Convolutional Neural Network for Classification of Malaria Parasite in Blood Smear

الموضوعات

علوم تطبيقية وتكنولوجية

الملخص

Malaria is a very dangerous disease caused by the infection of red blood cells with Plasmodium, which is transmitted by female mosquitoes (Anopheles) throughout their feeding period from human blood. According to the World Health Organization, due to the biological evolution of the malaria parasite, which led to a rise in cases of infection to three million cases per year, while deaths are about four hundred thousand cases per year. In Sudan, the results of malaria examination differ from one laboratory to another, and the results are inaccurate, in addition to the long examination time and the lack of trained and qualified human element in villages and remote places from cities. The aim of this study to examine malaria automatically by CNN creating a computer vision for a computer system that gives results of high accuracy and provides time and serves the remote areas of the cities. Methodology is done by deep learning the data by using the CNN structure to classify the images of the blood sample into infected and uninfected, using standard images and dividing them into three infected and uninfected files and a test file, and then processing the images in the environment of the program (Anaconda) and using a high programming language (Python) and then Initialize image data to fit the neural network environment by converting it to grayscale. After that, the layers of the CNN were used on the images, and the model was evaluated by testing it with a random experiment for a set of images that are easy to draw and categorize into injured and uninjured. The model showed a high accuracy of 94.9% when training part of the images with a total of 19840 samples and when the value increased to seven, the result was 96.5% as shown the model produced two images that were randomly selected, drawn and classified into infected and uninfected, This study recommends Using computers with a high processor to train the largest size of images and data to obtain a higher accuracy rate of up to 100% and the model can be developed to determine the type of parasite, as well as a unified database that links between diagnosis and treatment at the same time.

روابط وملفات

التعريف والنوع

رقم الوثيقة

f248eb71-853b-4779-9926-150109277989

رقم العقد

0

نوع الوسائط

Crawler

نوع المحتوى

الرسائل العلمية

صيغة المصدر

رسائل ماجيستير

نوع الملف

pdf text

أسماء الملفات

2202769_1.pdf

بيانات النشر

ألقاب المؤلفين

[{"name_ar":" Mawahib Gafare Abdalrahman Ahmed ","title_ar":"اشراف","title_en":"Supervision"},{"name_ar":"Suliman Abdarahman Abdalla","title_ar":"اشراف","title_en":"Supervision"},{"name_ar":"Wejdan Abdalwhab Ali Ahmed","title_ar":"اعداد","title_en":"Preparation"}]

اللغة

English

المصدر والدورية

اسم المصدر

Convolutional Neural Network for Classification of Malaria Parasite in Blood Smear

المحتوى والصفحات

عدد الصفحات

0

كلمات الباحثين

Malaria -disease - malaria examination

إشراف وإعداد

الإشراف

Mawahib Gafare Abdalrahman Ahmed , Suliman Abdarahman Abdalla

الإعداد

Wejdan Abdalwhab Ali Ahmed

الاقتباسات الببليوغرافية