基于卷积神经网络的汉服关键尺寸自动测量

doi:10.13475/j.fzxb.20200505006

纺织学报 ›› 2020, Vol. 41 ›› Issue (12): 124-129.doi: 10.13475/j.fzxb.20200505006

基于卷积神经网络的汉服关键尺寸自动测量

王奕文¹, 罗戎蕾²^,³(), 康宇哲⁴

1.浙江理工大学服装学院, 浙江杭州 310018
2.浙江理工大学国际教育学院, 浙江杭州 310018
3.浙江省丝绸与时尚文化研究中心, 浙江杭州 310018
4.浙江理工大学信息学院, 浙江杭州 310018

收稿日期:2020-05-26 修回日期:2020-08-31 出版日期:2020-12-15 发布日期:2020-12-23
通讯作者: 罗戎蕾
作者简介:王奕文(1993—),女,硕士生。主要研究方向为服装数字化复原与研究。
基金资助:
浙江省“十三五”高校虚拟仿真实验教学项目(浙教办函[2019]365号)

Automatic measurement of key dimensions for Han-style costumes based on use of convolutional neural network

WANG Yiwen¹, LUO Ronglei²^,³(), KANG Yuzhe⁴

1. School of Fashion Design & Engineering, Zhejiang Sci-Tech University, Hangzhou, Zhejiang 310018, China
2. School of International Education, Zhejiang Sci-Tech University, Hangzhou, Zhejiang 310018, China
3. Silk and Fashion Culture Center, Hangzhou, Zhejiang 310018, China
4. School of Information Science and Technology, Zhejiang Sci-Tech University, Hangzhou, Zhejiang 310018, China

Received:2020-05-26 Revised:2020-08-31 Online:2020-12-15 Published:2020-12-23
Contact: LUO Ronglei

摘要/Abstract

摘要：

在样本数据稀缺的情况下,为快速准确地获取古代汉服的关键尺寸数据,提出一种基于卷积神经网络的汉服尺寸测量方案。首先搭建1个二阶段卷积神经网络模型GlobalNet-RefineNet进行关键点检测,通过2次迁移学习和反复迭代训练提高关键点识别准确率;再利用算法得到坐标点的像素距离,结合博物馆或发掘报告中给出的汉服平铺图和至少1个真实测量尺寸,通过比例映射得到全衣的尺寸数据。以汉服上衣为例进行实验验证,结果表明:经过2次迁移学习,卷积神经网络模型的收敛程度高,训练效果好,通过该方案测得的汉服上衣尺寸相对误差在0.59%~4.17%之间;该方案为传统服饰的复原研究和文物尺寸测量工作提供了新思路。

关键词: 尺寸测量, 服装关键尺寸, 汉服, 卷积神经网络, 迁移学习

Abstract:

In order to quickly and accurately obtain the key dimensions of the ancient Chinese Han-style costumes with scarce sample data, a clothing size measurement scheme based on the use of convolutional neural network was proposed in this paper. Firstly, a two-stage convolutional neural network model GlobalNet-RefineNet was built for detecting the key points. The accuracy of the key point recognition was improved through twice transfer learning and repeated iterative training. An algorithm was used to get the pixel distance between coordinate points. Combined with the tiles of Han-style costume and at least one real measurement size given in the museum or excavation report, the size data of the whole garment were obtained through proportional mapping. This research used the top of a Han-style costume as an example for experiments. The research results show that after two times of transfer learning, the model has a high degree of convergence and good training effect. The relative error of costume top size measured by this scheme is between 0.59%-4.17%. This research provides new ideas for the restoration research of traditional clothing and the measurement of cultural relics.

Key words: dimensional measurement, key dimensions of costume, Han-style costume, convolutional neural network, transfer learning

中图分类号:

TS941.79

王奕文, 罗戎蕾, 康宇哲. 基于卷积神经网络的汉服关键尺寸自动测量[J]. 纺织学报, 2020, 41(12): 124-129.

WANG Yiwen, LUO Ronglei, KANG Yuzhe. Automatic measurement of key dimensions for Han-style costumes based on use of convolutional neural network[J]. Journal of Textile Research, 2020, 41(12): 124-129.

图/表 10

图1

图2

图3

表1

表2

图4

图5

表3

表4

表5

参考文献 15

[1]	曹丽, 汪亚明, 包晓敏. 机器视觉在服装尺寸自动测量中的应用[J]. 纺织学报, 2003,24(1):19-21.
	CAO Li, WANG Yaming, BAO Xiaomin. Application of machine vision in automatic measurement of clothing size[J]. Journal of Textile Research, 2003,24(1):19-21.
[2]	王生伟, 张凯健. 基于角点检测的服装尺寸在线测量技术[J]. 信息技术与信息化, 2018,12:73-75.
	WANG Shengwei, ZHANG Kaijian. Online measurement technology for garment size based on image process-ing[J]. Information Technology and Informatization, 2018,12:73-75.
[3]	董建明, 胡觉亮. 一种有效的服装尺寸自动测量方法[J]. 纺织学报, 2008,29(5):98-101.
	DONG Jianming, HU Jueliang. An efficient method for automatic measurement of garment dimensions[J]. Journal of Textile Research, 2008,29(5):98-101.
[4]	肖祎. 基于拍照的服装和人体尺寸测量系统设计与研发[D]. 杭州:浙江大学, 2019: 25-37.
	XIAO Yi. Design and development of a garment and human body measurement system based on photo-graphy[D]. Hangzhou:Zhejiang University, 2019: 25-37.
[5]	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE Computer Society, 2016: 770-778.
[6]	LIN T Y, DOLLA P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. State of California: IEEE Computer Society, 2017: 106.
[7]	GROOS D, RAMAMPIARO H, IHLEN E. Efficientpose: scalable single-person pose estima-tion[C]// Proceedings of the Association for the Advance of Artificial Intelligence. New York: Applied Intelligence, 2020: 1-14.
[8]	ZHANG Q, LU Z, ZHOU X, et al. Automatic removal of false image stars in disk-resolved images of the cassini imaging science subsystem[J]. Research in Astronomy and Astrophysics, 2020,20(7):95-104. doi: 10.1088/1674-4527/20/6/95
[9]	段萌. 基于卷积神经网络的图像识别方法研究[D]. 郑州:郑州大学, 2017: 42.
	DUAN Meng. The research on the mothed of image recognition based on convolutional neural networks[D]. Zhengzhou:Zhengzhou University, 2017: 42.
[10]	Fashion AI赛题与数据[DB/OL]. 阿里云天池[2020-04-25]. https://tianchi.aliyun.com/competition/entrance/231648/introduction.
	Fashion AI competition questions and data[DB/OL]. Alibaba Tianchi[2020-04-25]. https://tianchi.aliyun.com/competition/entrance/231648/introduction.
[11]	LIU Ziwei, LUO Ping, QIU Shi, et al. Large-scale fashion (deepfashion) database[DB/OL]. The Chinese University of Hong Kong [2020-04-25]. http://mmlab.ie.cuhk.edu.hk/projects/DeepFashion.html.
[12]	XU Luohongshang. FashionAI全球挑战赛-服饰属性识别赛后技术分享[EB/OL]. [2020-04-25]. https://blog.csdn.net/xuluohongshang/article/details/80616331.
	XU Luohongshang. Fashion AI global challenge-apparel attribute recognition technology sharing after the game [EB/OL]. [2020-04-25]. https://blog.csdn.net/xuluohongshang/article/details/80616331.
[13]	王淑娟, 李一泉, 蒋玉秋, 等. 钱家衣橱无锡七房桥明墓出土服饰保护修复展[M]. 杭州:中国丝绸博物馆, 2017: 14-16.
	WANG Shujuan, LI Yiquan, JIANG Yuqiu, et al. Qian family's wardrobe costume found in the tomb of Qian Zhang(1486-1505) and his wife[M]. Hangzhou: China National Silk Museum, 2017: 14-16.
[14]	泰州市博物馆. 江苏泰州森森庄明墓发掘简报[J]. 文物, 2013,11:36-49.
	Taizhou Museum. The excavation of a tomb of the Ming dynasty at Sensenzhuang in Taizhou city, Jiangsu[J]. Cultural Relics, 2013,11:36-49.
[15]	山东博物馆. 斯文在兹孔府旧藏服饰[M]. 济南:山东博物馆, 2012: 64.
	Shandong Museum. Clothing from the collections of Kong family mansion[M]. Ji'nan:Shandong Museum, 2012: 64.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

类型	测量对象张数	含不同个数关键点图片张数					单点准确率/%	漏识别率/ %	识别准确率/%
类型	测量对象张数	13 个	10~13 个		10~7 个	7个以下	单点准确率/%	漏识别率/ %	识别准确率/%
现代汉服	50	20		15	15	0	68	0	40
古代汉服	50	17		14	19	0	61	0	34

类型	测量对象张数	含不同个数关键点图片张数					单点准确率/%	漏识别率/ %	识别准确率/%
类型	测量对象张数	13个	10~13 个		7~10 个	7个以下	单点准确率/%	漏识别率/ %	识别准确率/%
现代汉服	50	44		6	0	0	95	0	92
古代汉服	50	40		8	2	0	86	0	80

上衣部位	实际值/ mm	像素距离/像素	测量值/mm	绝对误差/mm	相对误差/%
领口宽	160	79	165	5	3.13
袖口宽	120	55	115	5	4.17
摆宽	1 120	530	1 108	9	0.80
衣长	725	344	719	6	0.83

上衣部位	实际值/ mm	像素距离/像素	测量值/ mm	绝对误差/mm	相对误差/%
领口宽	100	55	97	3	3.00
袖宽	680	384	676	4	0.59
衣长	810	456	803	7	0.86

上衣部位	实际值/ mm	像素距离/像素	测量值/ mm	绝对误差/mm	相对误差/%
袖宽	360	203	366	6	1.67
腰宽	590	331	597	7	1.19
衣长	630	344	621	9	1.43

基于卷积神经网络的汉服关键尺寸自动测量

Automatic measurement of key dimensions for Han-style costumes based on use of convolutional neural network

RichHTML

PDF (PC)

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 15

相关文章 12

Metrics

本文评价

推荐阅读 0

[1]	邵金鑫, 张宝昌, 曹继鹏. 基于图像处理与深度学习方法的棉纤维梳理过程纤维检测识别技术[J]. 纺织学报, 2020, 41(07): 40-46.
[2]	夏海浜, 黄鸿云, 丁佐华. 基于迁移学习与支持向量机的服装舒适度评估[J]. 纺织学报, 2020, 41(06): 125-131.
[3]	王泽霞, 陈革, 陈振中. 基于改进卷积神经网络的化纤丝饼表面缺陷识别[J]. 纺织学报, 2020, 41(04): 39-44.
[4]	王晓华, 姚炜铭, 王文杰, 张蕾, 李鹏飞. 基于改进YOLO深度卷积神经网络的缝纫手势检测[J]. 纺织学报, 2020, 41(04): 142-148.
[5]	贾小军, 叶利华, 邓洪涛, 刘子豪, 陆锋杰. 基于卷积神经网络的蓝印花布纹样基元分类 [J]. 纺织学报, 2020, 41(01): 110-117.
[6]	孙洁, 丁笑君, 杜磊, 李秦曼, 邹奉元. 基于卷积神经网络的织物图像特征提取与检索研究进展[J]. 纺织学报, 2019, 40(12): 146-151.
[7]	刘正东, 刘以涵, 王首人. 西装识别的深度学习方法[J]. 纺织学报, 2019, 40(04): 158-164.
[8]	吴欢, 丁笑君, 李秦曼, 杜磊, 邹奉元. 采用卷积神经网络 CaffeNet 模型的女裤廓形分类[J]. 纺织学报, 2019, 40(04): 117-121.
[9]	汪珊娜张华熊康锋. 基于卷积神经网络的领带花型情感分类[J]. 纺织学报, 2018, 39(08): 117-123.
[10]	王雯雯高畅刘基宏. 应用卷积神经网络的细纱断纱锭位识别[J]. 纺织学报, 2018, 39(06): 136-141.
[11]	景军锋范晓婷李鹏飞洪良. 应用深度卷积神经网络的色织物缺陷检测[J]. 纺织学报, 2017, 38(02): 68-74.
[12]	葛宝臻, 郭华婷, 彭博, 田庆国, 李云鹏. 基于人体特征提取的模特体形尺寸自动测量方法[J]. 纺织学报, 2012, 33(4): 129-135.