:: Volume 7, Issue 1 (Vol.7, No.1,spring 2019) ::
J Neyshabur Univ Med Sci 2019, 7(1): 36-56 Back to browse issues page
A novel technique to diagnose malaria using image processing
Mohammad Rostami Mr * 1, Elahe Neshat Mrs2
1- Member of Young Researchers and elite Club, Dehaghan Branch, Islamic Azad University, Dehaghan, Isfahan, Iran , mohamad.rostami10@yahoo.com
2- Department of Computer Engineering and IT, Tehran branch Islamic Azad University, north-branch, Iran
Abstract:   (557 Views)

Introduction
Malaria is an infectious disease which has caused serious problems in regard to human health. In order to control and cure Malaria, we need the correct diagnosis and required treatments. In the present report, we present a new automatic method for diagnosing Malaria and its parasite by using an image processing technique and nervous system.
Materials and Methods
In this method, we started with pre-processing to separate the red globules from the image with the help of an active contour model. Then by using wavelet function, we separated 840 features from the image. Afterward, with the advantage of vector supporting machine, we categorized the globules into normal and abnormal.
Results
In comparison to the past, the process of diagnosis improved and we have reached the result of 99.78%.
Conclusion
In the pre-processing level, the noiser which has been recognized for categorization of red globules separated from the choice and then by using wavelet, we extract the features. We have tested and analyzed the data by using some nervous system and vector supporting machine.

Keywords: active contour model, categorization, extracting features, malaria, vector supporting machine, wavelet
Full-Text [PDF 1930 kb]   (165 Downloads)    
Type of Study: Orginal Article | Subject: Nursing
Received: 2017/03/19 | Revised: 2019/04/29 | Accepted: 2019/02/25 | ePublished ahead of print: 2019/04/27 | Published: 2019/04/29 | ePublished: 2019/04/29


XML   Persian Abstract   Print



Volume 7, Issue 1 (Vol.7, No.1,spring 2019) Back to browse issues page