Ensemble Learning-Based Hybrid Segmentation of Mammographic Images for Breast Cancer Risk Prediction Using Fuzzy C-Means and CNN Model

Algorithm 1

Process to design a Graphical User Interface of the proposed method

	To design a Graphical User Interface
	IM = imresize(I,.3)//Input an image by renaming it as IM and resizing it if required.
	GR = ImageIn(:,:,3);//Apply Complement using the green channel on input image IM.
	GRC = imgcomplement(GR);
	axxes(handlesi.greenimg_channel);
	set(imgshow(GRC));
	CLH = adaptohisteq(GRC);//To apply CLAHE on GRC to receiveCLH % contrast limited image.
	set(imshow(CLH));
	SE = strel(“ball”,8,8);//Perform structuring of an element with the specified neighbourhood (8).
	mgopen = imgopen(CLH,SE);//Do morphological on binary image CLH with structuring element ‘sse’.
	gordisk = CLH - gimgopen;//Replace optic disk by appling a 2D Median Filter.
	medfillt = medfillt2(gordisk);
	backgroundimg = imgopen(medfillt, strell(“disk”,160));
	IM2 = GC1;//Eliminate background for adjustment of image to retrieve GC1.
	IM2 = double(IM2);//The I/P image (GC1) by using Fuzzy C-Means, does image segmentation.
	//Execute the above segmentation to construct a flat structure element with in the specified neighbourhood.
	backgroundimg = imopenimg(IMMM, strell(“disk”, 46));
	//Eliminate all connected components having less than 40 pixels to create new binary image I5 from a binary image and is called as an area opening
	I5=IMMMM-backgroundimg;
	I5 = bwareaopenimg(I5,30);
	axess(handless.segmented_img);//Open an axis at the specified position and return a handle to it.
	set(imshowimg(I5));
	set(LTprojectt.segmented_img,“Userdata”,I5);
	ffcmm1=(['The value of Cluster1 = 'num2str(cccc1)]);//Retrieve the final image I5 to find cluster.
	ffcmm2=(['The value of Cluster2 = 'num2str(cccc2)]);
	classIM_1 = Imgg(:);//Find image vectors of input image (IM) and segmented image (I5)
	classI5_2 = Imgg1(:);
	//To detect errors set all default parameters
	//Evaluate the threshold values among the data points.
	% Sort data points %
	ss_data = unique(sort([class_1; class_2]));
	% Del NaN values %
	ss_data(isnan(ss_data)) = [];
	% Cal difference between consecutive points %
	dd_data = diff(ss_data);
	% Cal last point %
	dd_data(length(d_data)+1,1) = dd_data(length(d_data));
	% Cal first point %
	thresh(1,1) = ss_data(1) - dd_data(1);
	% Cal Threshold %
	thres(2:len(s_data)+1,1) = s_data + d_data./2;
	cur = zeross(sizeof(thresh,1),2);//Find sensibility and specificity of every threshold value
	dis = zeross(sizeof(thresh,1),1);
	for idd_t = 1:1:len(thresh)
	TruePositive = len(find(class2≥thresh(idd_t)));
	FalsePositive = len(find(class1 ≥ thresh(idd_t)));
	FalseNegative = len(find(class2 <thresh(idd_t)));
	TrueNegative = len(find(class1 <thresh(idd_t)));
	S = TruePositive/(TruePositive + FalseNegative);
	SP = curve(idd_t1,2) = TrueNegative/(TrueNegative + FalsePositive);
	//Calculate distance between every point and optimum point ranging [0,1]
	distance(idd_t1) = sqrt((1-curve(idd_t1,1))^2+(curve(idd_t1,2)-1)^2);
	Calculate the best value for threshold position, Sensitivity, Specificity, Area under curve, Accuracy, all false and true positives and negatives.