基于学习功能的人体模型表达与实现

doi:10.13475/j.fzxb.20200802605

纺织学报 ›› 2021, Vol. 42 ›› Issue (10): 146-149.doi: 10.13475/j.fzxb.20200802605

基于学习功能的人体模型表达与实现

季勇¹^,², 蒋高明¹^,³()

1.江南大学针织技术教育部工程研究中心, 江苏无锡 214122
2.南通大学杏林学院, 江苏南通 226001
3.生态纺织教育部重点实验室(江南大学), 江苏无锡 214122

收稿日期:2020-08-14 修回日期:2021-06-30 出版日期:2021-10-15 发布日期:2021-10-29
通讯作者: 蒋高明
作者简介:季勇(1987—),男,博士生。主要研究方向为数字化纺织技术。
基金资助:
国家自然科学基金项目(61772238);中央高校基本科研业务费专项资金项目(JUSRP52013B);泰山产业领军人才项目(tscy20180224)

Expression and realization of human body model based on learning model

JI Yong¹^,², JIANG Gaoming¹^,³()

1. Engineering Research Center for Knitting Technology, Ministry of Education, Jiangnan University, Wuxi, Jiangsu 214122, China
2. Xinlin College, Nantong University, Natong, Jiangsu 226001, China
3. Key Laboratory of Eco-Textiles(Jiangnan University), Ministry of Education, Wuxi, Jiangsu 214122, China

Received:2020-08-14 Revised:2021-06-30 Published:2021-10-15 Online:2021-10-29
Contact: JIANG Gaoming

摘要/Abstract

摘要：

针对复杂结构的人体重建问题,提出了一个基于学习功能的人体模型的表达方法,人体线性模型采用校正人体形状的网格顶点算法。根据标准的人体网格创建方法,人体模型的形状由向量代表的平均模板形成,通过学习不同人体形状的回归矩阵,对人体线性模型进行分割和深度估计,隐式地建立人体模型的空间关系,估计人体线性模型中的关节位置,使人体模型的躯干表达更加自然并且清晰。结果表明,基于学习功能的人体模型的表达算法可靠,模型输出结果较为准确,该模型能实现高效的人体模型重构与优化,为人体模型的有效表达提供技术依据和理论参考。

关键词: 学习模型, 人体模型, 三维人体, 网格, 网络结构

Abstract:

In order to solve the problem of human body reconstruction with complex structure, this paper proposes a representation method of human body model based on learning model. The linear model of human body adopts the mesh vertex algorithm to correct the shape of human body. According to the standard method of creating human body mesh, the shape of human body model is formed by the average template represented by vector. By learning the regression matrix of different human body shapes, the linear model of human body is segmented and the depth is estimated. The spatial relationship of human body model is implicitly established, and the joint position in the linear model of human body is estimated, so that the trunk expression of human body model is more natural and clear. The results show that the human model expression algorithm based on learning model is reliable, and the output of the model is more accurate. The model can achieve efficient human model reconstruction and optimization, and provide technical basis and theoretical reference for the effective expression of human model.

Key words: learning model, human body model, three-dimensional human body, grid, network structure

中图分类号:

TS941

季勇, 蒋高明. 基于学习功能的人体模型表达与实现[J]. 纺织学报, 2021, 42(10): 146-149.

JI Yong, JIANG Gaoming. Expression and realization of human body model based on learning model[J]. Journal of Textile Research, 2021, 42(10): 146-149.

图/表 5

图1

图2

图3

表1

图4

参考文献 15

[1]	任洋甫, 普园媛, 徐丹, 等. 语义值驱动的三维直立状态人体模型变形[J]. 系统仿真学报, 2019, 31(3):476-485.
	REN Yangfu, PU Yuanyuan, XU Dan, et al. Semantic value-driven deformation of three-dimensional upright human model[J]. Journal of System Simulation, 2019, 31(3):476-485.
[2]	ZHOU S Z, FU H B, LIU L G, et al. Parametric reshaping of human bodies in images[J]. Acm Transactions on Graphics, 2010, 29(4):10.
[3]	CHEN X W, GUO Y, ZHOU B, et al. Deformable model for estimating clothed and naked human shapes from a single image[J]. Visual Computer, 2013, 29(11):1187-1196. doi: 10.1007/s00371-013-0775-7
[4]	BOISVERT Jonathan, SHU Chang, WUHRER Stefanie, et al. Three-dimensional human shape inference from silhouettes: reconstruction and validation[J]. Machine Vision and Applications, 2013, 24(1):145-157. doi: 10.1007/s00138-011-0353-9
[5]	刘咏梅, 张小雪, 郭云昕. 国内外服装用人体数据库调查与分析[J]. 纺织学报, 2015, 36(6):141-147.
	LIU Yongmei, ZHANG Xiaoxue, GUO Yunxin. Investigation and analysis of human body database for clothing at home and abroad[J]. Journal of Textile Research, 2015, 36(6):141-147.
[6]	修毅, 王银辉. 数字人体模型中腰部剖面曲线参数化变形算法[J]. 纺织学报, 2017, 38(4):97-102.
	XIU Yi, WANG Yinhui. Parametric deformation algorithm of waist profile curve in digital human model[J]. Journal of Textile Research, 2017, 38(4):97-102.
[7]	HUANG Y, SHENG H, ZHENG Y, et al. Deep diff: learning deep difference features on human body parts for person re-identification[J]. Neurocomputing, 2017(12):191-203.
[8]	王佳蕾, 刘奇. 人体个性化模型快速构建及骨骼匹配与运动仿真[J]. 中国医学影像技术, 2014, 30(10):1579-1582.
	WANG Jialei, LIU Qi. Rapid construction of personalized human body model and bone Matching and motion simulation[J]. Chinese Journal of Medical Imaging Technology, 2014, 30(10):1579-1582.
[9]	于瑞云, 苏展, 谢青, 等. 基于空间分割的人体模型骨骼提取算法[J]. 计算机学报, 2019(9):2049-2061.
	YU Ruiyun, SU Zhan, XIE Qing, et al. Skeletal extraction algorithm of human body model based on spatial segmentation[J].Acta Computerica Sinica, 2019(9):2049-2061.
[10]	VAUGHAN N, DUBEY V N, WEE M Y K, et al. Parametric model of human body shape and ligaments for patient-specific epidural simulation[J]. Artificial Intelligence in Medicine, 2014, 62(2):129-140. doi: 10.1016/j.artmed.2014.08.005
[11]	WALTHER T, WUERTZ R P . Unsupervised construction of human body models[J]. Cognitive Systems Research, 2017, 47(1):68-84. doi: 10.1016/j.cogsys.2017.08.001
[12]	曾志超, 李桂清, 邹歆仪, 等. 三维人体模型姿态与形状重构[J]. 计算机辅助设计与图形学学报, 2019, 31(9):1485-1493.
	ZENG Zhichao, LI Guiqing, ZOU Xinyi, et al. Attitude and shape reconstruction of three-dimensional human body model[J]. Journal of Computer-aided Design and Graphics, 2019, 31(9):1485-1493.
[13]	NING G, ZHANG Z, HE Z, et al. Knowledge-guided deep fractal neural networks for human pose estimation[J]. IEEE Transactions on Multimedia, 2018, 20(5):1246-1259. doi: 10.1109/TMM.2017.2762010
[14]	黄新, 马希荣, 赵子平. 一种基于图元方向变形的人体模型生成新算法[J]. 计算机工程与科学, 2017, 39(9):1714-1720.
	HUANG Xin, MA Xirong, ZHAO Ziping. A new algorithm for generating human body model based on graph meta-directional deformation[J]. Computer Engineering and Science, 2017, 39(9):1714-1720.
[15]	VAVALLE N A, DAVIS M L, STIZEL J D, et al. Quantitative validation of a human body finite element model using rigid body impacts.[J]. Annals of Biomedical Engineering, 2015, 43(9):1-12. doi: 10.1007/s10439-014-1225-z

相关文章 15

[1]	季勇蒋高明. 基于学习模型的人体模型表达与实现研究[J]. , 2021, 42(10): 0-0.
[2]	余志才, 钟跃崎, GONGRHugh, 谢昊洋, HUSSAINAzmat. 基于三维悬垂模型和织物面密度的织物匹配[J]. 纺织学报, 2020, 41(10): 46-51.
[3]	尉腾祥, 李敏, 彭虹云, 付少海. 纬平针棉针织物平幅丝光条件与其线圈结构的关系[J]. 纺织学报, 2020, 41(04): 98-105.
[4]	杨恩惠, 邱华, 代文杰. 基于六边形网格结构的针织物三维建模[J]. 纺织学报, 2019, 40(11): 69-74.
[5]	尉腾祥, 李敏, 彭虹云, 付少海. 纬平棉针织物双向拉伸线圈形态分析[J]. 纺织学报, 2019, 40(11): 64-68.
[6]	张芳芳, 段永川. 二维三轴编织复合材料预压单胞模型建立及其弹性规律数值预测[J]. 纺织学报, 2019, 40(10): 85-91.
[7]	王朝晖, 吴雨曦, 杨敏. 基于三维测量数据的个性化女装原型样板生成[J]. 纺织学报, 2019, 40(09): 143-149.
[8]	鲍陈, 缪永伟, 孙瑜亮, 张旭东. 基于散乱点云的三维人体自动测量[J]. 纺织学报, 2019, 40(01): 120-129.
[9]	张雷季祖鹏顾萍萍. 基于结构拓扑优化的绣花机压脚与针杆机构结构优化[J]. 纺织学报, 2018, 39(12): 113-117.
[10]	杨晓波. 基于语义分析的三维服装曲面生成算法[J]. 纺织学报, 2018, 39(11): 163-167.
[11]	赖军张梦莹张华李俊. 消防服衣下空气层的作用与测定方法研究进展[J]. 纺织学报, 2017, 38(06): 151-156.
[12]	姜生倪诗吟张利娟. 锦纶对纺织废胶-受阻酚微观形态与性能的影响[J]. 纺织学报, 2017, 38(03): 72-77.
[13]	梁翠芳陈霞傅婷卜佳仙万贤福汪军. 基于图像处理的网格圈织物孔隙率检测[J]. 纺织学报, 2014, 35(5): 49-0.
[14]	邵光伟蒋金华陈南梁. 柔性经编金属网格材料的切口敏感性[J]. 纺织学报, 2014, 35(3): 46-0.
[15]	张吉生吴志明储云明. 经编无缝女束身上衣的塑身性能[J]. 纺织学报, 2014, 35(3): 109-0.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

试验方法	交并比均值	优化Obj 文件效率/%	精度均值	时间/s
Optimal	49.56	100	47.61	103.2
Baseline-greedy	35.42	69	36.48	60.5
本文方法	65.27	83	62.15	69.1

基于学习功能的人体模型表达与实现

Expression and realization of human body model based on learning model

RichHTML

PDF (PC)

摘要/Abstract

引用本文

使用本文

图/表 5

参考文献 15

相关文章 15

Metrics

本文评价

推荐阅读 0