Author/Authors :
EL KOURD, K. Preparatory school of sience technical of Algiers-Algeriabc - Department of physical, Algeria , AZZIZI, A. University of Med khider of Biskra - Electronic of engineering, Algeria , BOUGOURZI, F. University of Med khider of Biskra - Electronic of engineering, Algeria , HAMMOUM, S. Clinical of radiology-MRI service-view Kouba-Algiers, Algeria
Abstract :
In this paper, we transform a nonlinear model to a linear one by using numerical analysis with ‘‘Runger-Kutta4(RK4)’’. Which is a mathematical technique to approximate solution of ordinary differential equations; this method is most popular where the step size H is working to increase the lighting of the image compared with the original picture. The new data (normal pathological images) obtained from this method is used in the statistical study of simple regression and “ANOVA” technique to detect the tumor of MRI images. After that, we study the linear regression and “ANOVA” technique by using ANOVA statistical test (equation of ANOVA: fcal) and compare it with ANOVA table(ftab) for probability p-value =0.01 (here for area 200x200, ftab=1) and see all pixels inferior to‘’1’’ that means the hypothesis ho is accepted. All these detail is to extract the place of the lesion on MRI ,(which contain matrix data of normal image and pathological ones), the extract the accepted ho pixels directly on the pathological image. The simulation program applied here is Matlab.
