Segmentation and Identification of Mycobacterium Tuberculosis (MTB)

This article addresses the detection of Mycobacterium tuberculosis (MTB) using digital image processing to enhance tuberculosis diagnosis. The microscopic-observation drugsusceptibility (MODS) method and computer vision strategies are employed, carrying out several preprocessing steps, such as grayscale conversion, binarization and morphological operations. Four computer vision models were trained, with the third model standing out for its accuracy by combining original images with segmented edges. This approach facilitates the rapid and accurate detection of bacilli, enhancing diagnosis and tuberculosis control.