精子多目标跟踪算法的设计与实现
|
摘要
多目标跟踪一直都是图像处理、计算机视觉和模式识别领域里研究的热点。在采集的图像序列中完成多目标跟踪涉及到各帧图像中目标的检测提取、重叠目标的分割、目标特征分析、目标预测与数据关联等内容。随着计算机技术的发展,多目标检测与跟踪的应用越来越广泛,已经渗透到工业、医疗保健、航空航天、军事等各个领域,在国民经济中发挥着越来越大的作用。
计算机辅助精液分析(CASA)是多目标跟踪技术在医学领域的典型应用,它提供了精子动力学的量化数据,并且速度快、简单、快捷、实验结果可重复、较人工精子检测方法具有高精度、数据可靠等优点,正是本课题的研究意义所在。
本文首先介绍了计算机辅助精子分析系统及应用现状,分析了国内外的系统的优势和存在问题,然后从两个方面讨论了几种经典的多目标跟踪算法(区域跟踪法、光流法、核跟踪法、轮廓跟踪法、特征匹配法、卡尔曼滤波和粒子滤波等),阐述了上述算法的基本原理,并分析了各自算法的优缺点和应用范围。
精子的视觉特征和运动信息是精子多目标跟踪的重要指标,本文根据精子视觉特征和运动信息,和实际应用中遇到的各种困难,利用矩量保持和Otsu方法对视频序列进行目标分割,用自适应滤波对运动目标进行预测,然后利用改进的模板匹配和特征匹配的算法对目标进行图像配准,得到了精子目标的运动轨迹。本文对算法进行了仿真实验,实验结果降低计算机辅助精子分析系统的平均识别时间,提高了精确度。最后对本课题研究进行了总结,并指出了下一步研究的主要方向。
Multiple Object detection and tracking related to computer image processing, pattern recognition, artificial intelligence and other areas, is widely applied in many aspects such as military, industrial, life, and so on.
The computer-aided sperm analysis (CASA) is important application of the image analysis technique in the biological and medical area. It provides a rapid and automated assessment of the parameters of sperm motion, together with improved standardization and quality control.
This paper summarizes and classifies two major components of a visual tracking system; target representation and localization and filtering and data association. Target representation and localization is mostly a bottom-up process. Typically the computational complexity for these algorithms is low. Filtering and data association is mostly a top-down process, which involves incorporating prior information about the scene or object, dealing with object dynamics, and evaluation of different hypotheses. The computational complexity for these algorithms is usually much higher.
Examines pros and cons of tracking system and based on the sperm movement characteristics, Otsu and feature matching key technologies, the paper presents an efficient and low-cost method for automatically detecting and tracking the moving object from sperm image sequences. This tracking algorithm was used after the background and extra particles were successfully removed through a two-step enhancement algorithm. The algorithm improves the effectiveness and accuracy of sperm detecting.
计算机辅助精液分析(CASA)是多目标跟踪技术在医学领域的典型应用,它提供了精子动力学的量化数据,并且速度快、简单、快捷、实验结果可重复、较人工精子检测方法具有高精度、数据可靠等优点,正是本课题的研究意义所在。
本文首先介绍了计算机辅助精子分析系统及应用现状,分析了国内外的系统的优势和存在问题,然后从两个方面讨论了几种经典的多目标跟踪算法(区域跟踪法、光流法、核跟踪法、轮廓跟踪法、特征匹配法、卡尔曼滤波和粒子滤波等),阐述了上述算法的基本原理,并分析了各自算法的优缺点和应用范围。
精子的视觉特征和运动信息是精子多目标跟踪的重要指标,本文根据精子视觉特征和运动信息,和实际应用中遇到的各种困难,利用矩量保持和Otsu方法对视频序列进行目标分割,用自适应滤波对运动目标进行预测,然后利用改进的模板匹配和特征匹配的算法对目标进行图像配准,得到了精子目标的运动轨迹。本文对算法进行了仿真实验,实验结果降低计算机辅助精子分析系统的平均识别时间,提高了精确度。最后对本课题研究进行了总结,并指出了下一步研究的主要方向。
Multiple Object detection and tracking related to computer image processing, pattern recognition, artificial intelligence and other areas, is widely applied in many aspects such as military, industrial, life, and so on.
The computer-aided sperm analysis (CASA) is important application of the image analysis technique in the biological and medical area. It provides a rapid and automated assessment of the parameters of sperm motion, together with improved standardization and quality control.
This paper summarizes and classifies two major components of a visual tracking system; target representation and localization and filtering and data association. Target representation and localization is mostly a bottom-up process. Typically the computational complexity for these algorithms is low. Filtering and data association is mostly a top-down process, which involves incorporating prior information about the scene or object, dealing with object dynamics, and evaluation of different hypotheses. The computational complexity for these algorithms is usually much higher.
Examines pros and cons of tracking system and based on the sperm movement characteristics, Otsu and feature matching key technologies, the paper presents an efficient and low-cost method for automatically detecting and tracking the moving object from sperm image sequences. This tracking algorithm was used after the background and extra particles were successfully removed through a two-step enhancement algorithm. The algorithm improves the effectiveness and accuracy of sperm detecting.
引文
[1] Alper Yilmaz,Omar Javed, Mubarak Shah. Object Tracking: A Survey [J], ACM Computing Surveys, 2006, 4(38): Article 13.
[2]于海滨,刘敬彪,刘济林.用于行人头部特征提取的目标区域匹配方法[J].中国图象图形学报,2009,14(3):482-487.
[3]崔一博,谭啸宇,崔鹏,等.视频监控系统中的多摄像头跟踪优化设计[J],哈尔滨工业大学学报,2008,40(9):1485-1487.
[4]苏兴俊,陈景波,香法伟.交通图像序列中的多目标跟踪[J].常熟理工学院学报(自然科学), 2009,23(4):82-86.
[5]谷红亮,史元春,申瑞民,等.一种用于智能空间的多目标跟踪室内定位系统[J].计算机学报,2007,30(9):1603-1605.
[6] Henry Carrillo, Jorge Villarreal, Miguel Sotaquira et al. A Computer Aided Tool for the Assessment of Human Sperm Morphology [J]. 7th IEEE International Conference on Bioinformatics and Bioengineering. . Boston; 2007:1152-1157.
[7] Corkidi G., Taboada B., Wood, C. D., et al. Three-dimensional image acquisition system for multi-sperm tracking [J]. 5th IEEE International Symposium on Biomedical Imaging, Macro: IEEE, 2008:145-148.
[8]肖松山,范世福,李彦芳.计算机辅助精子运动分析系统的研究[J],天津大学学报.2000, 33 (6):811-813.
[9] S.T. Young, W.L. Tzeng, Y.L.Kuo, M.L. Hsiao, S.R. Chiang. Real-time Tracing of Spermatozoa -A System for Improved evaluation of Sperm Progression [J]. IEEE Engineering in medicine and biology, 1996, 39-51.
[10] World Health Organization. WHO laboratory manual for examination of human semen and sperm cervical mucus interaction[S]. Cambridge: Cambridge University press, 1999:81-82.
[11] KRUGER, T.F., COETZEE, K. The role of sperm morphology in assisted reproduction[J], 1999, Human reproduction update, 1999,l 5(2), 172-178.
[12] Hurowitz, E.H., Leung, A., Wang, C. Evaluation of the CellTrak computer-assisted sperm analysis system in comparison to the Cellsoft system to measure human sperm hyperactivation[J]. Fertility and sterility, 1995, 64: 427-432.
[13] J. F. Moruzzi, A. J. Wyrobek, B. H. Mayall, B. L. Gledhill, Quantification and classification of human sperm morphology by computer-assisted image analysis[J],Fertil Steril, 142-152, 1988.
[14] D.E.Kime, K.J.W.Van Look, B.G.McAllister, G.Huyskensb, et al. Computer-assisted sperm analysis (CASA) as a tool for monitoring sperm quality in fish [J]. Comparative Biochemistry and Physiology, 2001, 130: 425-433.
[15] Chanyuth Tretipskul, Kakanang Buranaamnuay, Seri Koonjaenak, et al. The Use of Computer-Assisted Sperm Analysis for Discriminating Series of Motility Pattern of Frozen-thawed Boar Semen [J]. Thai J. Vet. Med. 40(1): 25-30.
[16] J. Verstegen, M. Iguer-Ouada, K. Onclin. Computer assisted semen analyzers in andrology research and veterinary practice [J]. THERIOGENOLOGY, 2002, Volume 57, Issue 1, Pages 149-179.
[17] Yu Huang, Thomas S. Huang, Heinrich Niemann. A Region-Based Method for Model-Free Object Tracking [J]. Proceedings of the 16th International Conference on Pattern Recognition (ICPR'02), 2002, 10592.
[18]邱建国,张建国,李凯.基于Harris与Sift算法的图像匹配方法[J].测试技术学报, 2009,23(3):271-274.
[19]巨西诺,孙继银.一种改进的基于MSER算法的不变量提取算法[J].贵州工业大学学报(自然科学版),2008, 37(6):49-52.
[20] David J. Fleet, Yair Weiss.Optical Flow Estimation. Handbook of Mathematical Models in Computer Vision [J]. Springer. 2006.
[21] John L. Barron, David J. Fleet, Steven Beauchemin. Performance of optical flow techniques [J]. International Journal of Computer Vision (Springer). 1994.
[22] Dorin Comaniciu,Visvanathan Ramesh, Peter Meer. Kernel-Based Object Tracking [J], IEEE Transactions on pattern analysis and machine intelligence, VOL. 25, NO. 5, MAY 2003.
[23] Fukunag, Keinosuke; Larry D. Hostetler. The Estimation of the Gradient of a Density Function with Applications in Pattern Recognition [J]. IEEE Transactions on Information Theory (IEEE), 1975, 21 (1): 32-40.
[24] Cheng, Yizong. Mean Shift, Mode Seeking, and Clustering [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (IEEE), 1995, 17 (8): 790-799.
[25] Dorin Comaniciu, Peter Meer. Mean Shift: A Robust Approach toward Feature Space Analysis
[26] Kass M, Witkin A, Terzopoulos D. Snakes: active contour models [J]. International Journal of Computer Vision, 1988, (1): 321-331.
[27] Gastaud M, Barlaud M, Aubert G. Tracking video objects using active contours [J]. Proceedings of the Workshop on Motion and Video Computing. 2002.90-95.
[28] Paragios N,Deriche R.Geodesic active contours and level sets for the detection and tracking of moving objects[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2000,22(3):266-280
[29] Johnson I Agbinya and David Rees, Multi-Object Tracking in Video [J], Real-Time Imaging 1999, 295-304
[30]付永会,张风超,张宪民.一种改进的基于颜色直方图的实时目标跟踪算法[J].数据采集与处, 2001,16( 3): 309-313.
[31]雷琳,李智勇,粟毅.基于RO I特征匹配融合的图像多目标跟踪算法[J].中国图象图形学报, 2008,13(3): 580-585.
[32]司锡才,陈玉坤,李志刚.数据关联算法的研究[J].哈尔滨工程大学学报, 2007, 28( 7): 813-816.
[33] Greg Welch, Gary Bishop. An Introduction to the Kalman Filter [J]. Los Angeles, SIGGRAPH 2001 Course Notes, 2001.
[34] Gordon N, Salmond D. Novel approach to non-linear and non-Gaussian Bayesian state estimation [J]. P roc of Institute Electric Engineering, 1993, 140 (2): 107-113.
[35] A. Doucet, S. Godsill and C. Andrieu. On sequential Monte Carlo sampling methods for Bayesian filtering. Statistics and Computing, 2000, 10:197-208.
[36]张卫明,张炎华,钟山.蒙特卡罗粒子滤波算法应用研究[J],微计算机信息,2007,23(1)
[37]胡士强,敬忠良.粒子滤波算法综述[J].控制与决策, 2005, 20(4): 364-365.
[38]吴德邦,邹利元.相差显微镜与干涉显微镜的比较[J],泸州医学院学报, 1999, 22(2): 110-113.
[39] Shitu Luo, Qi Zhang, Feilu Luo et al. An improved moment-preserving auto threshold image segmentation algorithm [J]. Proceedings of 2004 International Conference on Information Acquisition. Heifei:IEEE, 2004:316-318.
[40] N Otsu. A Threshold Selection Method from Gray-Level Histograms [J]. IEEE Transactions on Systems, Man, and Cybernetics. 1979, Vol. 9, No. 1:62-66.
[41] Jianlai Wang, Chunling Yang, Min Zhu et al. Implementation of Otsu's thresholding process based on FPGA [J].ICIEA 2009. 4th IEEE Conference on Industrial Electronics and Applications, Xi’an:IEEE, 2009: 479 - 483.
[42]何斌,马天予,王运坚,等.Visual C++数字图像处理(第二版)[M].第二版.北京:人民邮电出版社. 2002:491-499.
[43]匡宇,游志胜,王开平,等.精子运动图像的多目标检测与分割[J].中国图像图学报, 1999,4(5):405-407.
[44]王义敏,秦永元.基于区域生长的SAR图像目标检测方法研究[J],计算机应用, 2009,29(1): 45-46.
[45] A.M. Groenewald, E.C. Botha. Preprocessing and tracking algorithms for automatic sperm analysis [J], Communications and Signal Processing, 1991.64-68.
[2]于海滨,刘敬彪,刘济林.用于行人头部特征提取的目标区域匹配方法[J].中国图象图形学报,2009,14(3):482-487.
[3]崔一博,谭啸宇,崔鹏,等.视频监控系统中的多摄像头跟踪优化设计[J],哈尔滨工业大学学报,2008,40(9):1485-1487.
[4]苏兴俊,陈景波,香法伟.交通图像序列中的多目标跟踪[J].常熟理工学院学报(自然科学), 2009,23(4):82-86.
[5]谷红亮,史元春,申瑞民,等.一种用于智能空间的多目标跟踪室内定位系统[J].计算机学报,2007,30(9):1603-1605.
[6] Henry Carrillo, Jorge Villarreal, Miguel Sotaquira et al. A Computer Aided Tool for the Assessment of Human Sperm Morphology [J]. 7th IEEE International Conference on Bioinformatics and Bioengineering. . Boston; 2007:1152-1157.
[7] Corkidi G., Taboada B., Wood, C. D., et al. Three-dimensional image acquisition system for multi-sperm tracking [J]. 5th IEEE International Symposium on Biomedical Imaging, Macro: IEEE, 2008:145-148.
[8]肖松山,范世福,李彦芳.计算机辅助精子运动分析系统的研究[J],天津大学学报.2000, 33 (6):811-813.
[9] S.T. Young, W.L. Tzeng, Y.L.Kuo, M.L. Hsiao, S.R. Chiang. Real-time Tracing of Spermatozoa -A System for Improved evaluation of Sperm Progression [J]. IEEE Engineering in medicine and biology, 1996, 39-51.
[10] World Health Organization. WHO laboratory manual for examination of human semen and sperm cervical mucus interaction[S]. Cambridge: Cambridge University press, 1999:81-82.
[11] KRUGER, T.F., COETZEE, K. The role of sperm morphology in assisted reproduction[J], 1999, Human reproduction update, 1999,l 5(2), 172-178.
[12] Hurowitz, E.H., Leung, A., Wang, C. Evaluation of the CellTrak computer-assisted sperm analysis system in comparison to the Cellsoft system to measure human sperm hyperactivation[J]. Fertility and sterility, 1995, 64: 427-432.
[13] J. F. Moruzzi, A. J. Wyrobek, B. H. Mayall, B. L. Gledhill, Quantification and classification of human sperm morphology by computer-assisted image analysis[J],Fertil Steril, 142-152, 1988.
[14] D.E.Kime, K.J.W.Van Look, B.G.McAllister, G.Huyskensb, et al. Computer-assisted sperm analysis (CASA) as a tool for monitoring sperm quality in fish [J]. Comparative Biochemistry and Physiology, 2001, 130: 425-433.
[15] Chanyuth Tretipskul, Kakanang Buranaamnuay, Seri Koonjaenak, et al. The Use of Computer-Assisted Sperm Analysis for Discriminating Series of Motility Pattern of Frozen-thawed Boar Semen [J]. Thai J. Vet. Med. 40(1): 25-30.
[16] J. Verstegen, M. Iguer-Ouada, K. Onclin. Computer assisted semen analyzers in andrology research and veterinary practice [J]. THERIOGENOLOGY, 2002, Volume 57, Issue 1, Pages 149-179.
[17] Yu Huang, Thomas S. Huang, Heinrich Niemann. A Region-Based Method for Model-Free Object Tracking [J]. Proceedings of the 16th International Conference on Pattern Recognition (ICPR'02), 2002, 10592.
[18]邱建国,张建国,李凯.基于Harris与Sift算法的图像匹配方法[J].测试技术学报, 2009,23(3):271-274.
[19]巨西诺,孙继银.一种改进的基于MSER算法的不变量提取算法[J].贵州工业大学学报(自然科学版),2008, 37(6):49-52.
[20] David J. Fleet, Yair Weiss.Optical Flow Estimation. Handbook of Mathematical Models in Computer Vision [J]. Springer. 2006.
[21] John L. Barron, David J. Fleet, Steven Beauchemin. Performance of optical flow techniques [J]. International Journal of Computer Vision (Springer). 1994.
[22] Dorin Comaniciu,Visvanathan Ramesh, Peter Meer. Kernel-Based Object Tracking [J], IEEE Transactions on pattern analysis and machine intelligence, VOL. 25, NO. 5, MAY 2003.
[23] Fukunag, Keinosuke; Larry D. Hostetler. The Estimation of the Gradient of a Density Function with Applications in Pattern Recognition [J]. IEEE Transactions on Information Theory (IEEE), 1975, 21 (1): 32-40.
[24] Cheng, Yizong. Mean Shift, Mode Seeking, and Clustering [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (IEEE), 1995, 17 (8): 790-799.
[25] Dorin Comaniciu, Peter Meer. Mean Shift: A Robust Approach toward Feature Space Analysis
[26] Kass M, Witkin A, Terzopoulos D. Snakes: active contour models [J]. International Journal of Computer Vision, 1988, (1): 321-331.
[27] Gastaud M, Barlaud M, Aubert G. Tracking video objects using active contours [J]. Proceedings of the Workshop on Motion and Video Computing. 2002.90-95.
[28] Paragios N,Deriche R.Geodesic active contours and level sets for the detection and tracking of moving objects[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2000,22(3):266-280
[29] Johnson I Agbinya and David Rees, Multi-Object Tracking in Video [J], Real-Time Imaging 1999, 295-304
[30]付永会,张风超,张宪民.一种改进的基于颜色直方图的实时目标跟踪算法[J].数据采集与处, 2001,16( 3): 309-313.
[31]雷琳,李智勇,粟毅.基于RO I特征匹配融合的图像多目标跟踪算法[J].中国图象图形学报, 2008,13(3): 580-585.
[32]司锡才,陈玉坤,李志刚.数据关联算法的研究[J].哈尔滨工程大学学报, 2007, 28( 7): 813-816.
[33] Greg Welch, Gary Bishop. An Introduction to the Kalman Filter [J]. Los Angeles, SIGGRAPH 2001 Course Notes, 2001.
[34] Gordon N, Salmond D. Novel approach to non-linear and non-Gaussian Bayesian state estimation [J]. P roc of Institute Electric Engineering, 1993, 140 (2): 107-113.
[35] A. Doucet, S. Godsill and C. Andrieu. On sequential Monte Carlo sampling methods for Bayesian filtering. Statistics and Computing, 2000, 10:197-208.
[36]张卫明,张炎华,钟山.蒙特卡罗粒子滤波算法应用研究[J],微计算机信息,2007,23(1)
[37]胡士强,敬忠良.粒子滤波算法综述[J].控制与决策, 2005, 20(4): 364-365.
[38]吴德邦,邹利元.相差显微镜与干涉显微镜的比较[J],泸州医学院学报, 1999, 22(2): 110-113.
[39] Shitu Luo, Qi Zhang, Feilu Luo et al. An improved moment-preserving auto threshold image segmentation algorithm [J]. Proceedings of 2004 International Conference on Information Acquisition. Heifei:IEEE, 2004:316-318.
[40] N Otsu. A Threshold Selection Method from Gray-Level Histograms [J]. IEEE Transactions on Systems, Man, and Cybernetics. 1979, Vol. 9, No. 1:62-66.
[41] Jianlai Wang, Chunling Yang, Min Zhu et al. Implementation of Otsu's thresholding process based on FPGA [J].ICIEA 2009. 4th IEEE Conference on Industrial Electronics and Applications, Xi’an:IEEE, 2009: 479 - 483.
[42]何斌,马天予,王运坚,等.Visual C++数字图像处理(第二版)[M].第二版.北京:人民邮电出版社. 2002:491-499.
[43]匡宇,游志胜,王开平,等.精子运动图像的多目标检测与分割[J].中国图像图学报, 1999,4(5):405-407.
[44]王义敏,秦永元.基于区域生长的SAR图像目标检测方法研究[J],计算机应用, 2009,29(1): 45-46.
[45] A.M. Groenewald, E.C. Botha. Preprocessing and tracking algorithms for automatic sperm analysis [J], Communications and Signal Processing, 1991.64-68.