JoWUA

Abstract

Tuberculosis is an infectious disease that requires serious treatment. Extrapulmonary tuberculosis is detected using a microscope, but currently, it takes a long time because the fluid preparations are carefully viewed one by one in a microscope, and there are 150 fields of vision in the fluid preparations. Examination for extrapulmonary tuberculosis by culture takes between 1–2 weeks or even more. Examination by biopsy will take a long time because the fluid preparations are looked at carefully under the microscope one by one. The image of tuberculosis is expressed as an image where there is a bacillus object in red, but it turns out that apart from the bacillus object there are also other objects in red. So, it is essential to use computer technology for examinations of tuberculosis to be more efficient. This research aims to compare the Long Short-Term Memory (LSTM) and Bidirectional Long Short-Term Memory (BiLSTM) classification methods in the detection of extra-pulmonary tuberculosis disease, to obtain better accuracy results. This research carried out HSI color space transformation, segmentation using global thresholding, and feature extraction of 13 features based on shape and texture using the Correlation Based Feature Selection (CFS) method. The results show that BiLSTM has the best accuracy with a value of 88.40% with the number of features = 3, which are Short Run High Gray-Level Emphasis, Run Length Nonuniformity, and Minor axis length, while LSTM produces an accuracy of 63.19% with the number of features = 5. BiLSTM is capable of detecting opposite features, i.e., BiLSTM can detect opposite features in data sequences and BiLSTM's ability to understand multiple contexts, so it tends to provide more accurate results in some data classification tasks.