Automatic Detection of Tuberculosis Bacilli from Microscopic Sputum Smear Images Using Faster R-CNN, Transfer Learning and Augmentation

Tuberculosis (TB) is a serious infectious disease that remains a global health problem with an enormous burden of disease. TB spreads widely in low- and middle-income countries, which depend primarily on sputum smear test using conventional light microscopy in disease diagnosis, in this paper we propose a new deep-learning approach for bacilli localization and classification in conventional ZN-stained microscopic images. The approach is based on the state of the art Faster Region-based Convolutional Neural Network (RCNN) framework. Our experimental results show significant improvement by the proposed approach compared to existing methods, thus helping in accurate disease diagnosis.