基于纹理与形状的图像对象分割
|
摘要
随着计算机技术的发展,数字图像处理技术己经成为视觉图像研究的主要手段。但是,迄今为止人类还未能构建出一种达到高等生物视觉性能标准的自主视觉系统。其中,有一个重要因素就是生物视觉系统能够很容易完成复杂的图像分割,而机器视觉的分割却很难达到相应的分割速度及精度。因此要使机器视觉技术进一步发展,必须研究快速有效的图像分割技术。
本论文借鉴人类视觉机理,将先验知识和客观信息相结合,主要进行以下三个方面的工作。
1.针对特定类自然图像分割提出了一种新的基于外观模板的图像对象分割方法,主要有三个改进之处,首先引入一种新的分片方法和分片记分方法生成形状片断,使高质量的形状码片产生效率更高,且更具代表性;其次鉴于以前的模型方法中特征提取或者抽象度不够,或者抛弃了太多的图像信息,提出一种新的包含了形状和纹理的对象特征模型,该特征模型具有更加良好的对象类别区分能力,而且能够很大程度容忍不同对象实例的外观差异和成像条件变化;最后将基于密度和层次的快速聚类方法引入到算法体系当中,初步解决多对象图像分割的问题,并且对遮挡对象图像的分割效率也有一定的程度的提高。
2.基于Snake模型的图像分割研究。探讨气球Snake模型和GVF Snake模型以及GVF-Balloon Snake模型的原理,比较分析他们的优点和缺陷。根据GVF-Balloon Snake模型的特点,选取基于外观模型的图像对象分割算法得到的轮廓线作为初始轮廓线,完成对象的精确分割。
3.基于Snake模型的分割方法与基于纹理和形状的图像对象分割方法(T&C-SEG)相结合的组合式分割方法研究。分析自底向上与自顶向下分割各自的优势与不足,根据特定类自然图像特点,设计了两种方式兼用的组合式对象分割方法B&T-IOSEG(Bottom-up and top-down combined image object segmentation)算法。实验表明该算法是有效的,且提高了分割效果。
With the development of computer technology, digital image technology has become the principal mean of the image vision research. But now human beings are still unable to build an autonomous vision system which is able to meet the standards of biological vision. There is one important factor that the biological visual system can easily complete complex image segmentation, on the other hand, the segmentation of machine vision is very difficult to meet the appropriate speed and accuracy. Therefore, for the further development of machine vision technology, it is necessary to study the image segmentation technique to find a rapid and effective method.
In this thesis, combined with the priori knowledge and objective information,and guided by mechanim of human vision, more works has been done as follows: First, the objective of this work is the detection of object classes. The thesis first develops a novel technique to extract class-discriminative boundary fragments and the texture features near the boundary then boosting is used to select discriminative boundary fragments (weak detectors) to form a strong“Boundary-Fragment-Model”detector. A new appearance model is built with those entire detectors and the texture features. And then, the boundary fragment and the texture features and used to complete detection. To the end, a new fast cluster algorithm is used to deal with the centroid image. The generative aspect of the model is used to determine an approximate segmentation. In addition, this thesis presents an extensive evaluation of the new method on a series of test images and compares its performance with the existed methods which are from the literature. As it is shown in the experiment, the new method outperforms previously published methods with the overlap part of the object in multiple-object scene.
And then, this thesis researches the differerce among the three kinds of snake model. According to the characteristics of the snake model, this thesis takes the results of the T&C-SEG method as the initial contour of the snake model. Then, the GVF-Balloon Snake model is choosen to complete the segmentation. Experiments show that this method can enhance the effect of segmentation.
Finally, this thesis analyzes the advantages and disadvantages between bottom-up segmentation method and top-down segmentation method, and researches the modular segmentation method between snake model and the template-based image segmentation method. According to the characteristics of natural image, this thesis develops a combined image segmentation method. Experiments show that the method can significantly enhance the efficiency of segmentation.
本论文借鉴人类视觉机理,将先验知识和客观信息相结合,主要进行以下三个方面的工作。
1.针对特定类自然图像分割提出了一种新的基于外观模板的图像对象分割方法,主要有三个改进之处,首先引入一种新的分片方法和分片记分方法生成形状片断,使高质量的形状码片产生效率更高,且更具代表性;其次鉴于以前的模型方法中特征提取或者抽象度不够,或者抛弃了太多的图像信息,提出一种新的包含了形状和纹理的对象特征模型,该特征模型具有更加良好的对象类别区分能力,而且能够很大程度容忍不同对象实例的外观差异和成像条件变化;最后将基于密度和层次的快速聚类方法引入到算法体系当中,初步解决多对象图像分割的问题,并且对遮挡对象图像的分割效率也有一定的程度的提高。
2.基于Snake模型的图像分割研究。探讨气球Snake模型和GVF Snake模型以及GVF-Balloon Snake模型的原理,比较分析他们的优点和缺陷。根据GVF-Balloon Snake模型的特点,选取基于外观模型的图像对象分割算法得到的轮廓线作为初始轮廓线,完成对象的精确分割。
3.基于Snake模型的分割方法与基于纹理和形状的图像对象分割方法(T&C-SEG)相结合的组合式分割方法研究。分析自底向上与自顶向下分割各自的优势与不足,根据特定类自然图像特点,设计了两种方式兼用的组合式对象分割方法B&T-IOSEG(Bottom-up and top-down combined image object segmentation)算法。实验表明该算法是有效的,且提高了分割效果。
With the development of computer technology, digital image technology has become the principal mean of the image vision research. But now human beings are still unable to build an autonomous vision system which is able to meet the standards of biological vision. There is one important factor that the biological visual system can easily complete complex image segmentation, on the other hand, the segmentation of machine vision is very difficult to meet the appropriate speed and accuracy. Therefore, for the further development of machine vision technology, it is necessary to study the image segmentation technique to find a rapid and effective method.
In this thesis, combined with the priori knowledge and objective information,and guided by mechanim of human vision, more works has been done as follows: First, the objective of this work is the detection of object classes. The thesis first develops a novel technique to extract class-discriminative boundary fragments and the texture features near the boundary then boosting is used to select discriminative boundary fragments (weak detectors) to form a strong“Boundary-Fragment-Model”detector. A new appearance model is built with those entire detectors and the texture features. And then, the boundary fragment and the texture features and used to complete detection. To the end, a new fast cluster algorithm is used to deal with the centroid image. The generative aspect of the model is used to determine an approximate segmentation. In addition, this thesis presents an extensive evaluation of the new method on a series of test images and compares its performance with the existed methods which are from the literature. As it is shown in the experiment, the new method outperforms previously published methods with the overlap part of the object in multiple-object scene.
And then, this thesis researches the differerce among the three kinds of snake model. According to the characteristics of the snake model, this thesis takes the results of the T&C-SEG method as the initial contour of the snake model. Then, the GVF-Balloon Snake model is choosen to complete the segmentation. Experiments show that this method can enhance the effect of segmentation.
Finally, this thesis analyzes the advantages and disadvantages between bottom-up segmentation method and top-down segmentation method, and researches the modular segmentation method between snake model and the template-based image segmentation method. According to the characteristics of natural image, this thesis develops a combined image segmentation method. Experiments show that the method can significantly enhance the efficiency of segmentation.
引文
[1]王润生.图像理解[M].长沙:国防科技大学出版社, 1995: 145-182.
[2]谢昭.图像理解的关键问题和方法研究[D].合肥:合肥工业大学, 2007: 14-24.
[3]章毓晋.基于内容的视觉信息检索[M].北京:科学出版社, 2003: 210-218.
[4] T SERRE, A OLIVA, T POGGIO. A feed for ward theory of visual cortex accounts for human Performance in rapid categorization[J]. Massachusetts Institute of Technology, 2006, 13(2): 20-24.
[5]张鑫.图像对象的组合式分割方法研究[D].长沙:国防科技大学, 2006: 50-58.
[6]林开颜,吴军辉,徐立鸿.彩色图像分割方法综述[J].中国图像图形学报, 2005, 10(1): 1-10.
[7]林开颜,徐立鸿,吴军辉.快速模糊C均值聚类彩色图像分割方法[J].中国图像图形学报, 2004, 9(2): 159-163.
[8]尹平,王润生.基于边缘信息的分开合并图像分割方法[J].中国图像图形学报, 1998, 3(6): 450-454.
[9]潘建江,杨勋年,汪国昭.基于模糊连接度的图像分割及算法[J].软件学报, 2005, 16(1): 67-76.
[10]杨晖,曲秀杰.图像分割方法综述[J].电脑开发与应用. 2005, 18(3): 21-23.
[11]余鹏,封举富,童行伟.一种新的基于高斯混合模型的纹理图像分割方法[J].武汉大学学报:信息科学版, 2005, 30(6): 323-334.
[12]张建国.不变性纹理分析[D].北京:中科院自动化所, 2002: 51-72.
[13]王晗,白雪冰,王辉.基于空间灰度共生矩阵木材纹理分类识别[J].森林工程, 2007, 23(1): 32-36.
[14]余鹏,封举富.基于多分辨率小波和高斯混合模型的纹理图像分割[J].北京大学学报:自然科学版, 2005, 41(3): 338-343.
[15]周昌雄,于盛林.一种图像纹理分割的小波-曲线演化方法[J].信息与控制, 2005, 34(2): 167-171.
[16]陈洪波,徐晓蓉.结基于改进Hough变换的角点提取[J].计算机应用与软件, 2008, 25(2): 57-61.
[17]王洪元,周则明,王平安.一种改进Snake模型的边缘检测算法[J].南京理工大学学报, 2003, 27(4): 395-399.
[18]周昌雄.基于活动轮廓模型的图像分割方法研究[D].南京:南京航空航天大学, 2005: 34-62.
[19] B SUMENGEN, B MANJUNATH. Graph partitioning active contours (GPAC) for image segmentation[J]. Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(4): 509-511.
[20]刘彩霞,范延滨,杨厚俊. GVFsnake模型中一种新的初始轮廓设置方法[J].计算机应用, 2006, 26(7): 114-119.
[21] COHEN I, COHEN L D, AYAEHE N. Using deformable surfaces to segement 3D images and infer differential structure[J]. CVGIP: Image Understanding, 1992, 56: 257-263.
[22]祖克举,周昌雄.基于方向气球力活动轮廓模型的图像分割[J].图像处理, 2006, 22(10): 301-303.
[23]刘苗苗,黎宁,吴园.基于水平集方法的脑部MR图像分割算法[C].第十三届全国图象图形学学术会议, 2006: 275-278.
[24] Y G SHI, W C KARL. Real-time tracking using level set[J], IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2005, 2: 34-41.
[25] C LEE, S HUN, T A KETTER et.al. Unsupervised connectivity-based thresholding segmentation of midsaggittal brain MR images[J], Computers in Biology and medicine, 1998, 28: 309-338.
[26] B APPLETON, H TALBOT. Globally minimal surfaces by continuous maximal flows[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(1): 106-112.
[27] J WINN, N JOJIC. LOCUS: Learning object classes with unsupervised segmentation[C]. Proc. of ICCV’05, Beijing, China, 2005: 756-763.
[28] A LEVIN, Y WEISS. Learning to combine bottom-up and top-down segmentation[C]. Proc. of ECCV’06, LNCS, Graz, Austria, 2006: 581-594.
[29] B K JEON, Y B JUNG, K S HONG. Image segmentation by unsupervised sparse clustering[J].Pattern Recognition Letters. 2006, 27(14): 650-664.
[30]汪荣贵,高隽,张佑生,等.一种新的面向对象的概率图模型[J].计算机研究与发展, 2005, 42(8): 283-292.
[31]朱国普,曾庆双,屈彦呈.一种基于HMRF模型的无监督图像分割算法[J].电子学报, 2006, 34(2): 374-379.
[32]李旭超,朱善安.图像分割中的马尔可夫随机场方法综述[J].中国图像图形学报, 2007, 12(5): 789-798.
[33] E BORENSTEIN, S ULLMAN. Learning to segment[C]. Proc. of ECCV’04. LNCS, Prague, Czechic republic, 2004: 319-324.
[34] S AGARWAL, D ROTH. Learning a sparse representation for object detection[C]. Proc. of ECCV’02, LNCS, Copenhagen, Denmark, 2002: 113-127.
[35] S X YU, J SHI. Object-specific figure-ground segregation[C]. Proc. of CVPR’03, Nice, France, 2003, 2: 39-45.
[36] B LEIBE, A LEONARDIS, B SCHIELE. Combined object categorization and segmentation with an implicit shape model[C]. Proc. of ECCV’04: Workshop on Stat. Learning in Computer Vision, Prague, Czechic republic, 2004: 17-32.
[37] J SHOTTON, A BLAKE, R CIPOLLA. Contour-based learning for object detection[C]. Proc. of ICCV’05, Beijing, China, 2005: 503-510.
[38] G BORGEFORS. HIERARCHICAL. Chamfer matching: a parametric edge matching algorithm[J]. IEEE Trans. On PAMI, 1988, 10(6): 849-865.
[39]沈学华,周志华,吴建鑫,等. Boosting和Bagging综述[J].计算机工程与应用, 2000, 12: 31-40.
[40] A OPELT, A PINZ, A ZISSERMAN. A boundary-fragment-model for object detection[C]. Proc. of ECCV’06. LNCS, Graz, Austria, 2006: 575-588.
[41] Q Z YE. The signed euclidean distance transform and its applications[C]. Proc. Of the 9th International Conference of Pattern Recognition (ICPR’88), Rome, Italy, 1988: 495-499.
[42] R FERGUS, P PERONA, A ZISSERMAN. A sparse object category model for efficient learning and exhaustive recognition[C]. P roc. Proc. CVPR, San Diego, CA, 2005: 253-264.
[43] Y FREUND, R E.SHAPIRE. Experiments with a new boosting algorithm[C]. Proc. of the 13th Conf on Machine Learning, Helsinki, Finland, 1996: 148-156.
[44]景晓军,李剑锋,张苏慧,等.图像智能化的目标检测技术—目标背景视觉模型[J].北京邮电大学学报, 2003, 26: 2-7.
[45]汪倩倩,申忠宇,张亮.基于双目视觉模型的彩色目标实时检测与定位[J].南京师范大学学报:工程技术版, 2008, 8(1): 18-23.
[46]欧阳鑫玉,赵楠楠,宋蕾,等.图像分割技术的发展[J].鞍山钢铁学院学报,2002, 25(5): 127-132.
[47] BIR BHANU, FELLOW, JING PENG. Adaptive integrated image segmentation and object recognition[J]. IEEE Trans.Systems Man and Cybernetics-Part C: Applications and Reviews, 2000, 30(4): 161-178.
[48] G JUN, Z QIN, X XIAOHONG. A novel evolutionary template-matching algorithm and researches on it[J]. Chinese Journal of Electronics, 2006, 15(2): 246-250.
[49] J TANG, S MILLINGTON, S T ACTON. Surface extraction and thickness measurement of the articular cartilage from MR images using directional gradient vector flow snakes[J]. IEEE Transactions on Biomedical Engineering, 2006, 53(5): 896-903.
[50]邱保志,沈钧毅.基于扩展和网格的多密度聚类算法[J].控制与决策, 2006, 21(9): 111-114.
[2]谢昭.图像理解的关键问题和方法研究[D].合肥:合肥工业大学, 2007: 14-24.
[3]章毓晋.基于内容的视觉信息检索[M].北京:科学出版社, 2003: 210-218.
[4] T SERRE, A OLIVA, T POGGIO. A feed for ward theory of visual cortex accounts for human Performance in rapid categorization[J]. Massachusetts Institute of Technology, 2006, 13(2): 20-24.
[5]张鑫.图像对象的组合式分割方法研究[D].长沙:国防科技大学, 2006: 50-58.
[6]林开颜,吴军辉,徐立鸿.彩色图像分割方法综述[J].中国图像图形学报, 2005, 10(1): 1-10.
[7]林开颜,徐立鸿,吴军辉.快速模糊C均值聚类彩色图像分割方法[J].中国图像图形学报, 2004, 9(2): 159-163.
[8]尹平,王润生.基于边缘信息的分开合并图像分割方法[J].中国图像图形学报, 1998, 3(6): 450-454.
[9]潘建江,杨勋年,汪国昭.基于模糊连接度的图像分割及算法[J].软件学报, 2005, 16(1): 67-76.
[10]杨晖,曲秀杰.图像分割方法综述[J].电脑开发与应用. 2005, 18(3): 21-23.
[11]余鹏,封举富,童行伟.一种新的基于高斯混合模型的纹理图像分割方法[J].武汉大学学报:信息科学版, 2005, 30(6): 323-334.
[12]张建国.不变性纹理分析[D].北京:中科院自动化所, 2002: 51-72.
[13]王晗,白雪冰,王辉.基于空间灰度共生矩阵木材纹理分类识别[J].森林工程, 2007, 23(1): 32-36.
[14]余鹏,封举富.基于多分辨率小波和高斯混合模型的纹理图像分割[J].北京大学学报:自然科学版, 2005, 41(3): 338-343.
[15]周昌雄,于盛林.一种图像纹理分割的小波-曲线演化方法[J].信息与控制, 2005, 34(2): 167-171.
[16]陈洪波,徐晓蓉.结基于改进Hough变换的角点提取[J].计算机应用与软件, 2008, 25(2): 57-61.
[17]王洪元,周则明,王平安.一种改进Snake模型的边缘检测算法[J].南京理工大学学报, 2003, 27(4): 395-399.
[18]周昌雄.基于活动轮廓模型的图像分割方法研究[D].南京:南京航空航天大学, 2005: 34-62.
[19] B SUMENGEN, B MANJUNATH. Graph partitioning active contours (GPAC) for image segmentation[J]. Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(4): 509-511.
[20]刘彩霞,范延滨,杨厚俊. GVFsnake模型中一种新的初始轮廓设置方法[J].计算机应用, 2006, 26(7): 114-119.
[21] COHEN I, COHEN L D, AYAEHE N. Using deformable surfaces to segement 3D images and infer differential structure[J]. CVGIP: Image Understanding, 1992, 56: 257-263.
[22]祖克举,周昌雄.基于方向气球力活动轮廓模型的图像分割[J].图像处理, 2006, 22(10): 301-303.
[23]刘苗苗,黎宁,吴园.基于水平集方法的脑部MR图像分割算法[C].第十三届全国图象图形学学术会议, 2006: 275-278.
[24] Y G SHI, W C KARL. Real-time tracking using level set[J], IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2005, 2: 34-41.
[25] C LEE, S HUN, T A KETTER et.al. Unsupervised connectivity-based thresholding segmentation of midsaggittal brain MR images[J], Computers in Biology and medicine, 1998, 28: 309-338.
[26] B APPLETON, H TALBOT. Globally minimal surfaces by continuous maximal flows[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(1): 106-112.
[27] J WINN, N JOJIC. LOCUS: Learning object classes with unsupervised segmentation[C]. Proc. of ICCV’05, Beijing, China, 2005: 756-763.
[28] A LEVIN, Y WEISS. Learning to combine bottom-up and top-down segmentation[C]. Proc. of ECCV’06, LNCS, Graz, Austria, 2006: 581-594.
[29] B K JEON, Y B JUNG, K S HONG. Image segmentation by unsupervised sparse clustering[J].Pattern Recognition Letters. 2006, 27(14): 650-664.
[30]汪荣贵,高隽,张佑生,等.一种新的面向对象的概率图模型[J].计算机研究与发展, 2005, 42(8): 283-292.
[31]朱国普,曾庆双,屈彦呈.一种基于HMRF模型的无监督图像分割算法[J].电子学报, 2006, 34(2): 374-379.
[32]李旭超,朱善安.图像分割中的马尔可夫随机场方法综述[J].中国图像图形学报, 2007, 12(5): 789-798.
[33] E BORENSTEIN, S ULLMAN. Learning to segment[C]. Proc. of ECCV’04. LNCS, Prague, Czechic republic, 2004: 319-324.
[34] S AGARWAL, D ROTH. Learning a sparse representation for object detection[C]. Proc. of ECCV’02, LNCS, Copenhagen, Denmark, 2002: 113-127.
[35] S X YU, J SHI. Object-specific figure-ground segregation[C]. Proc. of CVPR’03, Nice, France, 2003, 2: 39-45.
[36] B LEIBE, A LEONARDIS, B SCHIELE. Combined object categorization and segmentation with an implicit shape model[C]. Proc. of ECCV’04: Workshop on Stat. Learning in Computer Vision, Prague, Czechic republic, 2004: 17-32.
[37] J SHOTTON, A BLAKE, R CIPOLLA. Contour-based learning for object detection[C]. Proc. of ICCV’05, Beijing, China, 2005: 503-510.
[38] G BORGEFORS. HIERARCHICAL. Chamfer matching: a parametric edge matching algorithm[J]. IEEE Trans. On PAMI, 1988, 10(6): 849-865.
[39]沈学华,周志华,吴建鑫,等. Boosting和Bagging综述[J].计算机工程与应用, 2000, 12: 31-40.
[40] A OPELT, A PINZ, A ZISSERMAN. A boundary-fragment-model for object detection[C]. Proc. of ECCV’06. LNCS, Graz, Austria, 2006: 575-588.
[41] Q Z YE. The signed euclidean distance transform and its applications[C]. Proc. Of the 9th International Conference of Pattern Recognition (ICPR’88), Rome, Italy, 1988: 495-499.
[42] R FERGUS, P PERONA, A ZISSERMAN. A sparse object category model for efficient learning and exhaustive recognition[C]. P roc. Proc. CVPR, San Diego, CA, 2005: 253-264.
[43] Y FREUND, R E.SHAPIRE. Experiments with a new boosting algorithm[C]. Proc. of the 13th Conf on Machine Learning, Helsinki, Finland, 1996: 148-156.
[44]景晓军,李剑锋,张苏慧,等.图像智能化的目标检测技术—目标背景视觉模型[J].北京邮电大学学报, 2003, 26: 2-7.
[45]汪倩倩,申忠宇,张亮.基于双目视觉模型的彩色目标实时检测与定位[J].南京师范大学学报:工程技术版, 2008, 8(1): 18-23.
[46]欧阳鑫玉,赵楠楠,宋蕾,等.图像分割技术的发展[J].鞍山钢铁学院学报,2002, 25(5): 127-132.
[47] BIR BHANU, FELLOW, JING PENG. Adaptive integrated image segmentation and object recognition[J]. IEEE Trans.Systems Man and Cybernetics-Part C: Applications and Reviews, 2000, 30(4): 161-178.
[48] G JUN, Z QIN, X XIAOHONG. A novel evolutionary template-matching algorithm and researches on it[J]. Chinese Journal of Electronics, 2006, 15(2): 246-250.
[49] J TANG, S MILLINGTON, S T ACTON. Surface extraction and thickness measurement of the articular cartilage from MR images using directional gradient vector flow snakes[J]. IEEE Transactions on Biomedical Engineering, 2006, 53(5): 896-903.
[50]邱保志,沈钧毅.基于扩展和网格的多密度聚类算法[J].控制与决策, 2006, 21(9): 111-114.