基于区域协方差的图像超像素生成

doi:10.11896/j.issn.1002-137X.2016.05.061

摘要/Abstract

摘要： 图像分割是指将图像分割成一些互不重叠的区域,各区域内部具有相同或相近的某些特定属性,而不同区域之间的属性则相差明显。在图像处理的许多应用中,由于像素级处理的方法因图像数据量庞大、运算规模较大而需要耗费大量的运行时间,因此对图像进行超像素分割预处理是很有必要的一个步骤。基于区域协方差分析,提出了一种新的像素块相似度度量方法；基于像素块相似度度量提出了一种图像超像素生成的鲁棒方法。该方法首先利用K-means算法对输入图像进行初始聚类分割成若干小区域,对每个小区域利用区域协方差矩阵描述其特征信息；然后利用小区域块之间的区域协方差距离来构造相似度矩阵,结合Graph-based与K-means方法对区域块聚类生成图像超像素。与其它方法相比,该方法在生成较紧凑超像素的同时能更好地保持图像边缘特征信息,改善了图像欠分割错误,减少了不必要的过分割现象。将图像超像素生成方法应用于图像风格化中可以快速生成油画风格的风格化图像。

关键词: 图像分割,超像素,区域协方差,聚类,风格化绘制

Abstract: Image segmentation is an important issue on image analysis and understanding,which means that the given image is segmented into some non-overlapping regions and each region has the same or similar intrinsic properties.These intrinsic properties are agreed in the same region,whilst they are different between different regions.In many ima-ge processing applications,due to its large amounts of image pixels,some pixel-based algorithms are always time-consuming and memory-demanding.The image superpixel-based scheme will be an efficient solution to alleviate the storage and time complexities.Based on the analysis of regional covariance,this paper presented a novel similarity measure for image regions and a robust scheme for generating image superpixels.Firstly,the input image is divided into some small regions using K-means algorithm,and then the intrinsic image properties are described by high-dimensional regional covariance matrix.Then,the similarity measure of different regions is determined by the regional covariance distance.Finally,combining the Graph-based scheme and K-means clustering,the final image superpixels are generated.Compared with other superpixel generation approaches,our proposed method is efficient and can reduce some unnecessary over-segmentation.Our algorithm can also keep the image edge information and reduce under-segmentation errors for generating compact image superpixels. The superpixel generation scheme can be applied to the stylized rendering,which will lead to the artistic oil painting.

Key words: Image segmentation,Superpixels,Region covariance,Clustering,Stylized rendering

张旭东,吕言言,缪永伟,杨东勇. 基于区域协方差的图像超像素生成[J]. 计算机科学, 2016, 43(5): 318-323. https://doi.org/10.11896/j.issn.1002-137X.2016.05.061

ZHANG Xu-dong, LV Yan-yan, MIAO Yong-wei and YANG Dong-yong. Regional Covariance Based Image Superpixels Generation[J]. Computer Science, 2016, 43(5): 318-323. https://doi.org/10.11896/j.issn.1002-137X.2016.05.061

参考文献

[1] Gonzalez R C,Woods R E.数字图像处理(第二版)[M].阮秋琦,阮宇智,等译.北京:电子工业出版社,2005
[2] Cao Yu,Zhao Jie,Yan Ji-hong.Image segmentation methodbased on custom color space model [J].Computer Science,2009,36(2):265-267(in Chinese) 曹宇,赵杰,闫继宏.基于自定义颜色空间的快速图像分割算法[J].计算机科学,2009,36(2):265-267
[3] Ren X,Malik J.Learning a classification model for segmentation [C]∥Proceedings of the IEEE International Conference on Computer Vision(ICCV).2003:10-17
[4] Han Shou-dong,Zhao Yong,Tao Wen-bing,et al.Gaussian super-pixel based fast image segmentation using graph cuts[J].Acta Automatica Sinica,2011,37(1):11-20(in Chinese) 韩守东,赵勇,陶文兵,等.基于高斯超像素的快速Graph Cuts图像分割方法[J].自动化学报,2011,37(1):11-20
[5] Zitnick C L,Kang S B.Stereo for image-based rendering using image over-segmentation [J].International Journal of Computer Vision,2007,75(1):49-65
[6] Fulkerson B,Vedaldi A,Soatto S.Class segmentation and object localization with superpixel neighborhoods [C]∥Proceedings of the IEEE International Conference on Computer Vision(ICCV).2009:670-677
[7] Wang S,Lu H,Yang F,et al.Superpixel tracking [C]∥Proceedings of the IEEE International Conference on Computer Vision(ICCV).2011:1323-1330
[8] Lucchi A,Smith K,Achanta R,et al.Superpixel-based segmentation of mitochondria in EM image stacks with learned shape features [J].IEEE Transactions on Medical Imaging,2012,31(2):474-486
[9] Tan Le-yi,Wang Shou-jue.A fast image segmentation based on path-based similarity and dual super-pixel sets[J].Acta Automatica Sinica,2013,39(10):1653-1664(in Chinese) 谭乐怡,王守觉.基于双重超像素集的快速路径相似度图像分割算法[J].自动化学报,2013,39(10):1653-1664
[10] Gong Qu,Fu Yun-feng,Ye Jian-ying,et al.Otsu image segmentation algorithm based on rebuilding of two-dimensional histogram [J].Computer Science,2013,40(8):313-315(in Chinese) 龚劬,付云凤,叶剑英,等.基于二维直方图重建的Otsu图像分割算法[J].计算机科学,2013,40(8):313-315
[11] Gao Shang-bing,Zhou Jing-bo,Yan Yun-yang.A new spuerpixel based spectral clustering for image segmentation [J].Journal of Najing University(Natural Sciences),2013,49(2):169-175(in Chinese) 高尚兵,周静波,严云洋.一种新的基于超像素的谱聚类图像分割算法[J].南京大学学报(自然科学版),2013,49(2):169-175
[12] Luo Hui-lan,Zhong Rui,Kong Fan-sheng.Method of point tra-cking based on superpixel [J].Journal of Image and Graphics,2014,19(3):428-438(in Chinese) 罗会兰,钟睿,孔繁胜.基于超像素的点追踪方法[J].中国图象图形学报,2014,19(3):428-438
[13] Felzenszwalb P F,Huttenlocher D P.Efficient graph-basedimage segmentation [J].International Journal of Computer Vision,2004,59(2):167-181
[14] Shi J,Malik J.Normalized cuts and image segmentation [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2000,22(8):888-905
[15] Moore A P,Prince S,Warrell J,et al.Superpixel lattices [C]∥Proceedings of IEEE Conference on Computer Vision and Pattern Recognition(CVPR).2008:1-8
[16] Wang Ai-qi,Qiu Tian-shuang.Superpixels construction methodbased on geodesic distance [J].Journal of Dalian University of Technology,2012,52(4):610-614(in Chinese) 王爱齐,邱天爽.基于测地距离的超像素生成方法[J].大连理工大学学报,2012,52(4):610-614
[17] Luo Xue-gang,Lv Jun-rui,Wang Hua-jun,et al.Color imagesegmentation based on superpixels and reciprocal nearest neighbors clustering [J].Journal of Guangxi University(Natural Scie-nces),2013,38(2):374-378(in Chinese) 罗学刚,吕俊瑞,王华军,等.基于超像素的互惠最近邻聚类彩色图像分割[J].广西大学学报(自然科学版),2013,38(2):374-378
[18] Vincent L,Soille P.Watersheds in digital spaces:an efficient algorithm based on immersion simulations [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1991,13(6):583-598
[19] Comaniciu D,Meer P.Mean shift:A robust approach towardfeature space analysis [J].IEEE Transactions on Pattern Ana-lysis and Machine Intelligence,2002,24(5):603-619
[20] Vedaldi A,Soatto S.Quick shift and kernel methods for modeseeking [M]∥Computer Vision-ECCV 2008.Berlin:Springer-Verlag,2008:705-718
[21] Achanta R,Shaji A,Smith K,et al.SLIC superpixels compared to state-of-the-art superpixel methods [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(11):2274-2282
[22] Nan Bing-fei,Mu Zhi-chun.SLIC0-based superpixel segmenta-tion method with texture fusion [J].Chinese Journal of Scienti-fic Instrument,2014,35(3):527-534(in Chinese) 南柄飞,穆志纯.基于SLIC0融合纹理信息的超像素分割方法[J].仪器仪表学报,2014,35(3):527-534
[23] Austvoll I,Kwolek B.Region covariance matrix-based objecttracking with occlusions handling [J].Lecture Notes in Compu-ter Science,2010,6374:201-208
[24] Khan S S,Ahmad A.Cluster center initialization algorithm for K-means clustering [J].Pattern recognition Letters,2004,25(11):1293-1302
[25] Von Luxburg U.A tutorial on spectral clustering [J].Statistics and Computing,2007,17(4):395-416
[26] Karacan L,Erdem E,Erdem A.Structure-preserving image smoo-thing via region covariances [J].ACM Transactions on Gra-phics,2013,32(6):176
[27] Liu M Y,Tuzel O,Ramalingam S,et al.Entropy rate superpixel segmentation [C]∥Proceedings of IEEE Conference on Computer Vision and Pattern Recognition(CVPR).2011:2097-2104
[28] Mori G.Guiding model search using segmentation [C]∥Proceedings of the IEEE International Conference on Computer Vision(ICCV).2005:1417-1423
[29] Grundland M,Gibbs C,Dodgson N A.Stylized rendering formultiresolution image representation [C]∥International Society for Optics and Photonics,Electronic Imaging.2005:280-292
[30] DeCarlo D,Santella A.Stylization and abstraction of photo- graphs[J].ACM Transactions on Graphics,2002,21(3):769-776

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed