COMPUTER BASED DIAGNOSIS OF MALARIA IN THIN BLOOD SMEARS USING THRESHOLDING BASED APPROACH









Abstract

Millions of people worldwide are diagnosed with malaria every year and a lot of them results in the death of the infected person. Malaria in caused by a plasmodium infected anopheles genus mosquito. The conventional method of detecting the plasmodium parasite through the microscope requires a significant amount of time and is still prone to errors. Thus, making it inefficient for analysis where a large number of samples needs to be checked for malaria. This paper aims to propose an efficient automated system which uses image processing methods to identify the presence of plasmodium parasite in red blood cells. This whole process consists of four parts namely preprocessing, segmentation, feature extraction and classification of the parasite. The whole project was carried out in Matlab 2019a environment. After the preprocessing that included greyscale conversion and application of adaptive mean filter for reduction of noise from the images, we used Zach Thresholding process for segmentation of thin blood slide images. The database for images that was used was MP-IDB which consisted of total of 229 images classified in each category of species. We applied this process on all 229 images and got an accuracy of about 99.49%, 92.48% sensitivity and 99.79% specificity showing that this procedure is very efficient and with high accuracy.


Modules


Algorithms


Software And Hardware

• Hardware: Processor: i3 ,i5 RAM: 4GB Hard disk: 16 GB • Software: operating System : Windws2000/XP/7/8/10 Anaconda,jupyter,spyder,flask Frontend :-python Backend:- MYSQL