection Style Transfer.门控 GAN：用于多集合风格迁移的对抗门控网络。
IEEE Trans Image Process. 2019 Feb;28(2):546-560. doi: 10.1109/TIP.2018.2869695. Epub 2018 Sep 12.
3
Deep learning.深度学习。
Nature. 2015 May 28;521(7553):436-44. doi: 10.1038/nature14539.
4
A generalized Laplacian of Gaussian filter for blob detection and its applications.用于斑点检测的广义拉普拉斯高斯滤波器及其应用。
IEEE Trans Cybern. 2013 Dec;43(6):1719-33. doi: 10.1109/TSMCB.2012.2228639.
5
Guided image filtering.引导图像滤波。
IEEE Trans Pattern Anal Mach Intell. 2013 Jun;35(6):1397-409. doi: 10.1109/TPAMI.2012.213.
6
State of the "art": a taxonomy of artistic stylization techniques for images and video.艺术的现状：图像和视频艺术风格化技术的分类法。
IEEE Trans Vis Comput Graph. 2013 May;19(5):866-85. doi: 10.1109/TVCG.2012.160.
7
Image quality assessment: from error visibility to structural similarity.图像质量评估：从误差可见性到结构相似性。
IEEE Trans Image Process. 2004 Apr;13(4):600-12. doi: 10.1109/tip.2003.819861.