Research and Implementation of 3D Reconstruction Algorithm for Multi-angle Monocular Garment Image

In order to reduce the difficulty in constructing garment modeling and improve the construction efficiency of garment modeling. In this paper, a method of garment 3D reconstruction based on monocular and multi view is proposed. Firstly, the garment image sequence is obtained, and then the contour information including garment part is obtained by instantiating and segmenting the garment image sequence. The feature points and matching of each image are extracted by SIFT algorithm, and the error matching is eliminated by adding double constraints. Then, sparse point cloud and dense point cloud are reconstructed. Finally, Poisson reconstruction is used to restore the surface details of clothing. The results show that the point cloud noise can be effectively reduced and the reconstruction speed can be accelerated by adding case segmentation and double constraints in the process of garment monocular multi view garment 3D reconstruction. This method can also restore the surface details of clothing in the process of 3D model reconstruction.

[1]  G LoweDavid,et al.  Distinctive Image Features from Scale-Invariant Keypoints , 2004 .

[2]  Jingjun Zhang,et al.  Passive 3D reconstruction based on binocular vision , 2019, International Conference on Graphic and Image Processing.

[3]  王欣 Wang Xin,et al.  Design of binocular vision 3D reconstruction system based on motion recovery , 2014 .

[4]  Yangyang Zhao,et al.  3D Reconstruction of Remote Sensing Image Using Region Growing Combining with CMVS-PMVS , 2016, MUE 2016.

[5]  Graham F. Carey,et al.  A high-order compact formulation for the 3D Poisson equation , 1996 .

[6]  Shanglin Yang,et al.  3D Virtual Garment Modeling from RGB Images , 2019, 2019 IEEE International Symposium on Mixed and Augmented Reality (ISMAR).

[7]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[8]  Wang Jining,et al.  Expanding research on CSG in 3D reconstruction from LiDAR , 2016 .

[9]  Jitendra Malik,et al.  End-to-End Recovery of Human Shape and Pose , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[10]  Jaekyung Kim,et al.  Physical parameter prediction by embedding human perceptual parameter for 3D garment modeling , 2019, 2019 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC).

[11]  Ioannis K. Brilakis,et al.  Automated sparse 3D point cloud generation of infrastructure using its distinctive visual features , 2011, Adv. Eng. Informatics.

[12]  Yong-Jin Liu,et al.  A survey on CAD methods in 3D garment design , 2010, Comput. Ind..

[13]  M. Westoby,et al.  ‘Structure-from-Motion’ photogrammetry: A low-cost, effective tool for geoscience applications , 2012 .

[14]  Ross B. Girshick,et al.  Mask R-CNN , 2017, 1703.06870.

[15]  Lin Wang,et al.  Garment modeling with a depth camera , 2015, ACM Trans. Graph..

[16]  Panagiotis Kyratsis,et al.  A Comparative Study of Open-Source and Licensed CAD Software to Support Garment Development Learning , 2019, Machines.

[17]  Ioannis Mariolis,et al.  Pose and category recognition of highly deformable objects using deep learning , 2015, 2015 International Conference on Advanced Robotics (ICAR).

[18]  V. R. Prakash.  An Enhanced Coding Algorithm for Efficient Video Coding , 2019, Journal of the Institute of Electronics and Computer.