病理切片图像拼接及误差修正技术的研究
|
摘要
在远程医疗会诊中,病理切片是会诊医生诊断的主要依据之一。病理切片的准确性直接决定了诊断结果的正确性,而病理切片的拼接质量又直接影响到病理切片的准确性。目前病理切片还只是基础应用,仅作为诊断参考。就技术而言,这是由于目前图像拼接技术的局限性。它导致在病理切片拼接过程中可能出现细微的拼接误差,而由于误差的存在,极可能导致某些内在因素被忽略,进而造成误诊。因此,研究病理切片的高精度拼接,解决病理切片拼接误差问题,从而提高根据病理切片进行医疗诊断的精确度,具有重要的现实意义。
本文研究了目前图像拼接研究中常用方法和技术,以及图像拼接在远程医疗中的应用。分析并描述了图像拼接中导致误差的因素。针对图像拼接的误差,解释了误差的分类,给出了拼接积累误差期望值的计算公式,从理论上证明了减少积累误差影响的方法的有效性。针对病理切片图像的特点,分析并提出了病理切片图像拼接与一般图像的区别。解释了病理切片图像采集过程;提出了基于SIFT算法的病理切片图像匹配点提取的方法,采用LOF异常检测算法去除了拼接中产生的误匹配点,并通过实验验证了拼接算法的正确性和有效性;针对积累误差,提出了一种基于单元拼接理论的图像拼接修正技术,基于图像拼接后相邻图像的位置关系,提出了拼接参数之间的修正公式,分析并提出了修正的简化规律;实验采用了16张4×4的图像,对其拼接产生的缝隙进行了修正,最终误差控制在一个像素以内。
该修正技术适用于对拼接精度要求高的应用领域。经拼接后可生成高质量、高精度的病理切片图像,不仅能够为远程的医生提高权威的诊疗参考,也可为当地医院研究以往的病理切片提供方便,具有较高的实用价值。
In the process of remote consultation, pathological images analysis is one of the main references for doctor to diagnosis. The accuracy of pathological images will have effect on the accuracy of diagnosis, and which is affected by the stitching quality of the pathological images. However, in current stage, remote pathology consultation is only treated as a reference source. From the technical view, it is due to the limitations of image mosaic technology. During the stitching process, slightly mismatch may be occurred. Because of the possible existing errors, some internal factors may be ignored, and then the patients may be misdiagnosed. Therefore, in order to improve pathological diagnosis based on the pathological images, it is necessary to study the height accurate stitching approach of pathology images.
This paper studies and summarizes the image mosaic theories and methods of image stitching, and the importance for the remote consultation. The errors classification and the nature of stitching errors in the image mosaic are analyzed and described. The expected value of the accumulation of errors is analyzed. According to the expected values of images stitching quality before and after the accumulation of errors, the effectiveness is theoretically proved. Combined with the characteristics of the pathological images, it is presented of the differences of image mosaic between pathological images and common images. The process of collecting pathological images is explained. An image stitching algorithm based on the SIFT is proposed, and then the LOF anomaly detection algorithm is used to remove the mismatching points. The method is tested with pathological images, and the results display that the method is correct and efficient. Focusing on the accumulation errors, an image mosaic correct method is proposed base on the cell mosaic theory. According to the position relationship of the stitched images, modified formulas of stitching parameters are proposed, then the simplify rule is proposed. The method is tested with sixteen images, four by four. The results of correction present that the mismatch errors are restrained little than one pixel.
The technology is suitable for high-accurate situation. It is a high-precision image stitching method, with high practical value. The pathological images with high-accuracy and high-quality have practical value. It not only can enhance the authority of doctors for remote diagnosis and treatment information, but also can provide convenience for the local hospitals to research the previous pathological images.
本文研究了目前图像拼接研究中常用方法和技术,以及图像拼接在远程医疗中的应用。分析并描述了图像拼接中导致误差的因素。针对图像拼接的误差,解释了误差的分类,给出了拼接积累误差期望值的计算公式,从理论上证明了减少积累误差影响的方法的有效性。针对病理切片图像的特点,分析并提出了病理切片图像拼接与一般图像的区别。解释了病理切片图像采集过程;提出了基于SIFT算法的病理切片图像匹配点提取的方法,采用LOF异常检测算法去除了拼接中产生的误匹配点,并通过实验验证了拼接算法的正确性和有效性;针对积累误差,提出了一种基于单元拼接理论的图像拼接修正技术,基于图像拼接后相邻图像的位置关系,提出了拼接参数之间的修正公式,分析并提出了修正的简化规律;实验采用了16张4×4的图像,对其拼接产生的缝隙进行了修正,最终误差控制在一个像素以内。
该修正技术适用于对拼接精度要求高的应用领域。经拼接后可生成高质量、高精度的病理切片图像,不仅能够为远程的医生提高权威的诊疗参考,也可为当地医院研究以往的病理切片提供方便,具有较高的实用价值。
In the process of remote consultation, pathological images analysis is one of the main references for doctor to diagnosis. The accuracy of pathological images will have effect on the accuracy of diagnosis, and which is affected by the stitching quality of the pathological images. However, in current stage, remote pathology consultation is only treated as a reference source. From the technical view, it is due to the limitations of image mosaic technology. During the stitching process, slightly mismatch may be occurred. Because of the possible existing errors, some internal factors may be ignored, and then the patients may be misdiagnosed. Therefore, in order to improve pathological diagnosis based on the pathological images, it is necessary to study the height accurate stitching approach of pathology images.
This paper studies and summarizes the image mosaic theories and methods of image stitching, and the importance for the remote consultation. The errors classification and the nature of stitching errors in the image mosaic are analyzed and described. The expected value of the accumulation of errors is analyzed. According to the expected values of images stitching quality before and after the accumulation of errors, the effectiveness is theoretically proved. Combined with the characteristics of the pathological images, it is presented of the differences of image mosaic between pathological images and common images. The process of collecting pathological images is explained. An image stitching algorithm based on the SIFT is proposed, and then the LOF anomaly detection algorithm is used to remove the mismatching points. The method is tested with pathological images, and the results display that the method is correct and efficient. Focusing on the accumulation errors, an image mosaic correct method is proposed base on the cell mosaic theory. According to the position relationship of the stitched images, modified formulas of stitching parameters are proposed, then the simplify rule is proposed. The method is tested with sixteen images, four by four. The results of correction present that the mismatch errors are restrained little than one pixel.
The technology is suitable for high-accurate situation. It is a high-precision image stitching method, with high practical value. The pathological images with high-accuracy and high-quality have practical value. It not only can enhance the authority of doctors for remote diagnosis and treatment information, but also can provide convenience for the local hospitals to research the previous pathological images.
引文
[1]刘翔,朱士俊,李信春.我国远程医疗发展现状、难点和对策分析[J].中国医院,2005,8(6):102-106.
[2]梁丹.我国远程医疗的影响因素分析及其对策研究[S].华中科技大学,2008(2):1105-1108.
[3]徐庐生,唐慧明.从信息技术看我国远程医疗的发展[J].中国医疗器械信息2006(1):35-39.
[4]冯振杰,任建芳,杜洁.远程医疗对医院建设的意义[J].西藏科技,2002(9):13-19.
[5]韩丽萍,杨永红,陈礼民.一种改进的病理切片图片拼接算法[J].计算机应用与软件, 2003,20(7):56-58.
[6]陈果,左洪福.显微图像的一种快速拼接算法[J].仪器仪表学报, 2003.8,Vol.24(4):343-347.
[7]王国钧,韩丽萍,陈礼民.病理切片图像的拼接技术[J].中国图像图形学报,2003,(10):56-58.
[8] Capel David,et al. Automated mosaicing with super-resolution zoom. //Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition [C]. Jun., 1998:885-891.
[9] Richard Szeliski, Shum,H.Y. Creating Full View Panoramic Image Mosaics and Environment Maps[C], SIGGRAPH’95, 29-37.
[10]孙巧珑,李川.远程医疗会诊病案收集与管理[J].医学影像学杂志,2008 18(9):21-26.
[11]王薇,蔡维生.浅议远程医疗中的法律问题[J].中国卫生事业管理,2006 22(6):5-10.
[12]梁庆丰,孙旭光.远程医疗与法律[J].眼科,2007 16(z1):12-14.
[13]何强,马颂德.图象镶嵌技术理论.难点及应用[J].高技术通讯,1998,3:20-24.
[14] BROWN L G. A survey of image registration techniques [J]. ACM Computing Surveys, 1992, 24(4):325-376.
[15] ZITOVA B, FLUSSER J. Image registration methods: a survey [J]. Image and Vision Computing, 2003, 21(11):977-1000.
[16]解凯,郭恒业,张田文.图像Mosaic技术综述[J].电子学报,2004,32(4):630-633.
[17]刘琼,倪国强,周生兵.图像配准中几种特征点提取方法的分析与实验[J].光学技术,2007,1: 1-33
[18] Jeff Orchard, Craig S. Kaplan Cut-out image mosaics[C]//Proceedings of the 6th international symposium on Non-photorealistic animation and rendering 2008.7.
[19] SZELISKI R.Video mosaics for virtual environments[J].IEEE Computer Graphics and Applications, 1996, 16(2):22-30.
[20]刘贵喜,刘冬梅,刘凤鹏,周亚平.一种稳健的特征点配准算法[J].光学学报,2008 28(3):454-461.
[21]罗诗途,王艳玲,张玘,罗飞路.车载图像配准稳定算法中的关键技术[J].光电工程,2005.5:5-32.
[22]周志强,汪渤,吕冀.不同分辨率图像的角点匹配方法[J].中山大学学报,2008 28(7):1024-1026.
[23]牛小兵等.基于特征的二维像拼接测量法的关键技术研究[J].计量学报,2002,23(2):147-150.
[24] HARRIS C, STEPHENS M. A combined corner and edge detector[C] Proc of the 4th Alvey Vision Conference. Manchester: [s.n], 1988:147-151.
[25]赵文彬,张艳宁.角点检测技术综述[J].计算机应用研究,2006,23(10):17-19.
[26]李志刚,纪玉波,薛全.边界重叠图像的一种快速拼接算法[J].计算机工程,2000,26(5):37-38.
[27]霍红卫,白帆.一种具有精确边界的重复体识别算法[J].科技导报,2008 31(2):214-219
[28]史亮.基于形状配准算法的研究与改进[J].安徽大学学报(自然科学版),2001,25(2):86-90.
[29]刘小军,周越.基于轮廓特征的SAR图像自动配准[J].计算机工程,2007,33(4):176-178.
[30] LI Hui, MANJUNATH B S, MITRA S K, A contour-based approach to multisensory image registration [J]. IEEE Trans on Image Processing, 1995, 4(3):320-334.
[31] BROWN M, LOWE D G. Invariant features from interest point groups[C]. //Proc of Brithish Machine Vision Conference. Cardiff,Wales:[s.n],2002:656-665.
[32] David G. Lowe, Distinctive image features from scale-invariant keypoints[C].//International Journal of Computer Vision, 60, 2 (2004):91-110.
[33]王立峰,潘志庚,石教英.全景图像的拼合及优化算法[J].计算机应用研究,1999.6:59-611.
[34]李智,周文晖,刘济林.动态场景下基于视差空间的立体视觉里程计[J].浙江大学学报,2008,42(10):1661-1665.
[35]钟力,张茂军,孙立峰,李云浩. 360度柱面全景图像生成算法及其实现[J].小型微型计算机系统,1999.12:899-903.
[36] Reddy, B. S.; Chatterji, B. N. An FFT-based technique for translation, rotation, and scale-invariant image registration [J]. IEEE Transactions on Image Processing, vol.5, issue 8,1996(8): 1266-1271.
[37] G. Wolberg and S. Zokai, Robust Image Registration Using Log-Polar Transform [J], in Proc. ICIP, 2000.
[38] KELLER Y.AVERBUCH A. ISAELI M. Pseudo polar based estimation of large translation [J], rotations and scaling in image. Vol.14 Issue.1, 2005(1):12-22.
[39] LIU Han-zhou, Guo Bao-long, FENG Zong-zhe. Pseudolog-polar Fourier transform for image registration [J]. IEEE Signal Processing Letters, 2006, 12(1):17-20.
[40] VIOLA P.WELLS W M.Alignment by maximization of mutual information [C].International Journal of Computer Vision .1997(02).
[41]胡冰,刘衡竹,王攀峰,周海.像素级融合并行算法的模型研究[J].计算机学报,2008(2): 6-8.
[42]李勤,俞信.适合于生物图像的图像融合算法研究[J].光学学报, 2000,20(4):494-500.
[43]吕晓琪,张炜.应用于医学图像的对比度调制融合方法[J].微机发展2004 ,14(12):26-30.
[44]肖李,卢凌,黄红星.基于改IHS变换的图像融合方法[J].武汉理工大学学报,2003(01):254-258.
[45] NI Wei.GUO Bao-long.YANG Liu Data fusion of multisensor remote sensing images using region based Contourlet contrast [J].Journal of Astronautics 2007(02):57-60.
[46]刘旭.光学元件表面疵病检测扫描拼接的误差分析[D].浙江大学.浙江.2008.
[47]李从利,韩裕生,冷晓艳,薛模根.序列图像拼接过程中误匹配的消除[J].计算机应用与软件. 2006 23(11):93-95.
[48] Intan R, Mukaidono M, Generalization of rough sets and its applications in information system [J]. Intelligent Data Analysia. 2006 6:323-339.
[49]丁迎,洪继光.图象拼接中伪匹配的判别和消解[J].中国图象图形学报. 1999 4(10):886-890.
[50]唐琎,谷士文,蔡自兴.图象拼接误差的理论分析[J].小型微型计算机系统。2003 24(12):2185-2189.
[51]李从利,薛模根,冷晓艳,卢伟.空中侦察序列图像连续拼接的累积误差分析[J].中国图象图形学报.2008 13(4). 814-819.
[52]李晓明,郑链,胡占义.基于SIFT特征的遥感影像自动配准[J].遥感学报,2006 10(6):885-892.
[53] Brown M, Lowe D G. Recognising Panoramas[A].//In the Proceedings of the 9th International Conference on Computer Vision (ICCV03)[C].Nice, October,2003.
[54] Lowe D G. Distinctive Image Features from Scale-invariant Key point [J]. International Journal of Computer Vision, 2004 60(2):91-110.
[55] Schaffalitzky F, Zisserman A. Multi-view Matching for Unordered Image Sets[C].//Proceedings of the 7th European Conference on Computer Vision.2002.
[56] Lindebery T. Scale-Space Theory in Computer Vision[M].The Kluwer International Series in Engineering and Computer Science, Kluwer Academy Publishers, Dordrrecht, Netherlands, 1994.
[57] Vassiliadis P, Simitsis A, Terrovitis M. A Generic and Customizable Framework for the Design of ETL Scenarios [J]. Information Systems, 2005, 30(5):492-525.
[58] Galhardas H, Florescu D, Shasha D, etc al. AJAX:An Extensible Data Cleaning Tool[C]//Proc. Of the ACM International Conference on the Management of Data.[S.I.]:ACM Press,2000.
[59]薛惠锋.智能数据挖掘技术[M].西安,西北工业大学出版杜,2005.
[60]孙明竹,赵新,程小燕,卢桂章.基于尺度不变特征变换特征的显微图像在线拼接方法[J].高技术通讯,2009 19(10):1061-1066.
[61]汤井田,王凯,肖嘉莹.基于SIFT特征检测的医学显微图像自动拼接[J].计算机工程与应用,2007,43(35):243-248.
[62]苗立刚,岳永娟,彭思龙.基于2D格状图的显微图像拼接[J].计算机工程,2007,33(12):49-51.
[63]王国钧,罗孝,陈礼民等.一种基于面结构特征的病理切片图像拼接算法[J].中国生物医学工程学报,2004,Vol.23(2):121-126.
[64] Jiaya Jia, Chi-Keung Tang. Image Stitching Using Structure Deformation[C]. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). 2008, 30(4):617-631.
[65] Anat Levin, Assaf Zomet, Shmuel Peleg, et al. Seamless Image Stitching in the Gradient Domain[C]. Proceedings of the European Conference on Computer Vision. 2006, 15(4):969 - 977.
[66] Shum H Y, Szeliski R. Construction and Refinement of Panoramic Mosaics with Global and Local alignment[C]. Proceedings of the 6th International Conference on Computer Vision, Bombay. 1998:953-956.
[67] Sawhney HS, Kumar R. VideoBrush: Experiences with Consumer Video Mosaicing[C] Proceedings of the Workshop on Applications of Computer Vision, Princeton. 1998:386-389.
[68] Brown M, Szeliski R, Winder S. Multi-image Matching Using Multi-scale Oriented Patches[C] Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Diego, USA. 2005:510-517.
[69]苗立刚,岳永娟,彭思龙.显微图像拼接的误差分析[J].小型微型计算机系统,2007 28(7):1255-1258.
[70]丁迎,洪继光.图像拼接中伪匹配的判别和消解[J].中国图象图形学报,1999, 4A(10):86-89.
[2]梁丹.我国远程医疗的影响因素分析及其对策研究[S].华中科技大学,2008(2):1105-1108.
[3]徐庐生,唐慧明.从信息技术看我国远程医疗的发展[J].中国医疗器械信息2006(1):35-39.
[4]冯振杰,任建芳,杜洁.远程医疗对医院建设的意义[J].西藏科技,2002(9):13-19.
[5]韩丽萍,杨永红,陈礼民.一种改进的病理切片图片拼接算法[J].计算机应用与软件, 2003,20(7):56-58.
[6]陈果,左洪福.显微图像的一种快速拼接算法[J].仪器仪表学报, 2003.8,Vol.24(4):343-347.
[7]王国钧,韩丽萍,陈礼民.病理切片图像的拼接技术[J].中国图像图形学报,2003,(10):56-58.
[8] Capel David,et al. Automated mosaicing with super-resolution zoom. //Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition [C]. Jun., 1998:885-891.
[9] Richard Szeliski, Shum,H.Y. Creating Full View Panoramic Image Mosaics and Environment Maps[C], SIGGRAPH’95, 29-37.
[10]孙巧珑,李川.远程医疗会诊病案收集与管理[J].医学影像学杂志,2008 18(9):21-26.
[11]王薇,蔡维生.浅议远程医疗中的法律问题[J].中国卫生事业管理,2006 22(6):5-10.
[12]梁庆丰,孙旭光.远程医疗与法律[J].眼科,2007 16(z1):12-14.
[13]何强,马颂德.图象镶嵌技术理论.难点及应用[J].高技术通讯,1998,3:20-24.
[14] BROWN L G. A survey of image registration techniques [J]. ACM Computing Surveys, 1992, 24(4):325-376.
[15] ZITOVA B, FLUSSER J. Image registration methods: a survey [J]. Image and Vision Computing, 2003, 21(11):977-1000.
[16]解凯,郭恒业,张田文.图像Mosaic技术综述[J].电子学报,2004,32(4):630-633.
[17]刘琼,倪国强,周生兵.图像配准中几种特征点提取方法的分析与实验[J].光学技术,2007,1: 1-33
[18] Jeff Orchard, Craig S. Kaplan Cut-out image mosaics[C]//Proceedings of the 6th international symposium on Non-photorealistic animation and rendering 2008.7.
[19] SZELISKI R.Video mosaics for virtual environments[J].IEEE Computer Graphics and Applications, 1996, 16(2):22-30.
[20]刘贵喜,刘冬梅,刘凤鹏,周亚平.一种稳健的特征点配准算法[J].光学学报,2008 28(3):454-461.
[21]罗诗途,王艳玲,张玘,罗飞路.车载图像配准稳定算法中的关键技术[J].光电工程,2005.5:5-32.
[22]周志强,汪渤,吕冀.不同分辨率图像的角点匹配方法[J].中山大学学报,2008 28(7):1024-1026.
[23]牛小兵等.基于特征的二维像拼接测量法的关键技术研究[J].计量学报,2002,23(2):147-150.
[24] HARRIS C, STEPHENS M. A combined corner and edge detector[C] Proc of the 4th Alvey Vision Conference. Manchester: [s.n], 1988:147-151.
[25]赵文彬,张艳宁.角点检测技术综述[J].计算机应用研究,2006,23(10):17-19.
[26]李志刚,纪玉波,薛全.边界重叠图像的一种快速拼接算法[J].计算机工程,2000,26(5):37-38.
[27]霍红卫,白帆.一种具有精确边界的重复体识别算法[J].科技导报,2008 31(2):214-219
[28]史亮.基于形状配准算法的研究与改进[J].安徽大学学报(自然科学版),2001,25(2):86-90.
[29]刘小军,周越.基于轮廓特征的SAR图像自动配准[J].计算机工程,2007,33(4):176-178.
[30] LI Hui, MANJUNATH B S, MITRA S K, A contour-based approach to multisensory image registration [J]. IEEE Trans on Image Processing, 1995, 4(3):320-334.
[31] BROWN M, LOWE D G. Invariant features from interest point groups[C]. //Proc of Brithish Machine Vision Conference. Cardiff,Wales:[s.n],2002:656-665.
[32] David G. Lowe, Distinctive image features from scale-invariant keypoints[C].//International Journal of Computer Vision, 60, 2 (2004):91-110.
[33]王立峰,潘志庚,石教英.全景图像的拼合及优化算法[J].计算机应用研究,1999.6:59-611.
[34]李智,周文晖,刘济林.动态场景下基于视差空间的立体视觉里程计[J].浙江大学学报,2008,42(10):1661-1665.
[35]钟力,张茂军,孙立峰,李云浩. 360度柱面全景图像生成算法及其实现[J].小型微型计算机系统,1999.12:899-903.
[36] Reddy, B. S.; Chatterji, B. N. An FFT-based technique for translation, rotation, and scale-invariant image registration [J]. IEEE Transactions on Image Processing, vol.5, issue 8,1996(8): 1266-1271.
[37] G. Wolberg and S. Zokai, Robust Image Registration Using Log-Polar Transform [J], in Proc. ICIP, 2000.
[38] KELLER Y.AVERBUCH A. ISAELI M. Pseudo polar based estimation of large translation [J], rotations and scaling in image. Vol.14 Issue.1, 2005(1):12-22.
[39] LIU Han-zhou, Guo Bao-long, FENG Zong-zhe. Pseudolog-polar Fourier transform for image registration [J]. IEEE Signal Processing Letters, 2006, 12(1):17-20.
[40] VIOLA P.WELLS W M.Alignment by maximization of mutual information [C].International Journal of Computer Vision .1997(02).
[41]胡冰,刘衡竹,王攀峰,周海.像素级融合并行算法的模型研究[J].计算机学报,2008(2): 6-8.
[42]李勤,俞信.适合于生物图像的图像融合算法研究[J].光学学报, 2000,20(4):494-500.
[43]吕晓琪,张炜.应用于医学图像的对比度调制融合方法[J].微机发展2004 ,14(12):26-30.
[44]肖李,卢凌,黄红星.基于改IHS变换的图像融合方法[J].武汉理工大学学报,2003(01):254-258.
[45] NI Wei.GUO Bao-long.YANG Liu Data fusion of multisensor remote sensing images using region based Contourlet contrast [J].Journal of Astronautics 2007(02):57-60.
[46]刘旭.光学元件表面疵病检测扫描拼接的误差分析[D].浙江大学.浙江.2008.
[47]李从利,韩裕生,冷晓艳,薛模根.序列图像拼接过程中误匹配的消除[J].计算机应用与软件. 2006 23(11):93-95.
[48] Intan R, Mukaidono M, Generalization of rough sets and its applications in information system [J]. Intelligent Data Analysia. 2006 6:323-339.
[49]丁迎,洪继光.图象拼接中伪匹配的判别和消解[J].中国图象图形学报. 1999 4(10):886-890.
[50]唐琎,谷士文,蔡自兴.图象拼接误差的理论分析[J].小型微型计算机系统。2003 24(12):2185-2189.
[51]李从利,薛模根,冷晓艳,卢伟.空中侦察序列图像连续拼接的累积误差分析[J].中国图象图形学报.2008 13(4). 814-819.
[52]李晓明,郑链,胡占义.基于SIFT特征的遥感影像自动配准[J].遥感学报,2006 10(6):885-892.
[53] Brown M, Lowe D G. Recognising Panoramas[A].//In the Proceedings of the 9th International Conference on Computer Vision (ICCV03)[C].Nice, October,2003.
[54] Lowe D G. Distinctive Image Features from Scale-invariant Key point [J]. International Journal of Computer Vision, 2004 60(2):91-110.
[55] Schaffalitzky F, Zisserman A. Multi-view Matching for Unordered Image Sets[C].//Proceedings of the 7th European Conference on Computer Vision.2002.
[56] Lindebery T. Scale-Space Theory in Computer Vision[M].The Kluwer International Series in Engineering and Computer Science, Kluwer Academy Publishers, Dordrrecht, Netherlands, 1994.
[57] Vassiliadis P, Simitsis A, Terrovitis M. A Generic and Customizable Framework for the Design of ETL Scenarios [J]. Information Systems, 2005, 30(5):492-525.
[58] Galhardas H, Florescu D, Shasha D, etc al. AJAX:An Extensible Data Cleaning Tool[C]//Proc. Of the ACM International Conference on the Management of Data.[S.I.]:ACM Press,2000.
[59]薛惠锋.智能数据挖掘技术[M].西安,西北工业大学出版杜,2005.
[60]孙明竹,赵新,程小燕,卢桂章.基于尺度不变特征变换特征的显微图像在线拼接方法[J].高技术通讯,2009 19(10):1061-1066.
[61]汤井田,王凯,肖嘉莹.基于SIFT特征检测的医学显微图像自动拼接[J].计算机工程与应用,2007,43(35):243-248.
[62]苗立刚,岳永娟,彭思龙.基于2D格状图的显微图像拼接[J].计算机工程,2007,33(12):49-51.
[63]王国钧,罗孝,陈礼民等.一种基于面结构特征的病理切片图像拼接算法[J].中国生物医学工程学报,2004,Vol.23(2):121-126.
[64] Jiaya Jia, Chi-Keung Tang. Image Stitching Using Structure Deformation[C]. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). 2008, 30(4):617-631.
[65] Anat Levin, Assaf Zomet, Shmuel Peleg, et al. Seamless Image Stitching in the Gradient Domain[C]. Proceedings of the European Conference on Computer Vision. 2006, 15(4):969 - 977.
[66] Shum H Y, Szeliski R. Construction and Refinement of Panoramic Mosaics with Global and Local alignment[C]. Proceedings of the 6th International Conference on Computer Vision, Bombay. 1998:953-956.
[67] Sawhney HS, Kumar R. VideoBrush: Experiences with Consumer Video Mosaicing[C] Proceedings of the Workshop on Applications of Computer Vision, Princeton. 1998:386-389.
[68] Brown M, Szeliski R, Winder S. Multi-image Matching Using Multi-scale Oriented Patches[C] Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Diego, USA. 2005:510-517.
[69]苗立刚,岳永娟,彭思龙.显微图像拼接的误差分析[J].小型微型计算机系统,2007 28(7):1255-1258.
[70]丁迎,洪继光.图像拼接中伪匹配的判别和消解[J].中国图象图形学报,1999, 4A(10):86-89.