Sputum Based Pneumonia Detection through Image Processing

Abstract

Sputum is a liquid composing saliva and mucus produced in the human lungs and in trachea that leads to lungs. The respiratory infection caused in human lungs could be investigates by processing and analyzing the sputum samples. Pneumonia is one kind of lung infection which is diagnosed by using sputum. Microscopic visualization and the inspection process for this detection process can be implemented by image processing. Image processing technique that accept the image as input, processes each pixel of the image and analyze each cell for detection and identification as the cell belong to infection causing bacteria or not. This process is very cumbersome when huge dataset is present. Hence, implementation of an automated system that takes microscopic image of sputum sample and process them by applying image processing techniques to detect and classify the type of bacteria present could be a solution for this problem.  The proposed method reduces the complexity involved in manual process and provides accessibility through automated system and helps to build a microbiological application. Sputum image has been tested for identification of Cocci and bacilli bacterial cell.