基于空间映射复Directionlet变换的图像纹理分类

白静; 贾建华; 焦李成

doi:10.3724/SP.J.1146.2008.00513

基于空间映射复Directionlet变换的图像纹理分类

doi: 10.3724/SP.J.1146.2008.00513

基金项目:

国家863计划项目(2007AA12Z136)，国家973规划项目(2006CB705700)，国家自然科学基金(60672126)，国家教育部博士点基金(20050701013)；教育部长江学者和创新团队支持计划(IRT0645)资助课题

计量
- 文章访问数: 3488
- HTML全文浏览量: 95
- PDF下载量: 1130
- 被引次数: 0
出版历程
- 收稿日期: 2008-04-28
- 修回日期: 2009-01-06
- 刊出日期: 2009-06-19

Texture Classification Based on Mapping Complex Directionlet

摘要

摘要: Directionlet变换具有多方向各向异性基函数，能有效捕捉图像的奇异性特征。该文在此基础上构造了一种空间映射的复Directionlet变换，使其具备了更为灵活的方向选择性和近似的平移不变性。利用空间映射方法获得Directionlet变换的复函数空间，对多尺度各方向子带系数提取能量特征用于图像纹理分类。通过对Brodatz图像库及真实SAR图像的纹理分类实验表明，该文算法较之小波分析及其它多尺度几何分析方法，具有更优的纹理分类性能，也验证了Directionlet工具在图像分析中的应用潜力。
- SAR图像;Brodatz图像;Directionlet变换;空间映射;纹理分类
Abstract: Directionlet transform can capture the image singularity due to possessing the multi-direction anisotropic basis functions. A texture classification algorithm based on the Mapping complex Directionlet Transform (M-DT) is proposed, which provides better directionality and approximate shift invariance. By space mapping for the texture image, then complex Directionlet transform is applied to the mapped image, and the multiscale subband coefficient energy feature is used for texture classification. The experiments using texture images from Brodatz and real SAR images indicate the proposed method outperforms wavelets and Multiscale Geometric Analysis (MGA) approaches, the potential application to image analysis by Directionlet is thus proved.

HTML全文

参考文献(1)

Meyer F G and Coifman R R. Directional image compressionwith Brushlets[C]. Proceedings of the IEEE-SP InternationalSymposium. Paris, France, June 1996: 18-21.[2]Cands E J. Ridgelets: theory and applications[D]. [Ph.D.dissertation]. Department of Statistics, Stanford University,1998.[3]Cands E J and Donoho D L. Curvelets and curvilinearintegrals[R]. Department of Statistics, Stanford University,CA, Tech. Rep., Dec. 1999.[4]Do M N and Vetterli M. The Contourlet transform: Anefficient directional multiresolution image representation[J].IEEE Trans. Image on Processing.2005, 14(12):2091-2106[5]Velisavljevic' V, Beferull-Lozano B, and Vetterli M, et al..Directionlets: anisotropic multi-directional representationwith separable filtering[J]. IEEE Trans. Image on Processing.2006, 15(7): 1916-1933.[6]Velisavljevic' V, Beferull-Lozano B, and Vetterli M, et al..Low-rate reduced complexity image compression usingDirectionlets[C]. IEEE International Conference on ImageProcessing. (ICIP), Atlanta, GA, October 2006: 1601-1604.[7]Velisavljevic' V, Beferull-Lozano B, and Vetterli M. Spacefrequencyquantization for image compression withDirectionlets [J].IEEE Trans. Image on Processing.2007,16(7):1761-1773[8]Jobanputra R and Clausi D A. Preserving boundaries forimage texture segmentation using grey level co-occurringprobabilities[J].Pattern Recognition.2006, 39(2):234-245[9]Fernandes F C A, Van Spaendonck R L C, and Burrus C S.Multidimensional, mapping-based complex wavelettransforms[J].IEEE Trans. Image on Processing.2005, 14 (1):110-124[10]Fernandes F C A. Directional, shift-insensitive, complexwavelet transforms with controllable redundancy[D]. [Ph.D.dissertation]. Rice Univ. Houston, TX. 1, 2002.[11]Simoncelli E P, Freeman W T, and Adelson E H, et al..Shiftable multiscale transforms[J].IEEE Trans. onInformation Theory.1992, 38(2):587-607[12]Tan Shan, Zhang Xiang-rong, and Jiao Li-cheng. ABrushlet-based feature set applied to texture classification[C].CIS. 2004, LNCS 3314: 1175-1180.[13]Hu Ying, Hou Biao, and Wang Shuang, et al.. Textureclassification via stationary-wavelet based Contourlettransform[C]. IWICPAS 2006, 2006, LNCS 4153: 485-494.

施引文献

资源附件(0)

访问统计