% two-dimensional segmenter combined with Deep Learning

clear all   

close all

clc

addpath('functions')

%% load the image 

disp('Load MRI images');

imset='training';

%imset='online';

%imset='validation';

name1=['matFiles/',imset,'_data'];

load (name1)

%%

patient=2;

min_sn=sum(slice_per_patient(1:patient-1))+1;

max_sn=sum(slice_per_patient(1:patient-1))+slice_per_patient(patient);

step1=ceil((max_sn-min_sn)/5);

sn_range=min_sn:step1:max_sn;

slice_num=sn_range(4);

%slice_num=110;

% maximum iteration of active contour

max_its = 150;

intEweight=0.5;

DLweight=0.1;

Dynamic_Window=0;

% enable plots

dis_ena=1;

% region of interest size

Mroi=100;

% get dicom info

dicom_path=['dcom/validation/ED',num2str(patient),'/images'];

para=get_dicominfo(dicom_path);

display(['processing slice# ',num2str(slice_num)])

I=t_I(:,:,slice_num);

% cut ROI from the image

subI=t_Iroi(:,:,slice_num);

% changing the size of the ROI if we want, useful for apex slices

roi_x=round((size(subI,1)-Mroi)/2)+1:round((size(subI,1)+Mroi)/2);

roi_y=round((size(subI,2)-Mroi)/2)+1:round((size(subI,2)+Mroi)/2);

subI2(:,:,slice_num)=subI(roi_y,roi_x);

m_cnt=t_centers{slice_num};

% run DL-LV segmentation to get phi_0

load DLconfigure/DL_LV_params.mat;

DL_prms_fn='DLconfigure/DL_LV_params.mat';

init_mask1=DLN(subI2(:,:,slice_num),stackedAEOptTheta,inputSize,hiddenSizeL1,hiddenSizeL2,outputSize,netconfig);

init_mask2=clean_segs(init_mask1);

init_mask(:,:,slice_num)=remap_mask(init_mask2,size(subI)/2+1,subI);

% run segmentation

[LV_seg1,phi,m_cnt] = region_seg_subPhi(I,subI2(:,:,slice_num),m_cnt,init_mask1,max_its,intEweight,DLweight,Dynamic_Window,0,DL_prms_fn);

LV_seg2=clean_segs(LV_seg1);

out_mask = remap_mask(LV_seg2,size(subI)/2+1,subI);

LV_seg(:,:,slice_num)=out_mask;

% compute metrics

manualPoints=t_contours{slice_num};

[dm1,dm2,PD1,PD2,HD1,HD2]=eval_metrics(I,m_cnt,LV_seg(:,:,slice_num),manualPoints,para); 

% special care for slices with large Perp Distance

if PD2>= 3 

    edited_seg=edit_prior_shape(subI2(:,:,slice_num),LV_seg(:,:,slice_num),0);

    [dm11,dm22,PD11,PD22,HD11,HD22]=eval_metrics(I,m_cnt,edited_seg,manualPoints,para); 

    if PD22<PD2

        dm1=dm11;dm2=dm22;PD1=PD11;PD2=PD22;HD1=HD11;HD2=HD22;

        LV_seg(:,:,slice_num)=edited_seg;

    end

end

%% evaluate segmentations 

DM1=dm1

DM2=dm2

%% show only one slice

C1=t_contours{slice_num};Ct=scaleContour(C1,t_centers{slice_num},100);

stn=[imset,':patient:',num2str(patient),' slice:',num2str(slice_num)];

figure 

showCurveAndPhi(t_Iroi(:,:,slice_num),Ct ,(LV_seg(:,:,slice_num)), slice_num);

ylabel(stn);

figure

showCurveAndPhi(t_Iroi(:,:,slice_num),Ct ,bwconvhull(LV_seg(:,:,slice_num)), slice_num);

ylabel(stn);