摘要
为了满足用户处理3D图像的需求,本文提出一种基于直线约束和不同相机间距的立体图像克隆算法,将源立体图像的指定区域通过迭代调整的方式精确克隆到目标立体图像中.首先采用一种轮廓线确定的方法来调整用户绘制的轮廓,然后提出一种基于相机间距和视差的缩放因子,并利用透视缩放函数、直线约束函数、视差约束函数和位置固定函数来调整源立体图像指定区域的大小,以保证精确克隆后立体图像视差的一致,且不发生结构失真.实验结果表明,所提算法能够对不同相机拍摄的立体图像进行精确克隆,显示效果更佳.
To meet the requirement of manipulating stereoscopic images for users,a stereoscopic image cloning algorithm based on linear constraint and different camera spacings was presented in this paper,which precisely clones the specified region of the source stereoscopic image into the target stereoscopic image by iterative adjustment.In order to better guide the image warping,a contour line determination method was adopted to adjust the contour drawn by the user.Then,a scale factor based on camera spacing and disparity was also proposed,and the perspective scaling function,linear constraint function,disparity constraint function and position fixing function were used to adjust the size of the specified region of the source stereoscopic image,ensuring that the precisely cloned stereoscopic images are disparity-matched,and protect the structure.Experimental results indicate that the proposed algorithm can better clone the stereoscopic images shot by different stereo cameras seamlessly.
引文
[1]Feng X,Liang X,Zhang Z.A Cluster Sampling Method for Image Matting via Sparse Coding[M].Germany:Springer International Publishing,2016.
[2]Pérez P,Gangnet M,Blake A.Poisson image editing[J].ACM Transactions on Graphics,2003,22(3):313-318.
[3]Hua M,Bie X,Zhang M,et al.Edge-aware gradient domain optimization framework for image filtering by local propagation[C]//IEEE Conference on Computer Vision and Pattern Recognition.Columbus,Ohio,USA,2014:2838-2845.
[4]Shahrian E,Rajan D,Price B,et al.Improving image matting using comprehensive sampling sets[C]//IEEE Conference on Computer Vision and Pattern Recognition.Oregon,Portland,2013:636-643.
[5]Lo W Y,van Baar J,Knaus C,et al.Stereoscopic 3D copy&paste[J].ACM Transactions on Graphics,2010,29(6):1-10.
[6]Luo S J,Shen I C,Chen B Y,et al.Perspective-aware warping for seamless stereoscopic image cloning[J].ACM Transactions on Graphics,2012,31(6):439-445.
[7]Tong R F,Zhang Y,Cheng K L.Stereopasting:Interactive composition in stereoscopic images[J].IEEE Transations on Visualization&Computer Graphics,2013,19(8):1375-1385.
[8]Lang M,Hornung A,Wang O,et al.Nonlinear disparity mapping for stereoscopic 3D[J].ACM Transactions on Graphics,2010,29(4):157-166.
[9]Yan W,Hou C,Wang B,et al.Content-aware dispari-ty adjustment for different stereo displays[J].Multimedia Tools&Applications,2017,76(8):10465-10479.
[10]Chang C H,Liang C K,Chuang Y Y.Content-aware display adaptation and editing for stereoscopic images[C]//International Conference on Computer Graphics and Interactive Techniques.Vancouver,Canada,2011:30.
[11]Kopf S,Guthier B,Hipp C,et al.Warping-based video retargeting for stereoscopic video[C]//IEEE International Conference on Image Processing.Paris,France,2014:2898-2902.
[12]Li B,Duan L,Lin C W,et al.Region-based depthpreserving stereoscopic image retargeting[C]//IEEE International Conference on Image Processing.Paris,France,2014:2903-2907.
[13]Yan W,Hou C,Zhou Y,et al.Content-and disparity adaptive stereoscopic image retargeting[J].Journal of Modern Optics,2016,63(3):207-218.
[14]Sun D,Roth S,Black M J.Secrets of optical flow estmation and their principles[C]//IEEE Conference on Computer Vision and Pattern Recognition.San Francisco,USA,2010:2432-2439.
[15]Smith B M,Zhang L,Jin H.Stereo matching with nonparametric smoothness priors in feature space[C]//IEEE Conference on Computer Vision and Pattern Recognition.San Francisco,USA,2009:485-492.
[16]Jeon H G,Lee J Y,Im S,et al.Stereo matching with color and monochrome cameras in low-light conditions[C]//IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas,USA,2016:4086-4094.