基于改进水平集的菌落图像智能计数算法
|
摘要
针对现有的菌落自动识别计数方法对背景敏感、对多菌种菌落分割普适性差的缺点,提出一种基于改进水平集的全自动菌落分割、计数方法.该方法利用偏置场对背景进行建模以消除背景灰度不均影响;构造含有终止条件的多相水平集算法实现菌落目标的自适应分割;通过极坐标空间中凹点检测实现粘连目标计数修正.由天津市食品安全检测技术研究院提供300例多菌种混杂菌落样本做为测试集,以专家人工鉴定结果为金标准,将本方法与传统形态学方法、迅数icount10两种定量方法进行对比,菌落密度在300 CFU内时,本方法计数准确率达到92.7%,对多菌种混杂菌落的计数精度、分割效果都优于其他两种方法.
The current colony counting methods are sensitive to background and are inefficient from the viewpoint of the universality of multi-species colony segmentation. To overcome this limitation,an automatic colony segmentation and counting method based on an improved level set is proposed. Here in,the bias field is used to model the background to eliminate the influence of intensity inhomogeneity. A multiphase level set algorithm with termination condition is constructed to realize adaptive segmentation of colony target. Through the concave point detection in polar coordinates,the correction of adhesion targets counting is achieved. The algorithm is tested on 300 samples of multispecies colonies provided by Tianjin Food Safety Inspection Technology Research Institute. The algorithm is compared with two quantitative methods:traditional morphological algorithm and Shineso icount10 algorithm,and the results of expert artificial identification are used as gold standards. When the colony density is under 300 CFU,the counting accuracy rate is up to 92.7%. Both the counting accuracy and segmentation effect of multi-species colonies using the proposed method are superior to those of the other two algorithms.
The current colony counting methods are sensitive to background and are inefficient from the viewpoint of the universality of multi-species colony segmentation. To overcome this limitation,an automatic colony segmentation and counting method based on an improved level set is proposed. Here in,the bias field is used to model the background to eliminate the influence of intensity inhomogeneity. A multiphase level set algorithm with termination condition is constructed to realize adaptive segmentation of colony target. Through the concave point detection in polar coordinates,the correction of adhesion targets counting is achieved. The algorithm is tested on 300 samples of multispecies colonies provided by Tianjin Food Safety Inspection Technology Research Institute. The algorithm is compared with two quantitative methods:traditional morphological algorithm and Shineso icount10 algorithm,and the results of expert artificial identification are used as gold standards. When the colony density is under 300 CFU,the counting accuracy rate is up to 92.7%. Both the counting accuracy and segmentation effect of multi-species colonies using the proposed method are superior to those of the other two algorithms.
引文
[1]陈可意.菌落计数与分类智能算法研究[D].天津:天津大学精密仪器与光电子工程学院,2014.Chen Keyi.Study on Intelligent Algorithm of Colony Counting and Classification[D].Tianjin:School of Precision Instrument and Opto-Electronics Engineering,Tianjin University,2014(in Chinese).
[2]中华人民共和国卫生部.GB4789.2食品安全国家标准[S].北京:食品微生物学检验,2016.Ministry of Health of the People's Republic of China.GB4789.2 National Food Safety Standard[S].Beijing:Microbiology Detection,2016(in Chinese).
[3]尹诗白.基于模糊集合理论的颗粒目标分割和识别[D].西安:长安大学,2013.Yin Shibai.Segmentation and Recognition of Particle Object Based on Fuzzy Set Theory[D].Xi’an:Chang’an University,2013(in Chinese).
[4]司秀娟,王红强,宋良图.基于图像处理的菌落自动计数系统[J].北京生物医学工程,2017,36(6):633-638.Si Xiujuan,Wang Hongqiang,Song Liangtu.Automatic colony counting system based on image processing[J].Beijing Biomedical Engineering,2017,36(6):633-638(in Chinese).
[5]武宗茜,王鹏,丁天怀.活动轮廓模型在重叠藻细胞计数中的应用[J].计算机工程,2012,38(3):209-211.Wu Zongqian,Wang Peng,Ding Tianhuai.Application of active contour model in overlapped algae cells counting[J].Computer Engineering,2012,38(3):209-211(in Chinese).
[6]Dorbe N,Jaundalders A,Kadikis R,et al.FCN and LSTM based computer vision system for recognition of vehicle type,license plate number,and registration country[J].Automatic Control&Computer Sciences,2018,52(2):146-154.
[7]朱翔宇.基于卷积神经网络的肺部结节检测系统研制[D].南京:南京大学,2017.Zhu Xiangyu.The Research of a Pulmonary Nodule Detection Method Based on the Conventional Neural Network[D].Nanjing:Nanjing University,2017(in Chinese).
[8]Roy K,Bhattacharya P,Suen C Y.Towards nonideal iris recognition based on level set method,genetic algorithms and adaptive asymmetrical SVMs[J].Engineering Applications of Artificial Intelligence,2010,24(3):458-475.
[9]吴永飞.图像分割的变分模型及数值实现[D].重庆:重庆大学,2016.Wu Yongfei.Variational Model and Numerical Implementation of Image Segmentation[D].Chongqing:Chongqing University,2016(in Chinese).
[10]Li C,Huang R,Ding Z,et al.A level set method for image segmentation in the presence of intensity inhomogeneities with application to MRI[J].IEEE Transactions on Image Processing,2011,20(7):2007.
[11]Li C,Xu C,Gui C,et al.Distance regularized level set evolution and its application to image segmentation[J].IEEE Transactions on Image Processing,2010,19(12):3243-3254.
[12]张国权,李战明,李向伟,等.HSV空间中彩色图像分割研究[J].计算机工程与应用,2010,46(26):179-181.Zhang Guoquan,Li Zhanming,Li Xiangwei,et al.Research on color image segmentation in HSV space[J].Computer Engineering and Applications,2010,46(26):179-181(in Chinese).
[13]Li C,Kao C Y,Gore J C,et al.Minimization of region-scalable fitting energy for image segmentation[J].IEEE Transactions on Image Processing,2008,17(10):1940-1949.
[2]中华人民共和国卫生部.GB4789.2食品安全国家标准[S].北京:食品微生物学检验,2016.Ministry of Health of the People's Republic of China.GB4789.2 National Food Safety Standard[S].Beijing:Microbiology Detection,2016(in Chinese).
[3]尹诗白.基于模糊集合理论的颗粒目标分割和识别[D].西安:长安大学,2013.Yin Shibai.Segmentation and Recognition of Particle Object Based on Fuzzy Set Theory[D].Xi’an:Chang’an University,2013(in Chinese).
[4]司秀娟,王红强,宋良图.基于图像处理的菌落自动计数系统[J].北京生物医学工程,2017,36(6):633-638.Si Xiujuan,Wang Hongqiang,Song Liangtu.Automatic colony counting system based on image processing[J].Beijing Biomedical Engineering,2017,36(6):633-638(in Chinese).
[5]武宗茜,王鹏,丁天怀.活动轮廓模型在重叠藻细胞计数中的应用[J].计算机工程,2012,38(3):209-211.Wu Zongqian,Wang Peng,Ding Tianhuai.Application of active contour model in overlapped algae cells counting[J].Computer Engineering,2012,38(3):209-211(in Chinese).
[6]Dorbe N,Jaundalders A,Kadikis R,et al.FCN and LSTM based computer vision system for recognition of vehicle type,license plate number,and registration country[J].Automatic Control&Computer Sciences,2018,52(2):146-154.
[7]朱翔宇.基于卷积神经网络的肺部结节检测系统研制[D].南京:南京大学,2017.Zhu Xiangyu.The Research of a Pulmonary Nodule Detection Method Based on the Conventional Neural Network[D].Nanjing:Nanjing University,2017(in Chinese).
[8]Roy K,Bhattacharya P,Suen C Y.Towards nonideal iris recognition based on level set method,genetic algorithms and adaptive asymmetrical SVMs[J].Engineering Applications of Artificial Intelligence,2010,24(3):458-475.
[9]吴永飞.图像分割的变分模型及数值实现[D].重庆:重庆大学,2016.Wu Yongfei.Variational Model and Numerical Implementation of Image Segmentation[D].Chongqing:Chongqing University,2016(in Chinese).
[10]Li C,Huang R,Ding Z,et al.A level set method for image segmentation in the presence of intensity inhomogeneities with application to MRI[J].IEEE Transactions on Image Processing,2011,20(7):2007.
[11]Li C,Xu C,Gui C,et al.Distance regularized level set evolution and its application to image segmentation[J].IEEE Transactions on Image Processing,2010,19(12):3243-3254.
[12]张国权,李战明,李向伟,等.HSV空间中彩色图像分割研究[J].计算机工程与应用,2010,46(26):179-181.Zhang Guoquan,Li Zhanming,Li Xiangwei,et al.Research on color image segmentation in HSV space[J].Computer Engineering and Applications,2010,46(26):179-181(in Chinese).
[13]Li C,Kao C Y,Gore J C,et al.Minimization of region-scalable fitting energy for image segmentation[J].IEEE Transactions on Image Processing,2008,17(10):1940-1949.