International Research Journal of Engineering and Technology (IRJET) Volume: 09 Issue: 07 | July 2022 www.irjet.net
e-ISSN: 2395-0056 p-ISSN: 2395-0072
MALARIAL PARASITES DETECTION IN THE BLOOD CELL USING CONVOLUTIONAL NEURAL NETWORK Deeksha Kotian1, Brina Paula2, Dhanush Hegde3, Darshith D N4, Annappa Swamy D R5 1,2,3,4
Computer Science & Engineering, MITE Moodabidri.
5
Associate Professor, Dept. of Computer Science & Engineering, MITE Moodabidri, Karnataka, India. ---------------------------------------------------------------------***-------------------------------------------------------------------medical professional. But RDT has few drawbacks like Abstract - Malaria is a life-threatening, infectious susceptibility to damage by heat and humidity and higher cost compared to a light microscope.
mosquito-borne disease caused by Plasmodium parasites. These parasites are transmitted by the bites of infected female Anopheles mosquitoes. It is a significant burden on our healthcare system and it is the major cause of death in many developing countries. Therefore, early testing is necessary to detect malaria and save lives. The standard diagnostic methods for malaria detection are Microscopy and Rapid Detection Test (RDT). Microscopy process requires a skilled microscopist which sometimes cannot be available in rural areas and it is impossible to manually detect the presence of parasites. The RDT may not be able to detect some infections with lower numbers of malaria parasites circulating in the patient’s bloodstream. Therefore, there is a need for specialized technology that proves essential to combat this problem. This proposed system uses a deep learning model based on convolutional neural network (CNN) to classify single blood smears whether it is parasitized or nonparasitized. A variety of techniques were performed under CNN model such as activation layer, relu , maxpool , dropout , flatten, dense layer to optimize and improve the model accuracy. Our model deep-learning model predicts malaria parasites from images with an accuracy of 95.34%.
The machine learning and deep learning approaches have proved to be successful in the diagnosis of a disease. Machine learning models require a lot of tuning, factor analysis, and feature engineering. The machine learning method is not scalable with more data provided. Machine learning requires feature engineering and feature training which is not a handy tool. Deep learning models are reliable and easily scalable with a higher accuracy rate. Convolutional Neural Networks (CNN) have provedto be really effective in a wide variety of computer vision tasks. Convolutional Neural Networks transform input image volume into an output volume holding a class label. The regular manual diagnosis of blood samples requires proper expertise in classifying and counting the infected and uninfected cells. This is a time-consuming task and accuracy will be very less. To overcome the problem faced during the detection of malarial parasites in the blood cell, there is a need to develop a system which predicts the presence of malarial parasites in the blood cell efficiently.
2. LITERATURE REVIEW
Key Words: CNN, Malaria detection, Deep Learning,
Snehal Suryawanshi et al. [1] In the proposed system Poisson distribution using minimum error thresholding is used to detect the presence of malarial parasites in blood cells using the technique of image segmentation. The work presented in this paper is based on some extended techniques. Poisson distribution based minimum error thresholding algorithm automatically binarizes images, which is refined by morphological opening and hence foreground is being extracted. Then by a novel method, the seed points are detected combining multiscale Laplacian of Gaussian with gabor filtering. Then the features extracted are compared with a database to check whether the blood cell images are infected by malaria parasites.
Sequential Model, Blood cells.
1.INTRODUCTION Malaria is a mosquito-borne disease caused by a plasmodium parasites. It is transmitted by the bite of infected mosquitoes. Worldwide, there is an estimated 300–500 million people who suffer from malaria each year, which results in 1.5–2.7 million deaths yearly. According to the World Health Organization (WHO), approximately 219 million cases were diagnosed with malaria resulting in 435,000 deaths globally in 2017. Malaria is a deadly disease which is more frequently found in rural areas where medical diagnosis and health care options are not easily accessible. For Malaria diagnosis, the RDT and microscopic diagnosis are the most used clinical methods. RDT is an effective and faster tool. Also, it does not require the presence of a trained
© 2022, IRJET
|
Impact Factor value: 7.529
Jigyasha Soni et al. [2] proposed another approach to differentiate between the simple RBC and malarial parasite affected blood cell. In order to maximize the productivity of the algorithm various approaches are
|
ISO 9001:2008 Certified Journal
|
Page 253