A comparative performance analysis of different denoising techniques in sputum smear images

Abstract

Tuberculosis (TB) is the deadliest, communicable disease and can nearly affect all the parts of our body. To diagnose this disease, there are various tests available such as chest X-rays and TB skin tests. The most commonly used tool for TB detection is the sputum smear microscopy (SSM) which is less costly. The specimens are stained using Ziehl–Neelsen and are then examined by the technicians for any microbes. Detection of bacilli from stained sputum images manually is a lengthy, far-reaching process which can lead to inaccuracy in the output. Therefore, automatic methods are provided which give an optimal solution in a short time, in the absence of skilled experts in disease diagnosis. This paper gives an overview of available preprocessing methods in various digital images and hence will benefit the researchers working in the smear microscopy field.