基于最优导向法则与距离约束的图像修复算法
|
摘要
为了解决当前图像修复算法利用置信度与数据项来完成图像修复时,忽略了优先修复块中已知信息量所占的比例,导致修复图像存在不连续以及块效应的不足,提出了一种基于最优导向法则耦合距离约束因子的图像修复算法。首先,将像素点的梯度信息引入到待修复块的优先权中,联合置信度与数据项,构造了优先权判定函数,从破损区域中选取优先修复块。以优先修复块中已知信息量所占的比例为依据,构造最优导向法则,对优先修复块中已知信息所占比例进行调整,以找出最佳的匹配块。然后,计算像素点的梯度信息,建立梯度直方图,确定待修复像素点的主方向,通过待修复块内已知像素点与待修复像素点的距离构造主方向上的距离约束因子,以对样本块大小进行动态调整。最后,在像素点之间棋盘距离的约束下,通过对像素点进行均方误差和度量,搜索最优匹配块,从而完成图像修复。实验结果与分析显示,与当前图像修复算法相比,所提算法具有更高的修复视觉质量。
In order to solve the defects as discontinuity and block effect of the repaired image in current image inpainting algorithm, which induced by neglecting the proportion of the known information in the priority repair block. a novel image inpainting method based on optimal guidance rule coupled with distance constraint is proposed in this paper. Firstly, the gradient information of pixels is introduced into the priority measurement of repaired blocks, and the priority decision function is constructed with confidence items and data items to select priority repair blocks. The proportion of the known information in the priority repair block is based on the optimal guidance law, and the proportion of the known information in the priority repair block is optimized. Then, by calculating the gradient information of pixels of the gradient histogram to determine the main direction of the inpainting pixel, the distance constraint factor in the main direction was constructed based on the distance between the known pixels and repaired pixels to dynamically adjust the sample size. Finally, under the constraint of the chessboard distance between pixels, the image restoration is completed by using the methods of the mean square error and measurement of the pixels to search the optimal matching block. The simulation experiment results and analysis show that compared with the current image restoration algorithm, the proposed algorithm has better visual effect.
In order to solve the defects as discontinuity and block effect of the repaired image in current image inpainting algorithm, which induced by neglecting the proportion of the known information in the priority repair block. a novel image inpainting method based on optimal guidance rule coupled with distance constraint is proposed in this paper. Firstly, the gradient information of pixels is introduced into the priority measurement of repaired blocks, and the priority decision function is constructed with confidence items and data items to select priority repair blocks. The proportion of the known information in the priority repair block is based on the optimal guidance law, and the proportion of the known information in the priority repair block is optimized. Then, by calculating the gradient information of pixels of the gradient histogram to determine the main direction of the inpainting pixel, the distance constraint factor in the main direction was constructed based on the distance between the known pixels and repaired pixels to dynamically adjust the sample size. Finally, under the constraint of the chessboard distance between pixels, the image restoration is completed by using the methods of the mean square error and measurement of the pixels to search the optimal matching block. The simulation experiment results and analysis show that compared with the current image restoration algorithm, the proposed algorithm has better visual effect.
引文
[1] JIAOA A M, TSANGA P M. Restoration of digital off-axis fresnel hologram by exemplar and search based image inpainting with enhanced computing speed [J]. Computer Physics Communications, 2015, 8(193): 30-37.
[2] LI S J, YANG X H. Novel image inpainting algorithm based on adaptive fourth-order partial differential equation [J]. IET Image Processing, 2017, 11(10): 870-879.
[3] SI W, GUO W H, ZHU H. Image inpainting using reproducing kernel Hilbert space and heaviside functions[J]. Journal of Computational and Applied Mathematics, 2017, 311(8):551-564.
[4] 李旭峰,王静,刘红敏.特征优先块匹配图像修复算法[J].计算机辅助设计与图形学学报,2016,28(7):1131-1137.LI X F, WANG J, LIU H M. Image inpainting using feature precedence and patch matching [J]. Journal of Computer-Aided Design & Computer Graphics, 2016, 28(7): 1131-1137.
[5] MULLER J S. A coupled variational problem of linear growth related to the denoising and inpainting of images[J]. Journal of Mathematical Sciences, 2017, 224(5): 709-734.
[6] 刘华明,毕学慧,叶中付.样本块搜索和优先权填充的弧形推进图像修复[J].中国图象图形学报,2016,21(8):993-1003.LIU H M, BI X H, YE ZH F. Arc promoting image inpainting using exemplar searching and priority filling [J]. Journal of Image and Graphics, 2016, 21(8): 993-1003.
[7] RAJESH P B, SANJIV V B. Image restoration using prioritized exemplar inpainting with automatic patch optimization [J]. Journal of the Institution of Engineers, 2017, 98(3): 311-319.
[8] VAHID K, FARZIN Y. Fast exemplar-based image inpainting using a new pruning technique [J]. International Journal of Pattern Recognition and Artificial Intelligence, 2017, 31(10):1754-1762.
[9] 梁淑芬,郭敏,梁湘群.改进的Criminisi算法的数字图像修复技术[J].计算机工程与设计, 2016, 37(5): 1314-1319.LIANG SH F, GUO M, LIANG X Q. Enhanced criminisi algorithm of digital image inpianting technology [J]. Computer Engineering and Design, 2016, 37(5):1314-1319.
[10] VADIM F, PABLO A, GABRIELE F. Exemplar-based image inpainting using an affine invariant similarity measure[J].Computer Vision, Imaging and Computer Graphics Theory and Applications, 2017, 693(2): 454-474.
[11] 王文豪,周静波,高尚兵.Criminisi图像修复算法的优化[J]. 现代电子技术, 2017, 40(11): 53-57.WANG W H, ZHOU J B, GAO SH B. Optimization of criminisi algorithm for image inpainting [J]. Modern Electronics Technique, 2017, 40(11): 53-57.
[12] 李玉峰,李广泽,谷绍湖.基于区域分块与尺度不变特征变换的图像拼接算法[J].光学精密工程,2016,24(5):1197-1205.LI Y F, LI G Z, GU SH H. Image mosaic algorithm based on area blocking and SIFT [J]. Optics and Precision Engineering, 2016, 24(5): 1197-1205.
[13] 贾银江,徐哲男,苏中滨.基于优化SIFT算法的无人机遥感作物影像拼接[J].农业工程学报,2017,33(10): 123-129.JIA Y J, XU ZH N, SU ZH B. Mosaic of crop remote sensing images from uav based on improved SIFT algorithm [J]. Transactions of the Chinese Society of Agricultural Engineering, 2017, 33(10): 123-129.
[14] 赵娜,王慧琴,吴萌. 基于马尔科夫随机场匹配准则的Criminisi修复算法[J].计算机科学与探索, 2017, 11(7): 1150-1158.ZHAO N, WANG H Q, WU M. Criminisi digital inpainting algorithm based on markov random field matching criterion [J].Journal of Frontiers of Computer Science and Technology, 2017, 11(7): 1150-1158.
[15] 曾接贤,王璨.基于优先权改进和块划分的图像修复[J].中国图象图形学报,2017,22(9): 1183-1193.ZENG J X, WANG C. Image completion based on redefined priority and image division [J]. Journal of Image and Graphics, 2017, 22(9):1183-1193.
[16] HE K, GAO J Q, LU W X. Image inpainting algorithm based on improved confidence function and matching criterion [J]. Journal of Tianjin University, 2017, 50(4): 399-404.
[17] 李志丹,和红杰,尹忠科. 基于 Curvelet 方向特征的样本块图像修复算法[J].电子学报, 2016, 44(1): 150-154.LI ZH D, HE H J, YIN ZH K. Exemplar based image inpainting algorithm using direction features of curvelet transform [J]. Acta Electronica Sinica, 2016, 44(1): 150-154.
[18] SU X. Image inpainting method based on total variation regularization [J]. Recent Advances in Electrical & Electronic Engineering, 2017, 10(3): 242-247.
[19] SEYED S M. Object removal by depth-wise image inpainting [J]. Signal, Image and Video Processing, 2015, 9(8):1785-1794.
[2] LI S J, YANG X H. Novel image inpainting algorithm based on adaptive fourth-order partial differential equation [J]. IET Image Processing, 2017, 11(10): 870-879.
[3] SI W, GUO W H, ZHU H. Image inpainting using reproducing kernel Hilbert space and heaviside functions[J]. Journal of Computational and Applied Mathematics, 2017, 311(8):551-564.
[4] 李旭峰,王静,刘红敏.特征优先块匹配图像修复算法[J].计算机辅助设计与图形学学报,2016,28(7):1131-1137.LI X F, WANG J, LIU H M. Image inpainting using feature precedence and patch matching [J]. Journal of Computer-Aided Design & Computer Graphics, 2016, 28(7): 1131-1137.
[5] MULLER J S. A coupled variational problem of linear growth related to the denoising and inpainting of images[J]. Journal of Mathematical Sciences, 2017, 224(5): 709-734.
[6] 刘华明,毕学慧,叶中付.样本块搜索和优先权填充的弧形推进图像修复[J].中国图象图形学报,2016,21(8):993-1003.LIU H M, BI X H, YE ZH F. Arc promoting image inpainting using exemplar searching and priority filling [J]. Journal of Image and Graphics, 2016, 21(8): 993-1003.
[7] RAJESH P B, SANJIV V B. Image restoration using prioritized exemplar inpainting with automatic patch optimization [J]. Journal of the Institution of Engineers, 2017, 98(3): 311-319.
[8] VAHID K, FARZIN Y. Fast exemplar-based image inpainting using a new pruning technique [J]. International Journal of Pattern Recognition and Artificial Intelligence, 2017, 31(10):1754-1762.
[9] 梁淑芬,郭敏,梁湘群.改进的Criminisi算法的数字图像修复技术[J].计算机工程与设计, 2016, 37(5): 1314-1319.LIANG SH F, GUO M, LIANG X Q. Enhanced criminisi algorithm of digital image inpianting technology [J]. Computer Engineering and Design, 2016, 37(5):1314-1319.
[10] VADIM F, PABLO A, GABRIELE F. Exemplar-based image inpainting using an affine invariant similarity measure[J].Computer Vision, Imaging and Computer Graphics Theory and Applications, 2017, 693(2): 454-474.
[11] 王文豪,周静波,高尚兵.Criminisi图像修复算法的优化[J]. 现代电子技术, 2017, 40(11): 53-57.WANG W H, ZHOU J B, GAO SH B. Optimization of criminisi algorithm for image inpainting [J]. Modern Electronics Technique, 2017, 40(11): 53-57.
[12] 李玉峰,李广泽,谷绍湖.基于区域分块与尺度不变特征变换的图像拼接算法[J].光学精密工程,2016,24(5):1197-1205.LI Y F, LI G Z, GU SH H. Image mosaic algorithm based on area blocking and SIFT [J]. Optics and Precision Engineering, 2016, 24(5): 1197-1205.
[13] 贾银江,徐哲男,苏中滨.基于优化SIFT算法的无人机遥感作物影像拼接[J].农业工程学报,2017,33(10): 123-129.JIA Y J, XU ZH N, SU ZH B. Mosaic of crop remote sensing images from uav based on improved SIFT algorithm [J]. Transactions of the Chinese Society of Agricultural Engineering, 2017, 33(10): 123-129.
[14] 赵娜,王慧琴,吴萌. 基于马尔科夫随机场匹配准则的Criminisi修复算法[J].计算机科学与探索, 2017, 11(7): 1150-1158.ZHAO N, WANG H Q, WU M. Criminisi digital inpainting algorithm based on markov random field matching criterion [J].Journal of Frontiers of Computer Science and Technology, 2017, 11(7): 1150-1158.
[15] 曾接贤,王璨.基于优先权改进和块划分的图像修复[J].中国图象图形学报,2017,22(9): 1183-1193.ZENG J X, WANG C. Image completion based on redefined priority and image division [J]. Journal of Image and Graphics, 2017, 22(9):1183-1193.
[16] HE K, GAO J Q, LU W X. Image inpainting algorithm based on improved confidence function and matching criterion [J]. Journal of Tianjin University, 2017, 50(4): 399-404.
[17] 李志丹,和红杰,尹忠科. 基于 Curvelet 方向特征的样本块图像修复算法[J].电子学报, 2016, 44(1): 150-154.LI ZH D, HE H J, YIN ZH K. Exemplar based image inpainting algorithm using direction features of curvelet transform [J]. Acta Electronica Sinica, 2016, 44(1): 150-154.
[18] SU X. Image inpainting method based on total variation regularization [J]. Recent Advances in Electrical & Electronic Engineering, 2017, 10(3): 242-247.
[19] SEYED S M. Object removal by depth-wise image inpainting [J]. Signal, Image and Video Processing, 2015, 9(8):1785-1794.