投影约束的遮挡点恢复方法

doi:10.3969/j.issn.0372-2112.2017.11.006

电子学报 ›› 2017, Vol. 45 ›› Issue (11): 2611-2616.DOI: 10.3969/j.issn.0372-2112.2017.11.006

投影约束的遮挡点恢复方法

刘侍刚^1,2,3, 彭亚丽^1,2

1. 现代教学技术教育部重点实验室, 陕西西安 710062;
2. 陕西省教学信息技术工程实验室, 陕西西安 710119;
3. 陕西师范大学计算机科学学院, 陕西西安 710119

收稿日期:2016-11-01 修回日期:2017-05-09 出版日期:2017-11-25 发布日期:2017-11-25
通讯作者: 彭亚丽
作者简介:刘侍刚,男,1973年11月出生于江西峡江,2001年在哈尔滨工程大学获得硕士学位,2005年在西安电子科技大学获得博士学位,现为陕西师范大学教授.从事计算机视觉、三维重建等方面的有关研究.E-mail:shgliu@gmail.com
基金资助:
国家自然科学基金（No.61672333，No.61402274）；陕西省工业科技攻关项目（No.2016GY-081）；陕西省科学研究发展计划项目（No.2016NY-176）；陕西省重点科技创新团队（No.2014KTS-18）；现代教学技术教育部重点实验室学习科学交叉学科培育计划资助

Occlusion Recovery Method Based on Projective Constraints

LIU Shi-gang^1,2,3, PENG Ya-li^1,2

1. Key Laboratory of Modern Teaching Technology, Ministry of Education, Xi'an, Shaanxi 710062;
2. Engineering Laboratory of Teaching Information Technology of Shaanxi Province, Xi'an, Shaanxi 710119;
3. School of Computer Science, Shaanxi Normal University, Xi'an, Shaanxi 710119

Received:2016-11-01 Revised:2017-05-09 Online:2017-11-25 Published:2017-11-25

摘要/Abstract

摘要： 为了恢复图像中的遮挡点，本文在相机为正投影模型下，提出了一种投影约束的遮挡点恢复方法.该方法利用图像矩阵的行空间和列空间都是三维子空间的特性，通过用矩阵奇异值分解分别得到图像矩阵行和列投影满足的约束条件，将遮挡点的求解转化为迭代求解二次型的极值问题.仿真实验和真实实验结果表明，本文方法具有收敛速度快，恢复精度高等优点.

关键词: 遮挡点, 三维子空间, 奇异值分解, 二次函数

Abstract: To recover the position of occlusion,an occlusion recovery method based on projective constraints under orthographic projection is presented.Utilizing the property that both the row and the column space of image matrix are projective subspaces,the method applies Singular Value Decomposition(SVD) to get the image matrix's row and column metric constraints,and replaces occlusion solution by iteratively solving the minimum of a quadratic function.The experiments with both simulated and real data show that the proposed method has the advantages of fast convergence speed and small error.

Key words: occlusion, projective, singular value decomposition, quadratic function

中图分类号:

刘侍刚, 彭亚丽. 投影约束的遮挡点恢复方法[J]. 电子学报, 2017, 45(11): 2611-2616.

LIU Shi-gang, PENG Ya-li. Occlusion Recovery Method Based on Projective Constraints[J]. Acta Electronica Sinica, 2017, 45(11): 2611-2616.

参考文献

[1] R Kennedy,L Balzano,S Wright,C Taylor.Online algorithms for factorization-based structure from motion[J].Computer Vision and Image Understanding,2016,150:139-152.
[2] T Liu,D Tao,M Song,S Maybank.Algorithm-dependent generalization bounds for multi-task learning[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(2):227-241.
[3] C Gong,T Liu,D Tao,K Fu,E Tu,J Yang.Deformed graph Laplacian for semi-supervised learning[J].IEEE Transactions on Neural Networks and Learning Systems.2015,26(10):2261-2274.
[4] T Liu,D Tao.Classification with noisy labels by importance reweighting[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2016,38(3):447-461.
[5] C Tomasi,T Kanade.Shape and motion from image streams under orthography:a factorization method[J].International Journal of Computer Vision,1992,9(2):137-154.
[6] D Jacobs.Linear fitting with missing data for structure-from-motion[J].Computer Vision and Image Understanding,2008,82(1):57-81.
[7] D Strelow.General and nested Wiberg minimization[J].Computer Vision & Pattern Recognition,2012,157(10):1584-1591.
[8] C Gong,D Tao,S Maybank,W Liu,G Kang,J Yang.Multi-modal curriculum learning for semi-supervised image classification[J].IEEE Transactions on Image Processing.2016,25(7):3249-3260.
[9] T Okatani,K Deguchi.On the wiberg algorithm for matrix factorization in the presence of missing components[J].International Journal of Computer Vision,2007,72(3):329-337
[10] C Gong,D Tao,K Fu,J Yang.Fick's law assisted rropagation for semi-supervised learning[J].IEEE Transactions on Neural Networks and Learning Systems,2015,26(9):2148-2162.
[11] C Xu,T Liu,D Tao,C Xu.Local rademacher complexity for multi-label learning[J].IEEE Transactions on Image Processing,2016,25(3):1495-1507.
[12] A Agudo,F Noguer,B Calvo,J Montiel.Sequential non-rigid structure from motion using physical priors[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2016,38(5):979-994.
[13] D Martinec,T Pajdla.Outlier detection for factorization-based reconstruction from perspective images with occlusions[A].In Proceedings of the Photogrammetric Computer Vision 2002(volume B)[C].Inffeldgasse 162.OG,A-8010,2002.161-164.
[14] 刘侍刚,彭亚丽,曾智勇,裘国永.基于正交补空间的遮挡点恢复方法[J].电子学报, 2015,43(5):911-915. LIU Shi-gang,PENG Ya-li,ZENG Zhi-Yong,QIU Guo-yong.An occlusion recovery method based on orthogonal complement subspace[J].Acta Electronica Sinica,2015,43(5):911-915.(in Chinese)
[15] F Noguer,J Porta.A Bayesian approach to simultaneously recover camera pose and non-rigid shape from monocular images[J].Image and Vision Computing,2016,52(C):141-153.
[16] E Candes,T Tao.The power of convex relaxation:Near-optimal matrix completion[J].IEEE Transactions on Information Theory,2010,56(5):2053-2080.
[17] S Lin,J Lin,C Chuang.Particle filter with occlusion handling for visual tracking[J].IET Image Processing,2015,9(11):959-968.
[18] C Gong,K Fu,A Loza,Q Wu,J Liu,J Yang.Pagerank tracker:from ranking to tracking[J].IEEE Transactions on Cybernetics,2014,44(6):882-93.
[19] B Nasihatkon,R Hartley,J Trumpf.A generalized projective reconstruction theorem and depth constraints for projective factorization[J].International Journal of Computer Vision,2015,115(2):87-114.
[20] D Lowe.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision,2004,60(2):91-110.

投影约束的遮挡点恢复方法

Occlusion Recovery Method Based on Projective Constraints

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	廖勇, 杨馨怡, 杜洁汝. 基于两阶段的毫米波大规模MIMO低复杂度混合预编码算法[J]. 电子学报, 2021, 49(7): 1298-1304.
[2]	郑晨, 席晓莉, 宋忠国, 王梦蕾. 基于子空间技术中奇异向量分析的穿墙雷达杂波抑制方法[J]. 电子学报, 2019, 47(4): 848-854.
[3]	王金铭, 叶时平, 徐振宇, 陈超祥, 蒋燕君. 半张量积低存储压缩感知方法研究[J]. 电子学报, 2018, 46(4): 797-804.
[4]	张丹威, 王晓东, 黄国勇. 相关系数SVD增强随机共振的单向阀故障诊断[J]. 电子学报, 2018, 46(11): 2696-2704.
[5]	彭亚丽, 刘侍刚, 孙增国, 洪灵, 曹菡. 迭代二次规划遮挡点恢复[J]. 电子学报, 2018, 46(11): 2733-2737.
[6]	周新平, 孙德刚, 王竹, 欧长海. 一种基于奇异值分解的功耗轨迹筛选方法[J]. 电子学报, 2017, 45(9): 2250-2255.
[7]	赵学智, 叶邦彦. 非零奇异值和频率的关系及其在信号分解中的应用[J]. 电子学报, 2017, 45(8): 2008-2018.
[8]	刘斌, 刘维杰, 罗益辉, 郭琳. 八通道MSVD构造及其在多聚焦图像融合中的应用[J]. 电子学报, 2016, 44(7): 1694-1701.
[9]	王录涛, 吴锡, 金钢, 邹见效. 一种基于奇异值谱加权的超声彩色多普勒成像杂波抑制算法[J]. 电子学报, 2016, 44(6): 1294-1299.
[10]	刘侍刚, 彭亚丽, 曾智勇, 裘国永. 基于正交补空间的遮挡点恢复方法[J]. 电子学报, 2015, 43(5): 911-915.
[11]	张晓帆, 李广军. 基于低硬件复杂度、高速CORDIC的SVD模块设计与实现[J]. 电子学报, 2015, 43(4): 738-742.
[12]	黄霞. 基于本体和奇异值分解的图像数据索引查询算法[J]. 电子学报, 2014, 42(2): 288-291.
[13]	王令赛, 姜淑娟, 张艳梅, 于巧. 基于正交搜索的粒子群优化测试用例生成方法[J]. 电子学报, 2014, 42(12): 2345-2351.
[14]	闫锋刚, 金铭, 乔晓林. 适用任意阵列的变换域二维波达角快速估计算法[J]. 电子学报, 2013, 41(5): 936-942.
[15]	袁晖坪. 泛延拓矩阵的奇异值分解[J]. 电子学报, 2012, 40(8): 1539-1543.