Evidence record · UDINUS · SDGs
- Unit
- Rektorat UDINUS
- Period
- 2022
- Goals
- Record
- 3b85d05d-2bc2-4880-b455-7db416522b67
- Published
- 2026-10-30
Universitas Dian Nuswantoro (UDINUS), through the participation of Wulandari, Sari Ayu; Purnama, I Ketut Eddy; and Purnomo, Mauridhi Hery, conducted research on automatic segmentation of sputum color images for tuberculosis diagnosis. The activity presented an automatic algorithm for detecting the number of Mycobacterium tuberculosis in acid-fast bacilli (AFB) smear images containing I-shaped and V-shaped colonies. The method applied fuzzy intuitionistic processing based on an auto-thresholding segmentation approach.
AFB are described in the source as bacteria that can prevent acid decolorization during staining. When sputum preparations are given a blue color, the AFB retain their red color. The research addressed the problem of detecting bacterial numbers from AFB segmentation in conditions involving differences in light intensity and contrast caused by different lighting distributions. The stated aim was to segment the AFB image data as a whole, without dividing one bacterium into several parts.
The segmentation process consisted of patch preparation, mask preparation, and the use of a UNet architecture. During mask preparation, the research compared three color-threshold models: grayscale followed by black and white, Adaptive Histogram followed by black and white, and Fuzzy Intuitionistic followed by black and white. Each model was then segmented using the UNet method. The source identifies the creation of an input image mask as the novelty of the research. An optimization method using a maximal entropy approach was also applied. This approach sought the maximum degree of disorder by calculating entropy on the modified input image matrix.
The experimental results reported that the fuzzy intuitionistic method achieved high accuracy in segmenting AFB in I-shaped and V-shaped colonies, with an accuracy rate of 94.78%. This UDINUS research is related to RAW goal 3 through its focus on tuberculosis diagnosis, AFB image segmentation, and the detection of Mycobacterium tuberculosis in sputum smear images.