Morphology and segmentation used inautomated screening of tuberculosis
- Autores: Balasubramani Perumal, Emil Molayil Paul
- Localización: Sustainable development in engineering and technology, 2022, ISBN 978-84-124943-4-1, págs. 313-328
- Idioma: inglés
- Enlaces
-
Resumen
Tuberculosis(TB) is a necessary hazard in several areas of the globe. The multi-drug-resistantmicroorganism lines have exacerbated the matter, whereas diagnosing Tuberculosiseven so stays a challenge.TB along with Covid- 19 brings prolonged adversaryeffects to the health of the affected person. Mortality rate will abruptly increase, ifthis disease is not diagnosed properly. In this paper, we deal with the morphologicalfilters,for image enhancement(for the input chest images after the preprocessing).Both top hat and bot hat filters are used for the proper ROI (Region Of Interest)selection,which can be further proceeded for the segmentation part.Here,we havealso compared different segmentation methodologies to analyse the performance ofsegmentation unit and found out that the marker based watershed segmentation excelsthe other segmentation methods such as Graph cut,KNN,Clustering methods,PDE(Partial Differential Equation),Region Based Method,Thresholding Method.Variusgraphical methods are used for the performance comparison.It yields an accuracy ofabout 81%.
