-
Notifications
You must be signed in to change notification settings - Fork 1
/
basic_mea.m
261 lines (200 loc) · 14.1 KB
/
basic_mea.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
%%% written by D.S.JOKHUN on 04/09/2018
function [Basic_Measurements,Boundary_RadOfCurva_K,Norm_Pix_Intensities]=basic_mea(segmented_raw_2d,MetaData)
MetaData.Num_of_Pixels_X=size(segmented_raw_2d,2);
MetaData.Num_of_Pixels_Y=size(segmented_raw_2d,1);
Filename=cell(size(segmented_raw_2d,3),1);
Nuc_label=zeros(size(segmented_raw_2d,3),1);
Centroid=zeros(size(segmented_raw_2d,3),2);
Pro_area=zeros(size(segmented_raw_2d,3),1);
Perimeter=zeros(size(segmented_raw_2d,3),1);
AR=zeros(size(segmented_raw_2d,3),1);
Shape_factor=zeros(size(segmented_raw_2d,3),1);
PDI=zeros(size(segmented_raw_2d,3),1);;
Centre_mismatch=zeros(size(segmented_raw_2d,3),1);
I80_by_I20=zeros(size(segmented_raw_2d,3),1);
nHigh_by_nLow=zeros(size(segmented_raw_2d,3),1);
Mean_Norm_Int=zeros(size(segmented_raw_2d,3),1);
Median_Norm_Int=zeros(size(segmented_raw_2d,3),1);
SD_Norm_Int=zeros(size(segmented_raw_2d,3),1);
p_of_IntDistri_being_Normal=zeros(size(segmented_raw_2d,3),1);
Entropy=zeros(size(segmented_raw_2d,3),1);
Relative_concavity=zeros(size(segmented_raw_2d,3),1);
AvgCurva_X_Peri=zeros(size(segmented_raw_2d,3),1);
SDinCurva=zeros(size(segmented_raw_2d,3),1);
AvgPosCurva_X_Peri=zeros(size(segmented_raw_2d,3),1);
AvgNegCurva_X_Peri=zeros(size(segmented_raw_2d,3),1);
n_CurvaChangesSign=zeros(size(segmented_raw_2d,3),1);
MaxPosCurva=zeros(size(segmented_raw_2d,3),1);
nPeaks_Within80PercOf_MaxPos=zeros(size(segmented_raw_2d,3),1);
AvgProminance_of_PeaksWithin80PercOf_MaxPos=zeros(size(segmented_raw_2d,3),1);
AvgWidth_of_PeaksWithin80PercOf_MaxPos=zeros(size(segmented_raw_2d,3),1);
MaxNegCurva=zeros(size(segmented_raw_2d,3),1);
nPeaks_Within80PercOf_MaxNeg=zeros(size(segmented_raw_2d,3),1);
AvgProminance_of_PeaksWithin80PercOf_MaxNeg=zeros(size(segmented_raw_2d,3),1);
AvgWidth_of_PeaksWithin80PercOf_MaxNeg=zeros(size(segmented_raw_2d,3),1);
Boundary_RadOfCurva_K=cell(size(segmented_raw_2d,3),1);
for nuc_count=1:size(segmented_raw_2d,3);
Filename{nuc_count,1}=MetaData.Filename;
raw=segmented_raw_2d(:,:,nuc_count);
bw=raw>0;
stats=regionprops(bw,raw,'PixelIdxList','PixelValues','Area','MajorAxisLength','MinorAxisLength','Perimeter', 'Centroid','WeightedCentroid', 'ConvexArea');
if size(stats,1)>1
area_check=[];
template=zeros(size(raw));
for area_count=1:size(stats,1)
area_check(area_count)=stats(area_count).Area;
end
[~,id]=max(area_check);
template(stats(id).PixelIdxList)=1;
raw=raw.*uint16(template);
end
bw=raw>0;
stats=regionprops(bw,raw,'PixelIdxList','PixelValues','Area','MajorAxisLength','MinorAxisLength','Perimeter', 'Centroid','WeightedCentroid', 'ConvexArea');
Norm_Pix_Intensities{1,nuc_count}=MetaData.Filename;
Norm_Pix_Intensities(3:2+size(stats.PixelValues,1),nuc_count)=num2cell(mat2gray(stats.PixelValues));
Mean_Norm_Int(nuc_count,1)=mean(mat2gray(stats.PixelValues));
Median_Norm_Int(nuc_count,1)=median(mat2gray(stats.PixelValues));
SD_Norm_Int(nuc_count,1)=std(mat2gray(stats.PixelValues));
[~,p_of_IntDistri_being_Normal(nuc_count,1)] = kstest(zscore(mat2gray(stats.PixelValues)));
Centroid(nuc_count,1)=stats.Centroid(1)*MetaData.Voxel_Size_X;
Centroid(nuc_count,2)=stats.Centroid(2)*MetaData.Voxel_Size_Y;
Pro_area(nuc_count,1)=(stats.Area * (MetaData.Voxel_Size_X*MetaData.Voxel_Size_Y));
Perimeter(nuc_count,1)=(stats.Perimeter * MetaData.Voxel_Size_X);
AR(nuc_count,1)=(stats.MajorAxisLength/stats.MinorAxisLength);
Shape_factor(nuc_count,1)=((stats.Perimeter^2)/(4*pi*stats.Area));
%% PDI
distance_frm_cen_squared=zeros(MetaData.Num_of_Pixels_Y, MetaData.Num_of_Pixels_X);
for countX=1:MetaData.Num_of_Pixels_X
for countY=1:MetaData.Num_of_Pixels_Y
distance_frm_cen_squared(countY,countX)=((sqrt(((countX-stats.Centroid(1))^2)...
+(((MetaData.Num_of_Pixels_Y-countY+1)-stats.Centroid(2))^2)))...
*MetaData.Voxel_Size_X).^2;
end
end
actual_IntMoment_2=distance_frm_cen_squared.*mat2gray(raw);
total_NormInt=sum(sum(mat2gray(raw)));
uniform_IntMoment_2=distance_frm_cen_squared.*(double(bw)*(total_NormInt/stats.Area));
PDI(nuc_count,1)=sum(sum(actual_IntMoment_2))/sum(sum(uniform_IntMoment_2)); %peripheral distributon index
%% Centre mismatch
Centre_mismatch(nuc_count,1)=(sqrt(((stats.WeightedCentroid(1)-stats.Centroid(1))^2)...
+((stats.WeightedCentroid(2)-stats.Centroid(2))^2)))...
*MetaData.Voxel_Size_X;
%% Intensity Histogram Analysis
I_exclude_percentiles=prctile(single(nonzeros(raw)),[0.1,99.9]); %elimimating extreme values from the image (e.g saturated pixels etc.)
aft_excl_extremes=raw.*uint16(raw>=I_exclude_percentiles(1)).*uint16(raw<=I_exclude_percentiles(2));
I_percentiles=prctile(single(nonzeros(aft_excl_extremes)),[20,80]);
I80_by_I20(nuc_count,1)=I_percentiles(2)/I_percentiles(1);
normalize_aft_excl_extremes=mat2gray(nonzeros(aft_excl_extremes));
nHigh_by_nLow(nuc_count,1)=sum(normalize_aft_excl_extremes>=0.8)/sum(normalize_aft_excl_extremes<=0.2);
Entropy(nuc_count,1)=entropy(normalize_aft_excl_extremes);
%% global curvature analysis
Relative_concavity(nuc_count,1)=(stats.ConvexArea-stats.Area)/stats.ConvexArea;
%% local curvature analysis
% % % boundary=bwboundaries(bw);
% % % boundary_xy=[];
% % % boundary_xy(:,1)=boundary{1,1}(1:end-1,2); %The last point is a repeat of the 1st point.
% % % boundary_xy(:,2)=boundary{1,1}(1:end-1,1);
% % % smooth_span=2/MetaData.Voxel_Size_X; %smooth span of 2um
% % % cond = mod(smooth_span,2)<1; % =1 if remainder is less than 1 (rounding down will bring it to an even num) and =0 if remainder is more than 1 (rounding down will bring it to an odd number).
% % % smooth_span = floor(smooth_span);
% % % smooth_span = smooth_span+cond; % if cond was 1, 1 will be added to floor(span). if cond was 0, floor(span) is already odd and nothing is added.
% % % %padding the x-y series for proper smoothing and proper subsequent fitting
% % % boundary_xy((2*smooth_span)+1:(2*smooth_span)+size(boundary_xy,1),:)=boundary_xy(:,:); %sifting the series by some rows
% % % boundary_xy(1:(2*smooth_span),:)=boundary_xy(end-(2*smooth_span)+1:end,:); %padding the top
% % % boundary_xy(end+1:end+(2*smooth_span),:)=boundary_xy((2*smooth_span)+1:(2*smooth_span)+(2*smooth_span),:); %padding the bottom
% % % boundary_xy(:,1)=smooth(boundary_xy(:,1),smooth_span,'lowess');
% % % boundary_xy(:,2)=smooth(boundary_xy(:,2),smooth_span,'lowess');
% % %
% % % [MetaData.Filename,' nuc No.',num2str(nuc_count),' of ',num2str(size(segmented_raw_2d,3)),' -boundary analysis']
% % % fit_span=1/MetaData.Voxel_Size_X; %%tangent 1 will be taken over 1um stretch before point P and tangent 2 will be take 1um stretch after point P.
% % % cond = mod(fit_span,2)>=1; % =1 if remainder is more or equal to 1 (rounding down will bring it to an odd num) and =0 if remainder is less than 1 (rounding down will bring it to an even number).
% % % fit_span = floor(fit_span);
% % % fit_span = fit_span+cond; % if cond was 1, 1 will be added to floor(span). if cond was 0, floor(span) is already even and nothing is added.
% % %
% % % for point_count=(2*smooth_span)+1:size(boundary_xy,1)-(2*smooth_span);
% % % % point_count
% % % line_1=fit(boundary_xy(point_count-fit_span:point_count,1),boundary_xy(point_count-fit_span:point_count,2),'poly1');
% % % coeffvalues_1=coeffvalues(line_1);
% % % syms x y %declaring system of equations
% % % if abs(coeffvalues_1(1,1))>10^10
% % % norm_eqn_1 = y==boundary_xy(point_count-(fit_span/2),2);
% % % norm_grad_1=0;
% % % end
% % % if abs(coeffvalues_1(1,1))<10^-10
% % % norm_eqn_1 = x==boundary_xy(point_count-(fit_span/2),1);
% % % norm_grad_1=Inf;
% % % end
% % % if abs(coeffvalues_1(1,1)) <=10^10
% % % if abs(coeffvalues_1(1,1)) >=10^-10
% % % norm_grad_1=-1/(coeffvalues_1(1,1));
% % % c1=(boundary_xy(point_count-(fit_span/2),2)) - (norm_grad_1*(boundary_xy(point_count-(fit_span/2),1))); %c=y-mx
% % % norm_eqn_1 = y==(norm_grad_1*x)+c1;
% % % end
% % % end
% % %
% % % line_2=fit(boundary_xy(point_count:point_count+fit_span,1),boundary_xy(point_count:point_count+fit_span,2),'poly1');
% % % coeffvalues_2=coeffvalues(line_2);
% % % if abs(coeffvalues_2(1,1))>10^10
% % % norm_eqn_2 = y==boundary_xy(point_count+(fit_span/2),2);
% % % norm_grad_2=0;
% % % end
% % % if abs(coeffvalues_2(1,1))<10^-10
% % % norm_eqn_2 = x==boundary_xy(point_count+(fit_span/2),1);
% % % norm_grad_2=Inf;
% % % end
% % % if abs(coeffvalues_2(1,1)) <=10^10
% % % if abs(coeffvalues_2(1,1)) >=10^-10
% % % norm_grad_2=-1/(coeffvalues_2(1,1));
% % % c2=(boundary_xy(point_count+(fit_span/2),2)) - (norm_grad_2*(boundary_xy(point_count+(fit_span/2),1))); %c=y-mx
% % % norm_eqn_2 = y==(norm_grad_2*x)+c2;
% % % end
% % % end
% % %
% % % if round(norm_grad_1,10) == round(norm_grad_2,10)
% % % radius_of_curvature=Inf;
% % % else
% % % sol = solve([norm_eqn_1, norm_eqn_2], [x, y]);
% % % centre_of_curvature=double([sol.x,sol.y]);
% % % radius_of_curvature=double((sqrt(((boundary_xy(point_count,1)-sol.x)^2)+((boundary_xy(point_count,2)-sol.y)^2))))*MetaData.Voxel_Size_X;
% % %
% % % cen_of_curva_to_pt=[boundary_xy(point_count,1)-centre_of_curvature(1,1),boundary_xy(point_count,2)-centre_of_curvature(1,2)]; %vector to move from centre of curvature to point P on the boundary
% % % unit_vec=cen_of_curva_to_pt./norm(cen_of_curva_to_pt); % unit vector to move from centre of curvature to point P on the boundary
% % % query_point=[boundary_xy(point_count,1)+(2*unit_vec(1,1)),boundary_xy(point_count,2)+(2*unit_vec(1,2))];
% % % if inpolygon(query_point(1,1),query_point(1,2),boundary_xy(smooth_span+1:length(boundary_xy)-smooth_span,1),boundary_xy(smooth_span+1:length(boundary_xy)-smooth_span,2))==1 %If point P+2*(unit vector from the centre of curvature to point P) lies inside the cell, this statement will be true
% % % radius_of_curvature=-radius_of_curvature;
% % % end
% % % end
% % % boundary_xy(point_count,3)=radius_of_curvature;
% % % boundary_xy(point_count,4)=1/radius_of_curvature;
% % % end
Boundary_RadOfCurva_K_header={'x(pix)' 'y(pix)' 'Radius of curvature(um)' 'curvature,K=1/r(um^-1)'};
pre_Boundary_RadOfCurva_K=[0 0 0 0];%boundary_xy((2*smooth_span)+1:end-(2*smooth_span),1:4);
Boundary_RadOfCurva_K{nuc_count,1}=vertcat(Boundary_RadOfCurva_K_header,num2cell(pre_Boundary_RadOfCurva_K));
AvgCurva_X_Peri(nuc_count,1)=0;%mean(pre_Boundary_RadOfCurva_K(:,4))*Perimeter(nuc_count,1);
SDinCurva(nuc_count,1)=0;%std(pre_Boundary_RadOfCurva_K(:,4));
AvgPosCurva_X_Peri(nuc_count,1)=0;% (mean(nonzeros(pre_Boundary_RadOfCurva_K(:,4).*(pre_Boundary_RadOfCurva_K(:,4)>0)))) *Perimeter(nuc_count,1);
AvgNegCurva_X_Peri(nuc_count,1)=0;% (mean(nonzeros(pre_Boundary_RadOfCurva_K(:,4).*(pre_Boundary_RadOfCurva_K(:,4)<0)))) *Perimeter(nuc_count,1);
% % % SignTest=diff((nonzeros(pre_Boundary_RadOfCurva_K(:,4)))>0);
n_CurvaChangesSign(nuc_count,1)=0;%numel(nonzeros(SignTest));
MaxPosCurva(nuc_count,1)= 0;%max(findpeaks(pre_Boundary_RadOfCurva_K(:,4)));
% % % [pks,~,w,p] =findpeaks(pre_Boundary_RadOfCurva_K(:,4),'MinPeakDistance',round(3/MetaData.Voxel_Size_X),'MinPeakHeight',(0.8*MaxPosCurva(nuc_count,1))); %If 2 peaks are reparated by less than 3um along the boundary, they are counted as 1
nPeaks_Within80PercOf_MaxPos(nuc_count,1)=0;% numel(pks);
AvgProminance_of_PeaksWithin80PercOf_MaxPos(nuc_count,1)=0;%mean(p);
AvgWidth_of_PeaksWithin80PercOf_MaxPos(nuc_count,1)=0;%mean(w);
MaxNegCurva(nuc_count,1)=0;% -1*(max(findpeaks(-1*(pre_Boundary_RadOfCurva_K(:,4)))));
% % % [pks,~,w,p] =findpeaks(-1*(pre_Boundary_RadOfCurva_K(:,4)),'MinPeakDistance',round(3/MetaData.Voxel_Size_X),'MinPeakHeight',abs(0.8*MaxNegCurva(nuc_count,1)));
nPeaks_Within80PercOf_MaxNeg(nuc_count,1)=0;% numel(pks);
AvgProminance_of_PeaksWithin80PercOf_MaxNeg(nuc_count,1)=0;%mean(p);
AvgWidth_of_PeaksWithin80PercOf_MaxNeg(nuc_count,1)=0;%mean(w);
% figure ('Name',['Curvature,K of ',MetaData.Filename])
% colormap jet
% patch(pre_Boundary_RadOfCurva_K(:,1),pre_Boundary_RadOfCurva_K(:,2),pre_Boundary_RadOfCurva_K(:,4),'EdgeColor','interp','FaceColor','none','LineWidth',5)
%
% figure ('Name',['Curvature,K of ',MetaData.Filename])
% plot(pre_Boundary_RadOfCurva_K(:,4))
end
Basic_Measurements=horzcat(Filename,num2cell(Nuc_label),num2cell(Centroid),num2cell(Pro_area),num2cell(Perimeter),num2cell(AR),num2cell(Shape_factor),num2cell(PDI),num2cell(Centre_mismatch),num2cell(I80_by_I20),num2cell(nHigh_by_nLow),num2cell(Mean_Norm_Int),num2cell(Median_Norm_Int),num2cell(SD_Norm_Int),num2cell(p_of_IntDistri_being_Normal),num2cell(Entropy),...
num2cell(Relative_concavity),num2cell(AvgCurva_X_Peri),num2cell(SDinCurva),num2cell(AvgPosCurva_X_Peri),num2cell(AvgNegCurva_X_Peri),num2cell(n_CurvaChangesSign)...
,num2cell(MaxPosCurva),num2cell(nPeaks_Within80PercOf_MaxPos),num2cell(AvgProminance_of_PeaksWithin80PercOf_MaxPos),num2cell(AvgWidth_of_PeaksWithin80PercOf_MaxPos)...
,num2cell(MaxNegCurva),num2cell(nPeaks_Within80PercOf_MaxNeg),num2cell(AvgProminance_of_PeaksWithin80PercOf_MaxNeg),num2cell(AvgWidth_of_PeaksWithin80PercOf_MaxNeg));
end