虚拟试穿中的模型分割与碰撞检测
|
摘要
为解决在计算机辅助服装设计中虚拟试穿的计算速度与实时效率瓶颈,提出了一种基于Span的三维人体模型区域生长优化分割算法,在区段结构上适当添加标记,增加区段信息,避免不必要的重复访问与回溯,提高运算效率;提出了基于几何位置的衣物三维立体模型与人体模型碰撞检测方法,使用椭圆包围盒算法来拟合人体模型分割区域,检测到衣物质点在与椭圆包围盒发生碰撞后,修正其到椭圆包围盒表面正确的碰撞位置上,以便通过实时仿真来观测虚拟设计的结果.仿真实验结果表明:衣物的虚拟试穿模拟效果逼真,与传统方法相比,碰撞检测处理时间减少了64%,具有一定的鲁棒性与实效性.
In virtual try-on, to solve the bottleneck of computational speed and real-time efficiency, an optimized algorithm is presented based on Span of range growing method to segment the 3D human body model. By adding tags to the section structure, this method avoids unnecessary backtracking, and improves the computational efficiency. A new method about collision detection of human body model and 3D cloth model is presented based on geometric position. It uses the ellipsoid surround box algorithm to fit the segmentation area of human body model. After detecting the clothing particle has a collision with the ellipsoid surround box, the correct collision position is fixed on the surface of the ellipsoid surround box, and the virtual design effects are viewed through the real-time simulation. The simulation results show that the virtual try-on of clothing is realistic, and the time of collision detection is reduced 64% compared with the traditional method. It demonstrates the robustness and timeliness.
In virtual try-on, to solve the bottleneck of computational speed and real-time efficiency, an optimized algorithm is presented based on Span of range growing method to segment the 3D human body model. By adding tags to the section structure, this method avoids unnecessary backtracking, and improves the computational efficiency. A new method about collision detection of human body model and 3D cloth model is presented based on geometric position. It uses the ellipsoid surround box algorithm to fit the segmentation area of human body model. After detecting the clothing particle has a collision with the ellipsoid surround box, the correct collision position is fixed on the surface of the ellipsoid surround box, and the virtual design effects are viewed through the real-time simulation. The simulation results show that the virtual try-on of clothing is realistic, and the time of collision detection is reduced 64% compared with the traditional method. It demonstrates the robustness and timeliness.
引文
[1]BORSCHE R,KLAR A,MEURER A,et al.Mean field models for interacting ellipsoidal particles[J].Computers and Mathematics with Applications,2016,72(3):704-719.
[2]孙劲光,吴素红.基于空间分割与椭球包围盒的碰撞检测算法[J].计算机工程与应用,2016,52(4):217-222.SUN J G,WU S H.Collision detection algorithm based on ellipsoid bounding box and spatial decomposition[J].Computer Engineering and Applications,2016,52(4):217-222(in Chinese).
[3]LIU J Z,LI J T,LU G D.Deformation similarity clustering based collision detection in clothing simulation[J].International Journal of Clothing Science&Technology,2014,26(5):395-411.
[4]ZHANG Y Z,ZHENG J M.Cloth Simulation and Virtual Tryon with Kinect Based on Human Body Adaptation[M].Singapore:Springer,2014:31-50.
[5]李国财,高振宇,蔡政.虚拟试衣系统研究与实现[J].计算机应用与软件,2014,31(12):206-209,241.LI G C,GAO Z Y,CAI Z.Research and implementation of virtual fitting system[J].Computer Applications and Software,2014,31(12):206-209,241(in Chinese).
[6]黎琳,韩丽,徐建国,等.融合语义与几何特征的人体模型结构分割[J].计算机应用研究,2016,33(1):316-320.LI L,HAN L,XU J G,et al.Structural segmentation of human models based on semantic and geometric characteristics[J].Application Research of Computers.2016,33(1):316-320(in Chinese).
[7]王华松.一种基于种子点区域生长的彩色图象分割算法[D].郑州:河南大学,2015.WANG H S.A color image segmentation algorithm based on seed region growing[D].Zhenzhou:Henan University,2015(in Chinese).
[8]甘凌云,侯进.一种基于模型分割的三维人体骨架提取方法[J].计算机应用与软件,2016,33(9):33-37.GAN L Y,HOU J.A method for extracting 3D human body skeleton based on model segmentation[J].Computer Applications and Software,2016,33(9):33-37(in Chinese).
[9]YUAN Y W,CHEN D,YAN L M.Interactive three-dimensional segmentation using region growing algorithms[J].Journal of Algorithms&Computational Technology,2015,9(2):199-213.
[10]陈河军.融合边缘与区域生长的彩色图像分割算法研究[D].杭州:浙江工业大学,2015.CHEN H J.Research on algorithm of color image segmentation integrating edge and region growing[D].Hangzhou:Zhejiang University of Technology,2015(in Chinese).
[11]张翔宇,田庆国,葛宝臻,等.三维人体点云模型多约束肢体分割[J].计算机应用,2015,35(3):830-834.ZHANG X Y,TIAN Q G,GE B Z,et al.Multi-constrained segmentation of 3D human point-cloud[J].Journal of Computer Applications,2015,35(3):830-834(in Chinese).
[12]LU L,CHOI Y K,WANG W P,et al.Variational 3D shape segmentation for bounding volume computation[J].Computer Graphics Forum,2007,26(3):329-38.
[13]梁秀霞,韩慧健,张彩明.基于物理仿真的布料动画研究综述[J].计算机研究与发展,2014,51(1):31-40.LIANG X X,HAN H J,ZHANG C M.Survey of cloth animation based on physical simulation[J].Computer Research and Development,2014,51(1):31-40(in Chinese).
[14]裴珊珊.自适应多层次网格布料仿真及碰撞处理[D].秦皇岛:燕山大学,2014.PEI S S.Cloth simulation using adaptive multilevel meshes and collision handing[D].Qinhuangdao:Yanshan University,2014(in Chinese).
[15]刘志华.自碰撞检测高层剔除算法研究与实现[D].杭州:浙江大学,2017.LIU Z Z.Reseatch and implementation of high-level culling algorithm for self-collision detection[D].Hangzhou:Zhejiang University,2017(in Chinese).
[16]YANG F,ZHOU F,WANG R M,et al.A fast and efficient mesh segmentation method based on improved region growing[J].Applied Mathematics:A Journal of Chinese Universities:Series B,2014,29(4):468-480.
[17]HAMAD M,THOMASSEY S,BRUNIAUX P.A new sizing system based on 3D morphology clustering[J].Computers&Industrial Engineering,2017,113(5):683-692.
[18]KIM D E.Psychophysical testing of garment size variation using three-dimensional virtual try-on technology[J].Textile Research Journal,2016,86(4):365-379.
[19]唐勇,杨偲偲,吕梦雅,等.自适应椭球包围盒改进织物碰撞检测方法[J].计算机辅助设计与图形学学报,2013,25(10):1589-1596.TANG Y,YANG S S,LYU M Y,et al.Collision detection for cloth based on adaptive enclosing ellipsoids[J].Journal of Computer-Aided Design&Computer Graphics,2013,25(10):1589-1596(in Chinese).
[20]梁成.基于最近点碰撞检测算法的研究[D].成都:西华大学,2016.LIANG C.Research of collision detection algorithm based on closest points[D].Chengdu:Xihua University,2016(in Chinese).
[2]孙劲光,吴素红.基于空间分割与椭球包围盒的碰撞检测算法[J].计算机工程与应用,2016,52(4):217-222.SUN J G,WU S H.Collision detection algorithm based on ellipsoid bounding box and spatial decomposition[J].Computer Engineering and Applications,2016,52(4):217-222(in Chinese).
[3]LIU J Z,LI J T,LU G D.Deformation similarity clustering based collision detection in clothing simulation[J].International Journal of Clothing Science&Technology,2014,26(5):395-411.
[4]ZHANG Y Z,ZHENG J M.Cloth Simulation and Virtual Tryon with Kinect Based on Human Body Adaptation[M].Singapore:Springer,2014:31-50.
[5]李国财,高振宇,蔡政.虚拟试衣系统研究与实现[J].计算机应用与软件,2014,31(12):206-209,241.LI G C,GAO Z Y,CAI Z.Research and implementation of virtual fitting system[J].Computer Applications and Software,2014,31(12):206-209,241(in Chinese).
[6]黎琳,韩丽,徐建国,等.融合语义与几何特征的人体模型结构分割[J].计算机应用研究,2016,33(1):316-320.LI L,HAN L,XU J G,et al.Structural segmentation of human models based on semantic and geometric characteristics[J].Application Research of Computers.2016,33(1):316-320(in Chinese).
[7]王华松.一种基于种子点区域生长的彩色图象分割算法[D].郑州:河南大学,2015.WANG H S.A color image segmentation algorithm based on seed region growing[D].Zhenzhou:Henan University,2015(in Chinese).
[8]甘凌云,侯进.一种基于模型分割的三维人体骨架提取方法[J].计算机应用与软件,2016,33(9):33-37.GAN L Y,HOU J.A method for extracting 3D human body skeleton based on model segmentation[J].Computer Applications and Software,2016,33(9):33-37(in Chinese).
[9]YUAN Y W,CHEN D,YAN L M.Interactive three-dimensional segmentation using region growing algorithms[J].Journal of Algorithms&Computational Technology,2015,9(2):199-213.
[10]陈河军.融合边缘与区域生长的彩色图像分割算法研究[D].杭州:浙江工业大学,2015.CHEN H J.Research on algorithm of color image segmentation integrating edge and region growing[D].Hangzhou:Zhejiang University of Technology,2015(in Chinese).
[11]张翔宇,田庆国,葛宝臻,等.三维人体点云模型多约束肢体分割[J].计算机应用,2015,35(3):830-834.ZHANG X Y,TIAN Q G,GE B Z,et al.Multi-constrained segmentation of 3D human point-cloud[J].Journal of Computer Applications,2015,35(3):830-834(in Chinese).
[12]LU L,CHOI Y K,WANG W P,et al.Variational 3D shape segmentation for bounding volume computation[J].Computer Graphics Forum,2007,26(3):329-38.
[13]梁秀霞,韩慧健,张彩明.基于物理仿真的布料动画研究综述[J].计算机研究与发展,2014,51(1):31-40.LIANG X X,HAN H J,ZHANG C M.Survey of cloth animation based on physical simulation[J].Computer Research and Development,2014,51(1):31-40(in Chinese).
[14]裴珊珊.自适应多层次网格布料仿真及碰撞处理[D].秦皇岛:燕山大学,2014.PEI S S.Cloth simulation using adaptive multilevel meshes and collision handing[D].Qinhuangdao:Yanshan University,2014(in Chinese).
[15]刘志华.自碰撞检测高层剔除算法研究与实现[D].杭州:浙江大学,2017.LIU Z Z.Reseatch and implementation of high-level culling algorithm for self-collision detection[D].Hangzhou:Zhejiang University,2017(in Chinese).
[16]YANG F,ZHOU F,WANG R M,et al.A fast and efficient mesh segmentation method based on improved region growing[J].Applied Mathematics:A Journal of Chinese Universities:Series B,2014,29(4):468-480.
[17]HAMAD M,THOMASSEY S,BRUNIAUX P.A new sizing system based on 3D morphology clustering[J].Computers&Industrial Engineering,2017,113(5):683-692.
[18]KIM D E.Psychophysical testing of garment size variation using three-dimensional virtual try-on technology[J].Textile Research Journal,2016,86(4):365-379.
[19]唐勇,杨偲偲,吕梦雅,等.自适应椭球包围盒改进织物碰撞检测方法[J].计算机辅助设计与图形学学报,2013,25(10):1589-1596.TANG Y,YANG S S,LYU M Y,et al.Collision detection for cloth based on adaptive enclosing ellipsoids[J].Journal of Computer-Aided Design&Computer Graphics,2013,25(10):1589-1596(in Chinese).
[20]梁成.基于最近点碰撞检测算法的研究[D].成都:西华大学,2016.LIANG C.Research of collision detection algorithm based on closest points[D].Chengdu:Xihua University,2016(in Chinese).