Microscope image processing for TB diagnosis using shape features and ellipse fitting

Tuberculosis (TB) is an infectious disease caused by the bacteria Mycobacterium tuberculosis or simply M. tuberculosis. It is primarily an infection of lungs, but it can also affect other parts of the body such as brain, intestine, kidney and spine. TB remains one of the leading cause of death in developing countries, although most are preventable if diagnosed early and treated. Among the available tools, Sputum smear microscopy is the most widely used one for TB detection. It is done manually and is often time consuming; a laboratory technician is expected to spend at least 25 minutes per slide, limiting the number of slides that can be screened. Also any incorrect diagnosis will leads to serious health issues. So a solution is Automatic screening methods. Many attempts have been made to develop automatic approaches to identify TB bacteria from microscopic sputum smear images. In this paper, we present an automatic TB diagnosis technique using morphological features and ellipse fitting. Microscopic images of sputum smear are collected from infected subjects. These images are transformed into HSV color space for a better analysis, which is thresholded using hue range of red pink color. The resultant images will contain TB bacilli along with non-TB objects. In order to identify the TB regions from the non-TB regions, shape features of every identified region is evaluated. Finally a new algorithm that make use of concave contour points as well as ellipse fitting is performed to separate out the overlapping bacilli region and add them to the total count of bacilli.