Classification of Mycobacterium tuberculosis in Images of ZN-Stained Sputum Smears

Abstract

Screening for tuberculosis (TB) in low-and middle-income countries is centered on the microscope. We present methods for the automated identification of Mycobacterium tuberculosis in images of Ziehl-Neelsen (ZN) stained sputum smears obtained using a bright-field microscope. We segment candidate bacillus objects using a combination of two-class pixel classifiers. The algorithm produces results that agree well with manual segmentations, as judged by the Hausdorff distance and the modified Williams index. The extraction of geometric-transformation-invariant features and optimization of the feature set by feature subset selection and Fisher transformation follow. Finally, different two-class object classifiers are compared. The sensitivity and specificity of all tested classifiers is above 95% for the identification of bacillus objects represented by Fisher-transformed features. Our results may be used to reduce technician involvement in screening for TB, and would be particularly useful in laboratories in countries with a high burden of TB, where, typically, ZN rather than auramine staining of sputum smears is the method of choice.

A survey for the applications of content-based microscopic image analysis in microorganism classification domains

“…However, by the development of CBMIA approaches, only accuracy is not enough to meet our technical demands. Hence, multiple evaluation methods are proposed, supporting a more diversified choice space, like the MAP in Li et al (2015b) and the sensitivity in Khutlang et al (2010a). For more information about the system evaluation, please refer to Kishida (2005), Theodoridis and Koutroumbas (2009).…”

“…In the experimental part, 2728 positive and 1648 negative images are used for classifier training, and 1157 positive and 1064 negative images are used for test, and finally the best sensitivity above 90% is obtained. As an extension of the above work, feature selecting methods and more classifiers are compared in Khutlang et al (2010a), including Bayesian, Euclidean distance linear, logistic linear and quadratic algorithms. Finally, an improved classification sensitivity about 95% is achieved.…”

“…Beneficial AMs can help farmers to increase the agricultural yields, for Shabtai et al (1996), Ronen et al (2002) IM 1998 Gerlach SR et al Gerlach et al (1998) IM 2001 Lee MS et al Lee et al (2001) IM 2002 Lomander A et al Lomander et al (2002) IM 2002 Park JP et al Park et al (2002) IM 2007 Lecault V et al Lecault et al (2007) IM 2011 Yu B et al Yu et al (2011) MM 1992 Reichl U et al Reichl et al (1992) MM 1994 Pichon D et al Pichon et al (1994) MM 1996 Oh B et al Oh et al (1996) MM 1997 Tamura S et al Tamura et al (1997) MM 1998 Veropoulos K et al Veropoulos et al (1998) MM 1998 Wit P et al Wit and Busscher (1998) MM 1999 Kay JW et al Kay et al (1999) MM 2000 Khutlang R et al Khutlang et al (2009Khutlang et al ( , 2010a MM 2011 Rulaningtyas R et al Rulaningtyas et al (2011) MM 2011 Pangilinan C et al Pangilinan et al (2011) MM 2011-2012Osman MK et al Osman et al (2011a, b, 2012 MM 2012 Chang J et al Chang et al (2012b) MM 2012-2016 Priya E et al Priya et al (2012), Srinivasan (2015a, b, 2016) Culverhouse et al (1994Culverhouse et al ( , 1996Culverhouse et al ( , 2000Culverhouse et al ( , 2003Culverhouse et al ( , 2006a WM 1995 Thiel S et al Thiel and Davies (1995), …”

Automatic identification of tuberculosis mycobacterium

“…The first methods for automatic bacillus screening in conventional microscopy were published in 2008 (Costa et al, 2008;Sadaphal et al, 2008;Raof et al, 2008). Recently, other methods for automatic bacillus screening were published (Forero et al, 2004(Forero et al, , 2006Khutlang et al, 2010;Lenseigne et al, 2007;Makkapati, et al, 2009;Osman et al, 2012;Sotaquirá et al, 2009).…”

Sputum Smear Microscopy for Tuberculosis: Evaluation of Autofocus Functions and Automatic Identification of Tuberculosis Mycobacterium

“…al., 2006;Sotaquira, 2009;Khutlang, 2010) claimed that the main advantages of an automatic bacilli screening over a manual one are better reproducible values for sensitivity and specificity and a faster screening process. Table 2 shows reported values for sensitivity, specificity and time waste for one image analysis using automatic methods.…”