The incidence of tuberculosis (TB) is rising in many low- and middle-income countries despite the availability of effective, inexpensive and practical technologies for control of the disease. Early detection of TB is required for the initiation and monitoring of treatment and for epidemiological tracking, and therefore is essential for disease control. Direct sputum smear microscopy is the most cost-effective method of screening suspected cases of TB in high-prevalence countries. While fluorescence microscopy of auramine stained sputum specimens for TB screening has greater sensitivity than light microscopy of Ziehl-Neelsen (ZN) stained specimens, the latter is the method of choice in developing countries due its low cost and simplicity. Trained personnel required for the analysis of specimen slides is often a scarce resource in laboratories with a high specimen throughput. Human resource requirements may be reduced and greater efficiencies and accuracies achieved by automation of parts of the TB screening process. Automation of ZN-microscopy may increase the sensitivity of the process to approach that of fluorescence microscopy. While automated microscopy systems have been developed for other applications, TB has received limited attention in this respect. This project proposes to design and construct a device that will obtain digital images of a large number of fields in ZN-stained sputum slides, process these images to detect TB bacilli, and store a subset of images for human verification. Algorithms will be developed for automated focusing of the device and automated detection and classification of TB bacilli. While automated slide scanning systems have relied on attaching digital cameras to light or fluorescence microscopes, the proposed system will replace the traditional microscope by attaching the necessary optical, mechanical and electronic components and a digital camera directly to a microprocessor that will control all aspects of slide analysis. This device will form the basis for the future development of a fully automated slide handling system capable of batch processing of large numbers of slides, for TB screening as well as other slide analysis applications. This project is aimed at the development of an automated microscope that will examine sputum specimens and determine whether tuberculosis is present. Automation of this process will enable laboratories in countries with a high TB burden to screen for the disease more effectively, thus contributing to the management and control of the disease world-wide. ? ? ?
| Patel, Bhavin; Douglas, Tania S (2012) Creating a virtual slide map from sputum smear images for region-of-interest localisation in automated microscopy. Comput Methods Programs Biomed 108:38-52 |
| Patel, Bhavin; Douglas, Tania S (2011) Creating a virtual slide map of sputum smears by auto-stitching. Conf Proc IEEE Eng Med Biol Soc 2011:5088-91 |
| Khutlang, Rethabile; Krishnan, Sriram; Dendere, Ronald et al. (2010) Classification of Mycobacterium tuberculosis in images of ZN-stained sputum smears. IEEE Trans Inf Technol Biomed 14:949-57 |
| Osibote, O A; Dendere, R; Krishnan, S et al. (2010) Automated focusing in bright-field microscopy for tuberculosis detection. J Microsc 240:155-63 |
| Khutlang, R; Krishnan, S; Whitelaw, A et al. (2010) Automated detection of tuberculosis in Ziehl-Neelsen-stained sputum smears using two one-class classifiers. J Microsc 237:96-102 |
