GVF-Snake模型的改进及其流水线处理方法
摘要
图像分割是计算机视觉与模式识别领域的一个重要研究方向。轮廓的提取在各个方面的作用越来越重要,因此这方面的研究也逐渐增多。主动轮廓模型(Active Contour Model,也称为Snake model)是一种全新的图像分割和目标提取方法,它充分利用高层信息自上而下来提取感兴趣的目标或特性。GVF Snake模型作为传统Snake模型的一个改进,是近年来图像理解领域的一个研究热点。
本文首先介绍了图像分割的基本概念,基于其分割原理以及数学模型的不同把图像分割分成两类:基于数据驱动的分割和基于模型驱动的分割。本文研究的GVF Snake属于基于模型驱动的图像分割方法;
接着详细分析了传统Snake方法和及其改进模型GVF Snake方法的优缺点;针对GVF Snake需要人工设置初始轮廓线的问题,本文改进OSTU(大津法)产生GVF Snake模型所需的初始轮廓曲线,新生成的初始轮廓曲线往往比较接近目标轮廓,当GVF Snake模型使用此曲线搜索时,往往能更快、更好的分割目标,并且使用OSTU方法实现了分割的自动化;
最后,针对GVF Snake模型分割目标费时的特点,本文使用VC++实现整个GVF‐Snake模型,结合Intel最新的并行处理技术,把流水线应用于GVF Snake处理模型,实现了对视频图像的流水线处理,并且生成的GVF‐Snake流水线处理方式充分利用了现在多核多处理器设备的性能,提高了对视频流图像的处理效率,更好的满足了实时性。
Image segmentation is one of the important research field in computer vision and pattern recognition. It plays the important roles in many fields and many methods about it are proposed. Active contour model (also known as Snake model) is a new image segmentation and object extraction method, which extract object or locate feature of interest in images is efficient and correct by using high-level information and top-down processing.GVF Snake model as an improvement method over the traditional Snake model, is a hot research topic in image processing in recent years.
Firstly, this paper introduces the basic concepts of image segmentation. based on its partition principle and the mathematical model, the paper divide image segmentation into two categories: Based on data-driven segmentation and model-driven segmentation. GVF Snake is based model-driven image segmentation;
Secondly, the paper analysis the advantages and disadvantages of the traditional Snake model and GVF Snake model. in order to solve the problem of GVF Snake that need to manually set the initial contour, this paper use OSTU method to get the required initial contour for GVF Snake model, when use this curve to search object with GVF Snake model it is better to approach the target . GVF Snake model can get a faster and better segmentation result, and can division the image automatically.
lastly, GVF Snake is a time-consuming method. combined with the Intel parallel processing technology, the paper apply pipeline to the GVF Snake model and fully use of the multi-core、multi-processor device performance, which speed up the processing speed of image segmentation .
本文首先介绍了图像分割的基本概念,基于其分割原理以及数学模型的不同把图像分割分成两类:基于数据驱动的分割和基于模型驱动的分割。本文研究的GVF Snake属于基于模型驱动的图像分割方法;
接着详细分析了传统Snake方法和及其改进模型GVF Snake方法的优缺点;针对GVF Snake需要人工设置初始轮廓线的问题,本文改进OSTU(大津法)产生GVF Snake模型所需的初始轮廓曲线,新生成的初始轮廓曲线往往比较接近目标轮廓,当GVF Snake模型使用此曲线搜索时,往往能更快、更好的分割目标,并且使用OSTU方法实现了分割的自动化;
最后,针对GVF Snake模型分割目标费时的特点,本文使用VC++实现整个GVF‐Snake模型,结合Intel最新的并行处理技术,把流水线应用于GVF Snake处理模型,实现了对视频图像的流水线处理,并且生成的GVF‐Snake流水线处理方式充分利用了现在多核多处理器设备的性能,提高了对视频流图像的处理效率,更好的满足了实时性。
Image segmentation is one of the important research field in computer vision and pattern recognition. It plays the important roles in many fields and many methods about it are proposed. Active contour model (also known as Snake model) is a new image segmentation and object extraction method, which extract object or locate feature of interest in images is efficient and correct by using high-level information and top-down processing.GVF Snake model as an improvement method over the traditional Snake model, is a hot research topic in image processing in recent years.
Firstly, this paper introduces the basic concepts of image segmentation. based on its partition principle and the mathematical model, the paper divide image segmentation into two categories: Based on data-driven segmentation and model-driven segmentation. GVF Snake is based model-driven image segmentation;
Secondly, the paper analysis the advantages and disadvantages of the traditional Snake model and GVF Snake model. in order to solve the problem of GVF Snake that need to manually set the initial contour, this paper use OSTU method to get the required initial contour for GVF Snake model, when use this curve to search object with GVF Snake model it is better to approach the target . GVF Snake model can get a faster and better segmentation result, and can division the image automatically.
lastly, GVF Snake is a time-consuming method. combined with the Intel parallel processing technology, the paper apply pipeline to the GVF Snake model and fully use of the multi-core、multi-processor device performance, which speed up the processing speed of image segmentation .
引文
【1】D. Marr. Vision.A Computational Investigation into the Human Representation and Processing of Visual Information. W. H. Freeman and Company, San Francisco,1982.视觉计算理论.姚国正,刘磊,汪云九.北京;科学出版社,1998.
【2】A. J. Davision, P. W. Muray.Mobile robot localization using active contour vision.Proc.5 European Conf. on Computer Vision.Springer-Verlag.1998.385-392.
【3】B. Espian, F. Chaumette, P. Rives.A new approach to visual severing in robotics.IEEE Tram.On Robotics and Automation.Jun.1993,8(3):313—320.
【4】M. Kass, A. witkin and D. Terzopoulos.Snake:Active contour model,Int.J.Computer Vision, 1987(1):321--331.
【5】S. H. Matth, A statistical active contour model for SAR image segmentation. Image and Vision Computing. 1999,17:213—224.
【6】Y. Itai,Kim Hyoungseop, S. Ishikawa, etc.Automatic segmentation of lung areas based on SNAKES and extraction of abnormal areas.Proc.of IEEE 17 Int.Conf. on Tools with Artificial Intelligence.Nov.2005, 377--381.
【7】V. Van Raad, Active contour models: a multiscale implementation for anatomical feature delineation in cervical images. IEEE. Int. Conf. on Image Processing, Oct, 2004. 1: 1557--560
【8】G. Hamameh, A. Chodorowski,T. Gustavsson.Active concur models:applicafion to oral lesion detection in color images.IEEE.Int.Conf.on Systems,Man and Cybernetics.Oct.2000A:2458-2463.
【9】周则明,王平安,夏德深.基于Snake模型的低对比度噪声心脏MRI图像分割.系统仿真学报,2004,16(11):2493--2500。
【10】B. Bascle,P. Bouthemy, R. Deriche, et al.Tracking complex primitives in an image sequence.Proc.of IEEE 12 Int.Conf.on Pattern Recognition.Oct l994,426～431.
【11】Hao Jiang,M. S. Drew.A predictive contour inertia snake model for general video tracking.Proc.of IEEE Int. Conf. on Image Processing. Jun.2002,2:413--416.
【12】Xin Li,Xiao Cao,Yi Murphey, et a1.A real-time vehicle detection and tracking system in outdoor traffic scenes.Proc.of IEEE 17 Int. Conf. on Pattern Recognition.Aug.2004,2:761-764.
【13】Jenchao Tai,Shungtsang Tseng,Chingpo Lin,et a1.Real-time image tracking for automatic traffic monitoring and enforcement applications . Image and Video Computing 2004.22(6):485-501.
【14】GI. Chiou,J. N. Hwang. Lipreading from color video.IEEE Trans. on Image Processing.1997.6(8):1192-1195.
【15】R. Kaucic,A. Blake. Accurate, Real-Time,Unadorner Lip Tracking. Int. Conf. on Computer Vision.IEEE Computer Society.1998,370-375.
【16】L. D. Cohen and L. Cohen.Finite-element method for active contour models and balloons for 2D and 3D images.IEEE Trans.On Pattern Analysis and Machine Intelligence,1993,15(11):1131-1141.
【17】L. D. Cohen.On active contour models and balloons,CVGIP:Image Understanding,V01.53,PP.211-218,March 1991.
【18】B. Leroy, I. Herliin,and L. D. Cohen,Multi-resolution algon-rithm for active contour models . In 12 Int. Conf.Analysis and optimisation of System,PP.58-65.1996.
【19】G. Xu,E. Segawa,S.Tsuji.Robust Active Contours with Insensitive Parameters. Pattern Recognition.1994,27(7):879-884.
【20】C. Xu and J.L.Prince.Snakes, Shapes, and Gradient Vector Flow .IEEE Trans.on Image Processing,1998,7(3):359-369.
【21】Chenyang Xu, J.L.Prince.Generalized Gradient Vector Flow External Force for Active Contours.Signal Processing.1998.71:131-139.
【22】Xu C,Prince J L,Yezzi A.On the Relationship between Parametric and Geometric Active Contours.Proc.of the 34th Asilomar Conference on Signals,Systems and Computers.2000,10.483—489.
【23】D.J. Williams,M. Shah,A Fast Algorithm for Active Contours and Curvature Estimation.CVGIP:Image Understanding.1990.55(1):14-26.
【24】AA.AINI,T.E.Weymouth. R. C. Jain.Using Dynamic programming for Solving Variational Problems in Vision,IEEE Transactions on PAML.l989,12(9):855-867.
【25】王爱民,沈兰荪.图像分割研究综述.测控技术.2000,5. 1-6.
【26】(美)冈萨雷斯(Rafael C.Gonzalez)等著,阮秋琦,阮宇智译等译.数字图像处理第二版[M].北京:电子工业出版社.2003.
【27】Bhandarkar S M, and Zhang Hui. Image Segmentation Using Evolutionary Computation. IEEE Trans. On Evolutionary Computation, 1999, 3(1): 1~21
【28】Enno Littmann, and Helge Ritter. Adaptive Color Segmentation-A Comparison of Neural and Statistical Methods. IEEE Trans. On Neural Networks, 1997, 8(1): 175~185
【29】李厚强,王宜主,刘政凯.一种适用于多类别遥感图像分类的新方法.遥感学报, 1997, 1(4): 257~261
【30】Otsu N. Discriminant and least square threshold selection. In: Proc 4IJCPR, 1978:592-596.
【31】Brink A B. Gray level thresholding of images using a correlation criterion. Pattern Recognition Letters, 1989,9:335~341
【32】Mardia K V and Hainsworth T J. A spatial thresholding method for image segmentation. IEEE Trans. Pattern Analysis and Machine Intelligence, 1988,10:919~927
【33】Pun T. Anewmethod for gray level picture thresholding using the entropy of the histogram. Signal Processing, 1980,2:223~237
【34】靳宏磊,朱蔚萍,李立源,陈维南.二维灰度直方图的最佳分割方法.模式识别与人工智能,1999,12(3):329~333
【35】刘文萍,吴立德.图像分割中阈值选取方法比较研究.模式识别与人工智能,1997,10(3):271~277
【36】Rohlfing T., Maurer C.R., Nonrigid image registration in shared-memory multiprocessor environments with application to brains, breasts, and bees, IEEE Trans. Inform. Technol. Biomed., 7(1), 2003, 16-25
【37】Butz T., Thiran J.-P., Affine registration with feature space mutual information, in Proceeding of the 4th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI’01), Leccture Notes in Computer Science, Vol. 2208, 2001, 549-556
【38】毛国超,王骕,丁友东,基于可编程流水线的水墨画效果实时绘制计算机工程与设计2009,30 (13) 3175
【39】何元清,孙世新,傅彦,并行编程模式及分析电子科技大学学报Vol.31 No.2 Apr. 2002
【2】A. J. Davision, P. W. Muray.Mobile robot localization using active contour vision.Proc.5 European Conf. on Computer Vision.Springer-Verlag.1998.385-392.
【3】B. Espian, F. Chaumette, P. Rives.A new approach to visual severing in robotics.IEEE Tram.On Robotics and Automation.Jun.1993,8(3):313—320.
【4】M. Kass, A. witkin and D. Terzopoulos.Snake:Active contour model,Int.J.Computer Vision, 1987(1):321--331.
【5】S. H. Matth, A statistical active contour model for SAR image segmentation. Image and Vision Computing. 1999,17:213—224.
【6】Y. Itai,Kim Hyoungseop, S. Ishikawa, etc.Automatic segmentation of lung areas based on SNAKES and extraction of abnormal areas.Proc.of IEEE 17 Int.Conf. on Tools with Artificial Intelligence.Nov.2005, 377--381.
【7】V. Van Raad, Active contour models: a multiscale implementation for anatomical feature delineation in cervical images. IEEE. Int. Conf. on Image Processing, Oct, 2004. 1: 1557--560
【8】G. Hamameh, A. Chodorowski,T. Gustavsson.Active concur models:applicafion to oral lesion detection in color images.IEEE.Int.Conf.on Systems,Man and Cybernetics.Oct.2000A:2458-2463.
【9】周则明,王平安,夏德深.基于Snake模型的低对比度噪声心脏MRI图像分割.系统仿真学报,2004,16(11):2493--2500。
【10】B. Bascle,P. Bouthemy, R. Deriche, et al.Tracking complex primitives in an image sequence.Proc.of IEEE 12 Int.Conf.on Pattern Recognition.Oct l994,426～431.
【11】Hao Jiang,M. S. Drew.A predictive contour inertia snake model for general video tracking.Proc.of IEEE Int. Conf. on Image Processing. Jun.2002,2:413--416.
【12】Xin Li,Xiao Cao,Yi Murphey, et a1.A real-time vehicle detection and tracking system in outdoor traffic scenes.Proc.of IEEE 17 Int. Conf. on Pattern Recognition.Aug.2004,2:761-764.
【13】Jenchao Tai,Shungtsang Tseng,Chingpo Lin,et a1.Real-time image tracking for automatic traffic monitoring and enforcement applications . Image and Video Computing 2004.22(6):485-501.
【14】GI. Chiou,J. N. Hwang. Lipreading from color video.IEEE Trans. on Image Processing.1997.6(8):1192-1195.
【15】R. Kaucic,A. Blake. Accurate, Real-Time,Unadorner Lip Tracking. Int. Conf. on Computer Vision.IEEE Computer Society.1998,370-375.
【16】L. D. Cohen and L. Cohen.Finite-element method for active contour models and balloons for 2D and 3D images.IEEE Trans.On Pattern Analysis and Machine Intelligence,1993,15(11):1131-1141.
【17】L. D. Cohen.On active contour models and balloons,CVGIP:Image Understanding,V01.53,PP.211-218,March 1991.
【18】B. Leroy, I. Herliin,and L. D. Cohen,Multi-resolution algon-rithm for active contour models . In 12 Int. Conf.Analysis and optimisation of System,PP.58-65.1996.
【19】G. Xu,E. Segawa,S.Tsuji.Robust Active Contours with Insensitive Parameters. Pattern Recognition.1994,27(7):879-884.
【20】C. Xu and J.L.Prince.Snakes, Shapes, and Gradient Vector Flow .IEEE Trans.on Image Processing,1998,7(3):359-369.
【21】Chenyang Xu, J.L.Prince.Generalized Gradient Vector Flow External Force for Active Contours.Signal Processing.1998.71:131-139.
【22】Xu C,Prince J L,Yezzi A.On the Relationship between Parametric and Geometric Active Contours.Proc.of the 34th Asilomar Conference on Signals,Systems and Computers.2000,10.483—489.
【23】D.J. Williams,M. Shah,A Fast Algorithm for Active Contours and Curvature Estimation.CVGIP:Image Understanding.1990.55(1):14-26.
【24】AA.AINI,T.E.Weymouth. R. C. Jain.Using Dynamic programming for Solving Variational Problems in Vision,IEEE Transactions on PAML.l989,12(9):855-867.
【25】王爱民,沈兰荪.图像分割研究综述.测控技术.2000,5. 1-6.
【26】(美)冈萨雷斯(Rafael C.Gonzalez)等著,阮秋琦,阮宇智译等译.数字图像处理第二版[M].北京:电子工业出版社.2003.
【27】Bhandarkar S M, and Zhang Hui. Image Segmentation Using Evolutionary Computation. IEEE Trans. On Evolutionary Computation, 1999, 3(1): 1~21
【28】Enno Littmann, and Helge Ritter. Adaptive Color Segmentation-A Comparison of Neural and Statistical Methods. IEEE Trans. On Neural Networks, 1997, 8(1): 175~185
【29】李厚强,王宜主,刘政凯.一种适用于多类别遥感图像分类的新方法.遥感学报, 1997, 1(4): 257~261
【30】Otsu N. Discriminant and least square threshold selection. In: Proc 4IJCPR, 1978:592-596.
【31】Brink A B. Gray level thresholding of images using a correlation criterion. Pattern Recognition Letters, 1989,9:335~341
【32】Mardia K V and Hainsworth T J. A spatial thresholding method for image segmentation. IEEE Trans. Pattern Analysis and Machine Intelligence, 1988,10:919~927
【33】Pun T. Anewmethod for gray level picture thresholding using the entropy of the histogram. Signal Processing, 1980,2:223~237
【34】靳宏磊,朱蔚萍,李立源,陈维南.二维灰度直方图的最佳分割方法.模式识别与人工智能,1999,12(3):329~333
【35】刘文萍,吴立德.图像分割中阈值选取方法比较研究.模式识别与人工智能,1997,10(3):271~277
【36】Rohlfing T., Maurer C.R., Nonrigid image registration in shared-memory multiprocessor environments with application to brains, breasts, and bees, IEEE Trans. Inform. Technol. Biomed., 7(1), 2003, 16-25
【37】Butz T., Thiran J.-P., Affine registration with feature space mutual information, in Proceeding of the 4th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI’01), Leccture Notes in Computer Science, Vol. 2208, 2001, 549-556
【38】毛国超,王骕,丁友东,基于可编程流水线的水墨画效果实时绘制计算机工程与设计2009,30 (13) 3175
【39】何元清,孙世新,傅彦,并行编程模式及分析电子科技大学学报Vol.31 No.2 Apr. 2002