7708

7180

كارشناسي ارشد

كنترل

اصفهان: دانشگاه صنعتي اصفهان، دانشكده برق و كامپيوتر

1391

ده،95ص.: مصور،جدول،نمودار﴿رنگي﴾

ص.ع.به فارسي و انگليسي

مريم ذكري،حسين رباني

اردشير طالبي

22/2/92

رسول امير فتاحي، بهزاد نظري

مهندسي برق و كامپيوتر

ID7180

به فارسي و انگليسي: قابل رويت در نسخه ديجيتالي

100 Automatic Diagnosis of Malaria Based on Complete Circle Ellipse Fitting Search Algorithm Monireh Sheikhhosseini m sheikhhosseini@ec iut ac ir Date of Submission January Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan Iran Degree M Sc Language Farsi Supervisor Dr Maryam Zekri m zekri@cc iut ac ir Dr Hossein Rabbani h rabbani@med mui ac ir Abstract Since half the world s population is at the risk of malaria it is considered as a serious world health problem The definite way to diagnose malaria is microscopic observation of blood films Blood films are techniques in which the position of parasites is fixed on the blood smears Malaria diagnosis by the microscopic method is a time consuming and tedious task for technicians In addition the wrong decision may be made especially in the early stages of malaria Some of the problems with manual microscopy can be overcome by exploring computer based methods which are employed in order to automation of diagnostic procedure Automatic malaria diagnosis from microscopic images belongs to image classification problems which are included preprocessing segmentation feature extraction and classification steps The segmentation step of automatic malaria diagnosis is subjected to segment stained objects in the image which are candidate to be infected by parasites The nonlinear diffusion filtering algorithm is employed in preprocessing phase in order to remove noise from the image In this thesis parasite shape or the infected region is extracted by introducing a complete searching process Searching algorithm is based on least square ellipse fitting method which completes the parasite s shape through a step by step searching process This searching process is composed nucleus circle and ellipse searching phases The nucleus searching process is based on intensity thresholding Also by the circle searching process the position and initial points of parasites are determined The ellipse fitting method is applied on the resulted initial points in the next step in order to determine the ellipse searching parameters The ellipse searching process uses the fitted ellipse parameters in order to do the searching task around the fitted ellipse and complete the parasite shape Introducing the searching process causes a fast and simple feature extraction and classification steps Furthermore it empowers the algorithm to apply on the overlapped cells without employing clump splitting methods The proposed method uses the geometric features and decision rules to make a decision on healthy or infected objects Applying the proposed algorithm on the provided image from thin blood films shows the sensitivity and specificity which are equal to and respectively Keywords Automatic malaria diagnosis nonlinear diffusion filtering ellipse fitting least square circle search

مريم ذكري،حسين رباني

اردشير طالبي

رسول امير فتاحي، بهزاد نظري