Journal of Textile Research ›› 2020, Vol. 41 ›› Issue (04): 123-128.doi: 10.13475/j.fzxb.20190600806

• Apparel Engineering • Previous Articles     Next Articles

Image-based three-dimensional garment reconstruction

PAN Bo1, ZHONG Yueqi1,2()   

  1. 1. College of Textiles, Donghua University, Shanghai 201620, China
    2. Key Laboratory of Textile Science & Technology, Ministry of Education, Donghua University, Shanghai 201620, China
  • Received:2019-06-04 Revised:2020-01-15 Online:2020-04-15 Published:2020-04-27
  • Contact: ZHONG Yueqi E-mail:zhyq@dhu.edu.cn

Abstract:

Due to mesh reconstruction error caused by the lack of point cloud during the Poisson reconstruction, a solution strategy that fuse dense point cloud reconstructed in batches by using the nearest iteration algorithm was proposed to restore three-dimensional(3-D) structure. The appropriate number of reconstructed images and shooting schemes were determined by comparing the effect of reconstructed dense point cloud model. The optimal parameter selection of octree depth during Possion reconstruction were analyzed, and the model accuracy based on these strategy was tested. The results indicate that it is sufficient to recover model surface details when octree depth has been set up to 11. The model demonstrates more integrity when the image quantity is greater than 60. Adopting "Hemispherical" shooting scheme has been proven effective in enhancing the model integrity. The error of final 3-D model is less than 5.8 mm by taking the point cloud obtained using depth camera sensor as a benchmark.

Key words: Poisson reconstruction, two-dimensional image, three-dimensional reconstruction, point cloud fusion, nearest iteration algorithm, octree depth, virtual display of garment

CLC Number: 

  • TS942.8

Fig.1

Garment reconstruction result display. (a) Sparse point cloud; (b) Dense point cloud; (c) Mesh; (d) Texture mapping"

Fig.2

Incremental structure from motion pipeline"

Fig.3

Poission reconstruction results by different octree depth"

Fig.4

Mesh reconstruction error. (a) Initial dense point cloud; (b) Reconstructed mesh effect before improvement; (c) Dense point after fusion;(d) Mesh effect after improvement"

Fig.5

Different wearing types during images shooting"

Fig.6

Realization of dense point cloud registration. (a) Dense point cloud before cropping; (b) Dense point cloud after cropping; (c) Reconstructed mesh after fusion;(d) Dense point cloud overall effect after registration"

Fig.7

Shooting scheme. (a) Ring style;(b) Hemisphere style"

Fig.8

Reconstruction effect comparison. (a) Ring/40; (b)Ring/60; (c) Ring/80;(d) Ring/100; (e) Hemisphere/80"

Fig.9

Point cloud comparison on shoulder parts with different shooting scheme. (a) Ring;(b) Hemisphere"

Fig.10

Reconstruction result error distribution"

[1] 董鹏辉, 柯良军. 基于图像的三维重建技术综述[J]. 无线电通信技术, 2019(2):115-119.
DONG Penghui, KE Liangjun. Overview of 3D reconstruction techniques based on images[J]. Radio Communications Technology, 2019(2):115-119.
[2] AGARWAL S, SNAVELY N, SIMON I, et al. Building rome in a day[J]. Communications of the ACM, 2011,54(10):105-112.
[3] SHEN T, ZHU S, FANG T, et al. Graph-based consistent matching for structure-from-motion [C]//European conference on computer vision. Berlin: Springer, 2016: 139-155.
[4] YANG Xiaobo, HUANG Xiubao. Wavelet analysis of fabric surface wrinkle and self-organized neural network grade assessment[J]. Journal of Image & Graphics, 2005,10(4):473-478.
[5] WOODHAM B P. Photometric method for determining surface orientation from multiple images[J]. Optical Engineering, 1980,19(1):1-22.
[6] 沙莎, 蒋高明. 纬编针织物三维模拟技术的研究现状与发展趋势[J]. 纺织学报, 2016,37(11):166-172.
SHA Sha, JIANG Gaoming. Research status and development trend of 3D simulation technology for weft knitted fabric[J]. Journal of Textile Research, 2016,37(11):166-172.
[7] 张家瑞. 多视图3D场景重建算法及应用[D]. 西安:西安电子科技大学, 2018: 19-29.
ZHANG Jiarui. Multiple view 3D reconstruction algorithm and application[D]. Xi'an: Xi'an Xidian University, 2018: 19-29.
[8] 于明, 齐菲菲, 于洋, 等. 基于立体视觉的三维重建算法[J]. 计算机工程与设计, 2013(2):358-361.
YU Ming, QI Feifei, YU Yang, et al. Research of the algorithm for three-dimensional reconstruction based on stereo vision[J]. Computer Engineering and Design, 2013(2):358-361.
[9] 张浩鹏, 魏全茂, 张威, 等. 基于序列图像的空间目标三维重建[J]. 北京航空航天大学学报, 2016,42(2):273-279.
ZHANG Haopeng, WEI Quanmao, ZHANG Wei, et al. Sequential-image-based space object 3D reconstruc-tion[J]. Journal of Beijing University of Aeronautics and Astronautics, 2016,42(2):273-279.
[10] BIANCO S, CIOCCA G, MARELLI D. Evaluating the performance of structure from motion pipelines[J]. Journal of Imaging, 2018,4(8):98.
[11] GOESELE M, SNAVELY N, CURLESS B, et al. Multi-view stereo for community photo collections [C]//International conference on computer vision. New York: IEEE, 2007: 1-8.
[12] FURUKAWA Y, CURLESS B, SEITZ S M, et al. Towards internet-scale multiview stereo [C]//Computer society conference on computer vision and pattern recognition. New York: IEEE, 2010: 1434-1441.
[13] FERRARI V, TUYTELAARS T, GOOL L V. Simultaneous object recognition and segmentation by image exploration [C]//European conference on computer vision. Berlin: Springer, 2004: 40-54.
[14] KAZHDAN M, BOLITHO M, HOPPE H. Poisson surface reconstruction [C]//Eurographics symposium on geometry processing. Goslar: Eurographics Association, 2006: 61-70.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!