[1]肖红梅,王伟珍*,房 媛.基于生成对抗网络的女上装图像属性编辑[J].服装学报,2024,9(01):42-47.
 XIAO Hongmei,WANG Weizhen*,FANG Yuan.Image Attribute Editing of Women’s Tops Based on Generating Adversarial Networks[J].Journal of Clothing Research,2024,9(01):42-47.
点击复制

基于生成对抗网络的女上装图像属性编辑()
分享到:

《服装学报》[ISSN:2096-1928/CN:32-1864/TS]

卷:
第9卷
期数:
2024年01期
页码:
42-47
栏目:
服装智造
出版日期:
2024-03-01

文章信息/Info

Title:
Image Attribute Editing of Women’s Tops Based on Generating Adversarial Networks
作者:
肖红梅1;  王伟珍*1; 2;  房 媛3
1.大连工业大学 服装学院,辽宁 大连 116034; 2.大连工业大学 服装人因与智能设计研究中心,辽宁 大连 116034; 3.大连工业大学 工程训练中心,辽宁 大连 116034
Author(s):
XIAO Hongmei1;  WANG Weizhen*1; 2;  FANG Yuan3
1.School of Fashion, Dalian Polytechnic University, Dalian 116034, China; 2.Clothing Human Factors and Intelligent Design Research Center, Dalian Polytechnic University, Dalian 116034, China; 3.Engineering Training Center, Dalian Polytechnic University, Dalian 116034, China
分类号:
TS 941.26
文献标志码:
A
摘要:
为解决当前服装图像属性编辑模型生成图像存在属性缺失或冗余的问题,提出一种基于Fashion-AttGAN的优化模型对女上装图像细节进行变换的设计方法; 通过优化特征提取网络,将结构相似性损失项加入重构损失,提高生成器的属性编辑能力; 使用CP-VTON数据集训练,对女上装图像中袖长和颜色的细节进行调整。结果表明,生成图像在袖型连贯性和颜色准确性方面得到提升,改进模型收敛趋势更平稳,重构图像的结构相似性指标提升了27.4%,峰值信噪比提高了2.8%。该优化模型有效减少了生成图像的属性冗余和残缺,为服装图像细节变换研究提供参考。
Abstract:
In order to solve the problem of attributes missing or redundant in the current clothing image attribute editing models of generate images, a design method based on Fashion-AttGAN model was conducted to transform the details of women’s tops. This paper optimized feature network and added structure similarity index measure to the reconstructed loss function to improve the attribute editing ability of generator. The CP-VTON dataset was used for training to ultimately achieve fine-grained editing of women’s tops sleeve length and color. The experimental results show that the generated image achieves the improvement in sleeve coherence and color accuracy, the improved model is shown to move more smoothly towards convergence trend, the reconstructed image structure similarity index measure realizes the growth of 27.4% and peak signal-to-noise ratio grows by 2.8%. The proposed model reduces attributes missing or redundant in generated images and provides a technical reference for its detail transformation.

参考文献/References:

[1] ZHU S Z, FIDLER S, URTASUN R, et al. Be your own prada: fashion synthesis with structural coherence[C]//2017 IEEE International Conference on Computer Vision(ICCV). Venice: IEEE, 2017: 1689-1697.
[2] RADFORD A, METZ L, CHINTALAS S. Unsupervised representation learning with deep convolutional gen-erative adversarial networks[EB/OL].(2016-01-07)[2023-05-28]. https://arxiv.org/pdf/1511.06434.pdf.
[3] PING Q, WU B, DING W Y, et al. Fashion-AttGAN: attribute-aware fashion editing with multi-objective GAN[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops(CVPRW). Long Beach: IEEE, 2019: 323-325.
[4] HE Z L, ZUO W M, KAN M N, et al. AttGAN: facial attribute editing by only changing what you want[J]. IEEE Transactions on Image Processing, 2019, 28(11): 5464-5478.
[5] YUAN C X, MOGHADDAM M. Attribute-aware generative design with generative adversarial networks[J]. IEEE Access, 2020(8): 190710-190721.
[6] AK K, LIM J H, THAM J, et al. Attribute manipulation generative adversarial networks for fashion images[C]//2019 IEEE/CVF International Conference on Computer Vision(ICCV). Seoul: IEEE, 2019: 10540-10549.
[7] CHEN L L, TIAN J, LI G, et al. TailorGAN: making user-defined fashion designs[C]//2020 IEEE Winter Conference on Applications of Computer Vision(WACV). Snowmass: IEEE, 2020: 3241-3250.
[8] KWON Y, PETRANGELI S, KIM D, et al. Tailor me: an editing network for fashion attribute shape manipulation[C]//2022 IEEE/CVF Winter Conference on Applications of Computer Vision(WACV). Waikoloa: IEEE, 2022: 3142-3151.
[9] LIU M, DING Y K, XIA M, et al. STGAN: a unified selective transfer network for arbitrary image attribute editing[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). Long Beach:IEEE, 2019: 3668-3677.
[10] ZHANG G, KAN M N, SHAN S G, et al. Generative adversarial network with spatial attention for face attribute editing[C]//Computer Vision-ECCV 2018: 15th European Conference. Munich: ACM, 2018: 422-437.
[11] CRESWELL A, WHITE T, DUMOULIN V, et al. Gene-rative adversarial networks: an overview[J]. IEEE Signal Processing Magazine, 2018, 35(1): 53- 65.
[12] WANG Z, BOVIK A C, SHEIKH H R, et al. Image quality assessment: from error visibility to structural similarity[J]. IEEE Transactions on Image Processing: a Publication of the IEEE Signal Processing Society, 2004, 13(4): 600- 612.
[13] WANG B C, ZHENG H B, LIANG X D, et al. Toward characteristic-preserving image-based virtual try-on network[C]//Computer Vision-ECCV 2018: 15th European Conference.Munich:ACM, 2018: 607- 623.
(责任编辑:张 雪)

相似文献/References:

[1]陈 璐,雷聪聪,陈 郁*.用户面部情绪与其对服装感兴趣度的关联分析[J].服装学报,2021,6(04):311.
 CHEN Lu,LEI Congcong,CHEN Yu*.Study on the Relationship Between User’s Facial Emotions and Clothing Interests Level[J].Journal of Clothing Research,2021,6(01):311.
[2]韩曙光,姜凯文,赵丽妍.基于深度学习的服装三要素识别[J].服装学报,2022,7(05):399.
 HAN Shuguang,JIANG Kaiwen,ZHAO Liyan.Recognition of Clothing "Three Elements" Based on Deep Learning[J].Journal of Clothing Research,2022,7(01):399.
[3]刘 康,马浩然,邢 乐*.基于生成对抗网络的中式婚服设计[J].服装学报,2024,9(03):208.
 LIU Kang,MA Haoran,XING Le*.Chinese Wedding Dress Design Based on Generative Adversarial Network[J].Journal of Clothing Research,2024,9(01):208.
[4]季 勇,仇明慧,魏 佳,等.人体模型建模中全局映射智能估计人体形状的方法[J].服装学报,2024,9(03):223.
 JI Yong,QIU MINGHui,WEI Jia,et al.Intelligent Estimation Method of Human Body Shape in Human Body Modeling[J].Journal of Clothing Research,2024,9(01):223.

更新日期/Last Update: 2024-02-29