使用深度卷积生成对抗网络为男性生成逼真的非洲时装设计

E. Ogbuju, Gabriel Yashim, Francisca Onaolapo Oladipo
{"title":"使用深度卷积生成对抗网络为男性生成逼真的非洲时装设计","authors":"E. Ogbuju, Gabriel Yashim, Francisca Onaolapo Oladipo","doi":"10.1109/SEB-SDG57117.2023.10124486","DOIUrl":null,"url":null,"abstract":"The African fashion industry has continued to thrive over the years and has gained recognition globally. However, the design and manufacturing process still follows the conventional methodology. Some studies have been done on image synthesis using deep learning techniques for the progress of the fashion industry with little or none of the studies on African fashion. African fashion, especially Nigerian fashion lacks an online presence in terms of artificially generated realistic fashion styles. Since fashion is a way a people express themselves through clothing, footwear, makeup or hairdo in a specific time, place, and environment, this work aims to develop a system that would promote the diversity of the Nigerian people, reduce inequality and promote economic growth in line with the sustainable development goals. In this work, we developed a system that can generate new African fashion designs for men by applying the Deep Convolutional Generative Adversarial Networks (DCGAN) using a locally curated image data from publicly accessible and open-source images and the AFRIFASHION40000 dataset. The method involves designing a Male Fashion Generative Framework (MFGF), building an image generation model that can synthesize new fashion images and a neural style transfer model that can style the generated image. The model trained on the AFRIFASHION40000 dataset was found to perform better than the locally curated dataset as indicated by a lower generator loss and a higher discriminator loss. This study provides a solution to the problem of generating new designs for men in the Nigerian fashion industry and highlights the importance of integrating deep learning into the African fashion industry. The result shows a generation of low resolution images that represents different fashion styles for African men. It also provides a framework for future research. The findings can be beneficial to fashion designers, manufacturers, and consumers looking for new and innovative designs.","PeriodicalId":185729,"journal":{"name":"2023 International Conference on Science, Engineering and Business for Sustainable Development Goals (SEB-SDG)","volume":"65 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Generating Realistic African Fashion Designs for Men using Deep Convolutional Generative Adversarial Networks\",\"authors\":\"E. Ogbuju, Gabriel Yashim, Francisca Onaolapo Oladipo\",\"doi\":\"10.1109/SEB-SDG57117.2023.10124486\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The African fashion industry has continued to thrive over the years and has gained recognition globally. However, the design and manufacturing process still follows the conventional methodology. Some studies have been done on image synthesis using deep learning techniques for the progress of the fashion industry with little or none of the studies on African fashion. African fashion, especially Nigerian fashion lacks an online presence in terms of artificially generated realistic fashion styles. Since fashion is a way a people express themselves through clothing, footwear, makeup or hairdo in a specific time, place, and environment, this work aims to develop a system that would promote the diversity of the Nigerian people, reduce inequality and promote economic growth in line with the sustainable development goals. In this work, we developed a system that can generate new African fashion designs for men by applying the Deep Convolutional Generative Adversarial Networks (DCGAN) using a locally curated image data from publicly accessible and open-source images and the AFRIFASHION40000 dataset. The method involves designing a Male Fashion Generative Framework (MFGF), building an image generation model that can synthesize new fashion images and a neural style transfer model that can style the generated image. The model trained on the AFRIFASHION40000 dataset was found to perform better than the locally curated dataset as indicated by a lower generator loss and a higher discriminator loss. This study provides a solution to the problem of generating new designs for men in the Nigerian fashion industry and highlights the importance of integrating deep learning into the African fashion industry. The result shows a generation of low resolution images that represents different fashion styles for African men. It also provides a framework for future research. The findings can be beneficial to fashion designers, manufacturers, and consumers looking for new and innovative designs.\",\"PeriodicalId\":185729,\"journal\":{\"name\":\"2023 International Conference on Science, Engineering and Business for Sustainable Development Goals (SEB-SDG)\",\"volume\":\"65 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2023-04-05\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2023 International Conference on Science, Engineering and Business for Sustainable Development Goals (SEB-SDG)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/SEB-SDG57117.2023.10124486\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2023 International Conference on Science, Engineering and Business for Sustainable Development Goals (SEB-SDG)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/SEB-SDG57117.2023.10124486","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

多年来,非洲时尚产业持续蓬勃发展,并获得了全球的认可。然而,设计和制造过程仍然遵循传统的方法。一些研究利用深度学习技术进行图像合成,以促进时尚产业的发展,但对非洲时尚的研究很少或根本没有。非洲时装,尤其是尼日利亚时装,在人工生成的现实时尚风格方面缺乏在线存在。由于时尚是人们在特定的时间、地点和环境中通过服装、鞋类、化妆或发型来表达自己的一种方式,因此这项工作旨在建立一个能够促进尼日利亚人民多样性、减少不平等并促进符合可持续发展目标的经济增长的系统。在这项工作中,我们开发了一个系统,通过应用深度卷积生成对抗网络(DCGAN),使用来自公开访问和开源图像的本地策划图像数据以及AFRIFASHION40000数据集,该系统可以为男性生成新的非洲时装设计。该方法包括设计一个男性时尚生成框架(MFGF),建立一个可以合成新时尚图像的图像生成模型和一个可以对生成的图像进行样式化的神经风格迁移模型。在AFRIFASHION40000数据集上训练的模型比本地策划的数据集表现得更好,这表明生成器损失更低,鉴别器损失更高。这项研究为尼日利亚时尚行业的男性提供了一个解决方案,并强调了将深度学习融入非洲时尚行业的重要性。结果显示了代表非洲男性不同时尚风格的低分辨率图像。它也为未来的研究提供了一个框架。这些发现对时尚设计师、制造商和寻求创新设计的消费者都是有益的。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Generating Realistic African Fashion Designs for Men using Deep Convolutional Generative Adversarial Networks
The African fashion industry has continued to thrive over the years and has gained recognition globally. However, the design and manufacturing process still follows the conventional methodology. Some studies have been done on image synthesis using deep learning techniques for the progress of the fashion industry with little or none of the studies on African fashion. African fashion, especially Nigerian fashion lacks an online presence in terms of artificially generated realistic fashion styles. Since fashion is a way a people express themselves through clothing, footwear, makeup or hairdo in a specific time, place, and environment, this work aims to develop a system that would promote the diversity of the Nigerian people, reduce inequality and promote economic growth in line with the sustainable development goals. In this work, we developed a system that can generate new African fashion designs for men by applying the Deep Convolutional Generative Adversarial Networks (DCGAN) using a locally curated image data from publicly accessible and open-source images and the AFRIFASHION40000 dataset. The method involves designing a Male Fashion Generative Framework (MFGF), building an image generation model that can synthesize new fashion images and a neural style transfer model that can style the generated image. The model trained on the AFRIFASHION40000 dataset was found to perform better than the locally curated dataset as indicated by a lower generator loss and a higher discriminator loss. This study provides a solution to the problem of generating new designs for men in the Nigerian fashion industry and highlights the importance of integrating deep learning into the African fashion industry. The result shows a generation of low resolution images that represents different fashion styles for African men. It also provides a framework for future research. The findings can be beneficial to fashion designers, manufacturers, and consumers looking for new and innovative designs.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信