基于改进边缘检测算法的服装款式识别

doi:10.13475/j.fzxb.20201205006

摘要/Abstract

摘要：

为快速识别服装款式类型,提高生产效率,针对现有传统边缘检测算法难以准确提取轮廓特征序列的不足,设计一种改进的边缘提取算法。通过定义一种新的优化卷积核,在使用传统边缘检测算法提取训练样本的服装轮廓基础上,将该卷积核与目标矩阵进行卷积得到新的外轮廓,将新轮廓序列的傅里叶描述子作为特征向量,进一步利用BP神经网络模型完成服装款式的自动分类与识别。为验证改进方法的有效性,建立一个包含4类服装500个不重复服装图像的样本库,选取281个作为训练样本,对剩余219个样本进行测试,测试识别准确率最低为93.48%,最高达到了100%。该改进算法提高了服装款式识别率,对服装智能化生产具有借鉴意义。

关键词: 服装款式识别, 边缘检测算法, 傅里叶描述子, BP神经网络

Abstract:

In order to quickly identify clothing styles and improve production efficiency, an improved edge extraction algorithm was designed to solve the problem that the existing traditional edge detection algorithm was difficult to accurately extract contour feature sequence. By defining a new optimization convolution kernels, the use of traditional edge detection algorithm based on clothing contours of the training sample, will the convolution convolution kernels and target matrix are new outer contour, a new contour sequence of Fourier descriptor as a feature vector, to increase the use of BP neural network model to complete the design of automatic classification and recognition. In order to verify the effectiveness of the improved method, a sample library containing 500 non-repeated clothing images of four categories of clothing was established. 281 samples were selected as training samples and the remaining 219 samples were tested. The recognition accuracy of the test was as low as 93.48% and as high as 100%. It is of reference significance to the intelligent production of clothing.

Key words: clothing style identification, edge detection algorithm, fourier descriptor, BP neural network

中图分类号:

TS941.26

庹武, 王哓玉, 高雅昆, 于媛媛, 郝潇潇, 刘永亮, 郭鑫. 基于改进边缘检测算法的服装款式识别[J]. 纺织学报, 2021, 42(10): 157-162.

TUO Wu, WANG Xiaoyu, GAO Yakun, YU Yuanyuan, HAO Xiaoxiao, LIU Yongliang, GUO Xin. Clothing style identification based on improved edge detection algorithm[J]. Journal of Textile Research, 2021, 42(10): 157-162.

图/表 15

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

图11

表1

图12

表2

表3

参考文献 19

[1]	郭鑫鹏, 黄元元, 胡作进. 融合颜色与纹理的复杂场景下的服装图像分割算法[J]. 计算机科学, 2017, 44(S2):228-232.
	GUO Xinpeng, HUANG Yuanyuan, HU Zuojin. A garment image segmentation algorithm based on color and texture fusion in complexscenes[J]. Computer Science, 2017, 44(S2):228-232.
[2]	郝雪洁. 基于深度神经网络的细粒度图像分类方法研究[D]. 成都:电子科技大学, 2020: 49-58.
	HAO Xuejie. Study on fine-grained image classification method based on deep neural network[D]. Chengdu: University of Electronic Science and Technology of China, 2020:49-58.
[3]	KRITI Bansal, ANAND Singh Jalal. Robust and effective clothes recognition system based on fusion of Haralick and HOG features[J]. International Journal of Signal and Imaging Systems Engineering, 2019, 11(5):288-299. doi: 10.1504/IJSISE.2019.100649
[4]	尹光灿, 罗戎蕾. 基于卷积神经网络的服装领型识别与分类研究[J]. 现代纺织技术, 2020, 28(3):48-53.
	YIN Guangcan, LUO Ronglei. Research on recognition and classification of Garment collar based on convolutional neural network[J]. Advanced Textile Technology, 2020, 28(3):48-53.
[5]	JIANG Meiqi, YANG Xing, FANG Xia. Scene-specific clothing recognition based on YOLOv3 algorithm[J]. Journal of Research in Science and Engineering, 2020, 2(1):23-26. doi: 10.26666/rmp.jesr
[6]	HUSSAIN K F, ALISA, MALEK S M. An efficient approach for automatic cloth panel extraction from pattern images[J]. International Journal of Computer Applications, 2014, 98(3) : 15-22.
[7]	李东, 万贤福, 汪军, 等. 基于轮廓曲率特征点的服装款式识别方法[J]. 东华大学学报(自然科学版), 2018, 44(1):87-92.
	Li Dong, WAN Xianfu, WANG Jun, et al. Clothing style identification method based on contour curvature feature points[J]. Journal of Donghua University (Natural Science), 2018, 44(1) :87-92.
[8]	SUN Feifei, XU Pinghua, DING Xuemei. Multi-core SVM optimized visual word package model for garment style classification[J]. Cluster Computing, 2019, 22(2):4141-4147. doi: 10.1007/s10586-017-1651-4
[9]	丁敏敏. 服装衣领款式图数字化识别研究与实现[D]. 上海:东华大学, 2007:14-18.
	DING Minmin. Research and Implementation of digital identification of garment collar pattern[D]. Shanghai:Donghua University, 2007:14-18.
[10]	SUN G L, WU X, CHEN H H, et al. Clothing style recognition using fashion attribute detection [C]//International Conference on Mobile Multimedia Communications. Chengdu: ICST (Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering), 2015:25-27.
[11]	李鹏飞, 郑明智, 景军锋. 基于机器视觉的服装尺寸在线测量系统[J]. 毛纺科技, 2017, 45(3):42-47.
	LI Pengfei, ZHENG Mingzhi, JING Junfeng. Online measurement system of garment size based on machine vision[J]. Wool Textile Journal, 2017, 45(3):42-47.
[12]	YU C, CHEN D R, YANG L, et al. Otsu's thresholding method based on gray level-gradient two-dimensional histogram [C]//2010 2nd International Asia Conference on Informatics in Control, Automation and Robotics (CAR 2010). New York: IEEE, 2010:282-285.
[13]	胡聪, 屈瑾瑾, 许川佩, 等. 基于自适应池化的神经网络的服装图像识别[J]. 计算机应用, 2018, 38(8):2211-2217.
	HU Cong, QU Jinjin, XU Chuanpei, et al. Clothing image recognition based on adaptive pooling neural network[J]. Computer Application, 2018, 38(8):2211-2217.
[14]	庄立锋, 林俊文. 基于改进Canny算法的服装款式结构特征识别与分类[J]. 实验室研究与探索, 2020, 39(5):264-268.
	ZHUANG Lifeng, LIN Junwen. Recognition and classification of clothing style structure characteristics based on improved Canny algorithm[J]. Laboratory Research and Exploration, 2020, 39(5):264-268.
[15]	HALSTEAD M A, DENMAN S, SRIDHARAN S, et al. Multimodal clothing recognition for semantic search in unconstrained surveillance imagery[J]. Journal of Visual Communication & Image Representation, 2019, 58:439-452.
[16]	PERSOON E, FU K S . Shape discrimination using Fourier descriptors[J]. IEEE Transactions on Systems Man & Cybernetics, 1977, 7(3):170-179.
[17]	王涛, 刘文印, 孙家广, 等. 傅立叶描述子识别物体的形状[J]. 计算机研究与发展, 2002(12):1714-1719.
	WANG Tao, LIU Wenyin, SUN Jiaguang, et al. Fourier descriptors recognize the shape of objects[J]. Computer Research and Development, 2002(12):1714-1719.
[18]	于雨桐. 基于多特征融合的服装图像款式识别的研究与应用[D]. 北京:北京工业大学, 2016:35-37.
	YU Yutong. Research and application of garment image style recognition based on multi-feature fusion[D]. Beijing:Beijing University of Technology, 2016:35-37.
[19]	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(4):158-164.
	LIU Zhengdong, LIU Yihan, WANG Shouren. Deep learning method for suit recognition[J]. Journal of Textil Research, 2019, 40(4):158-164. doi: 10.1177/004051757004000209

相关文章 15

[1]	庹武王哓玉高雅昆郝潇潇刘永亮郭鑫. 基于改进边缘检测算法的服装款式识别[J]. , 2021, 42(10): 0-0.
[2]	许雪梅. 基于模拟退火算法改进遗传算法的织物智能配色[J]. 纺织学报, 2021, 42(07): 123-128.
[3]	张晓侠, 刘凤坤, 买巍, 马崇启. 基于BP神经网络及其改进算法的织机效率预测[J]. 纺织学报, 2020, 41(08): 121-127.
[4]	金守峰, 林强强, 马秋瑞, 张浩. 基于BP神经网络的织物表面绒毛质量的检测方法[J]. 纺织学报, 2020, 41(02): 69-76.
[5]	王晓晖, 刘月刚, 孟婥, 孙以泽. 基于遗传算法和神经网络的3D增材印花工艺参数优化[J]. 纺织学报, 2019, 40(11): 168-174.
[6]	周捷, 马秋瑞. 基于BP神经网络的运动文胸肩带属性与乳房振幅的函数关系[J]. 纺织学报, 2019, 40(09): 186-191.
[7]	周捷, 马秋瑞. BP神经网络在塑身内衣压力预测中的应用[J]. 纺织学报, 2019, 40(04): 111-116.
[8]	查刘根, 谢春萍. 应用四层BP神经网络的棉纱成纱质量预测[J]. 纺织学报, 2019, 40(01): 52-56.
[9]	巫莹柱　单颖法　黄伯熹　林广茂　梁家豪　张晓利. 聚对苯二甲酸丙二醇酯与聚对苯二甲酸丁二醇酯混纺纤维的智能识别[J]. 纺织学报, 2018, 39(09): 169-175.
[10]	李东万贤福汪军. 采用傅里叶描述子和支持向量机的服装款式识别方法[J]. 纺织学报, 2017, 38(05): 122-127.
[11]	张宁潘如如高卫东. 采用图像处理的织物缝纫平整度自动评估[J]. 纺织学报, 2017, 38(04): 145-150.
[12]	夏明陈益松张文斌. 基于椭圆傅里叶的人体胸围断面形状研究[J]. 纺织学报, 2014, 35(7): 107-0.
[13]	项明姚雪存崔振东华铨平. 利用相关反馈技术和形状特征的服装图像检索算法[J]. 纺织学报, 2014, 35(2): 94-0.
[14]	杨梦星张毅郑秋婷. BP神经网络预测环境对丝织物力学性能的影响[J]. 纺织学报, 2013, 34(8): 39-0.
[15]	张海波黄铁军修毅赵野军章江华. 基于神经网络的男西装图像情感语义识别[J]. 纺织学报, 2013, 34(12): 138-0.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

服装类型编号	测试集款式	总数量	训练集数量	测试集数量
1	连衣裙	116	58	58
2	长裤	116	61	55
3	衬衫	127	66	61
4	短袖	141	96	45

服装类型编号	测试集款式	识别率/%
服装类型编号	测试集款式	改进前	改进后
1	连衣裙	74.42	95.35
2	长裤	86.84	100.00
3	衬衫	76.09	93.48
4	短袖	73.33	96.67
	平均值	77.67	96.375

算法	平均识别率/%		平均运行时间/ms
改进前算法	77.670	84
改进后算法	96.375	98