forked from e0404/matRad
-
Notifications
You must be signed in to change notification settings - Fork 0
/
matRad_engelLeafSequencing.m
386 lines (293 loc) · 14.3 KB
/
matRad_engelLeafSequencing.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
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
function resultGUI = matRad_engelLeafSequencing(resultGUI,stf,dij,numOfLevels,visBool)
% multileaf collimator leaf sequencing algorithm for intensity modulated
% beams with multiple static segments accroding to Engel et al. 2005
% Discrete Applied Mathematics
%
% call
% resultSequencing = matRad_engelSequencing(w,stf,pln,numOfLevels,visBool)
%
% input
% resultGUI: resultGUI struct to which the output data will be added, if
% this field is empty resultGUI struct will be created
% stf: matRad steering information struct
% dij: matRad's dij matrix
% numOfLevels: number of stratification levels
% visBool: toggle on/off visualization (optional)
%
% output
% resultGUI: matRad result struct containing the new dose cube
% as well as the corresponding weights
%
% References
% [1] http://www.sciencedirect.com/science/article/pii/S0166218X05001411
%
% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
% Copyright 2015 the matRad development team.
%
% This file is part of the matRad project. It is subject to the license
% terms in the LICENSE file found in the top-level directory of this
% distribution and at https://github.com/e0404/matRad/LICENSES.txt. No part
% of the matRad project, including this file, may be copied, modified,
% propagated, or distributed except according to the terms contained in the
% LICENSE file.
%
% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% if visBool not set toogle off visualization
if nargin < 5
visBool = 0;
end
numOfBeams = numel(stf);
if visBool
% create the sequencing figure
sz = [800 1000]; % figure size
screensize = get(0,'ScreenSize');
xpos = ceil((screensize(3)-sz(2))/2); % center the figure on the screen horizontally
ypos = ceil((screensize(4)-sz(1))/2); % center the figure on the screen vertically
seqFig = figure('position',[xpos,ypos,sz(2),sz(1)]);
end
offset = 0;
for i = 1:numOfBeams
numOfRaysPerBeam = stf(i).numOfRays;
% get relevant weights for current beam
wOfCurrBeams = resultGUI.w(1+offset:numOfRaysPerBeam+offset);
X = ones(numOfRaysPerBeam,1)*NaN;
Z = ones(numOfRaysPerBeam,1)*NaN;
for j=1:stf(i).numOfRays
X(j) = stf(i).ray(j).rayPos_bev(:,1);
Z(j) = stf(i).ray(j).rayPos_bev(:,3);
end
% sort bixels into matrix
minX = min(X);
maxX = max(X);
minZ = min(Z);
maxZ = max(Z);
dimOfFluenceMxX = (maxX-minX)/stf(i).bixelWidth + 1;
dimOfFluenceMxZ = (maxZ-minZ)/stf(i).bixelWidth + 1;
%Create the fluence matrix.
fluenceMx = zeros(dimOfFluenceMxZ,dimOfFluenceMxX);
% Calculate X and Z position of every fluence's matrix spot
% z axis = axis of leaf movement!
xPos = (X-minX)/stf(i).bixelWidth+1;
zPos = (Z-minZ)/stf(i).bixelWidth+1;
% Make subscripts for fluence matrix
indInFluenceMx = zPos + (xPos-1)*dimOfFluenceMxZ;
%Save weights in fluence matrix.
fluenceMx(indInFluenceMx) = wOfCurrBeams;
% Stratification
calFac = max(fluenceMx(:));
D_k = round(fluenceMx/calFac*numOfLevels);
% Save the stratification in the initial intensity matrix D_0.
D_0 = D_k;
% container to remember generated shapes; allocate space for 10000 shapes
shapes = NaN*ones(dimOfFluenceMxZ,dimOfFluenceMxX,10000);
k = 0;
if visBool
clf(seqFig);
colormap(seqFig,'jet');
seqSubPlots(1) = subplot(2,2,1,'parent',seqFig);
imagesc(D_k,'parent',seqSubPlots(1));
set(seqSubPlots(1),'CLim',[0 numOfLevels],'YDir','normal');
title(seqSubPlots(1),['Beam # ' num2str(i) ': max(D_0) = ' num2str(max(D_0(:))) ' - ' num2str(numel(unique(D_0))) ' intensity levels']);
xlabel(seqSubPlots(1),'x - direction parallel to leaf motion ')
ylabel(seqSubPlots(1),'z - direction perpendicular to leaf motion ')
colorbar;
drawnow
end
% start sequencer
while max(D_k(:) > 0)
%calculate the difference matrix diffMat
diffMat = diff([zeros(size(D_k,1),1) D_k zeros(size(D_k,1),1)],[],2);
%calculate complexities
c = sum(max(0,diffMat),2); %TNMU-row-complexity
com = max(c); %TNMU complexity
g = com - c; %row complexity gap
%initialize segment
segment = zeros(size(D_k));
k = k + 1;
%Plot residual intensity matrix.
if visBool
seqSubPlots(2) = subplot(2,2,2,'parent',seqFig);
imagesc(D_k,'parent',seqSubPlots(2));
set(seqSubPlots(2),'CLim',[0 numOfLevels],'YDir','normal');
title(seqSubPlots(2),['k = ' num2str(k)]);
colorbar
drawnow
end
%loop over all rows
for j=1:size(D_0,1)
%determine essential intervals
data(j).left(1) = 0; %left interval limit, actual for an empty interval
data(j).right(1) = 0; %right interal limit, actual for an empty interval
data(j).v(1) = g(j); %greatest number such that the inequalities (6) resp. (7) is satisfied with u=v
data(j).w(1) = inf; %smallest number in the interval
data(j).u(1) = data(j).v(1); %min(v,w)
[~, pos, ~] = find(diffMat(j,:) > 0); % indices of all positive elements in the j. row of diffmat
[~, neg, ~] = find(diffMat(j,:) < 0); % indices of all negative elements in the j. row of diffMat
n=2;
%loop over the positive elements in the j. row of diffmat ->
%possible left interval limits
for m=1:size(pos,2)
%loop over the negative elements in the j. row of diffMat ->
%possible right interval limit
for l=1:size(neg,2)
%take only intervals I=[l,r] with l<=r
if pos(m) <= neg(l)-1
%set interval limits
data(j).left(n) = pos(m);
data(j).right(n) = neg(l)-1;
%calculate v according to Lemma 8
if g(j) <= abs( diffMat(j,pos(m)) + diffMat(j,neg(l)) )
data(j).v(n) = min( diffMat(j,pos(m)), -diffMat(j,neg(l)) ) + g(j);
else
data(j).v(n) = ( diffMat(j, pos(m)) - diffMat(j, neg(l)) + g(j)) / 2;
end
%calculate w and u according to equality (11) and
%(12)
data(j).w(n) = min(D_k(j,pos(m):(neg(l)-1)));
data(j).u(n) = min(data(j).v(n), data(j).w(n));
n = n+1;
end
end
end
u(j) = max(data(j).u);
end
%calculate u_max from theorem 9
d_k = min(u);
%loop over all rows
for j=1:size(D_0,1)
%find all possible (and essential) intervals
candidate = find(data(j).u >= d_k);
%calculate the potential of the possible intervals
%initialize p as -Inf
data(j).p(1:length(data(j).left)) = -Inf;
%loop over all possible intervals
for s=1:size(candidate,2)
if (s==1 && data(j).left(candidate(s)) == 0)
data(j).p(candidate(1)) = 0;
else
%calculate p1 according to equality (17)
if (d_k == diffMat(j, data(j).left(candidate(s))) && d_k ~= D_k(j, data(j).left(candidate(s))))
p1 = 1;
else
p1 = 0;
end
%calculate p2 according to equalitiy (18)
% if data(j).right(candidate(s)) < size(D_0, 2)
if (d_k == -diffMat(j, data(j).right(candidate(s))+1) && d_k ~= D_k(j, data(j).right(candidate(s))))
p2 = 1;
else
p2 = 0;
end
% else
%
% if d_k == -diffMat(j, data(j).right(candidate(s))+1)
% p2 = 1;
% else
% p2 = 0;
% end
%
% end
%calculate p3 according to equality (19)
p3 = size(find(D_k(j, data(j).left(candidate(s)):data(j).right(candidate(s))) == d_k),2);
data(j).p(candidate(s)) = p1 + p2+ p3;
end
end
%determinate intervals with maximum potential
maxPot = find(data(j).p == max(data(j).p));
%if several intervals have maximum potential, select
%the interval which has maximum length
if size(maxPot,2) > 1
for t=1:size(maxPot,2)
if t==1 && data(j).left(maxPot(t)) == 0
data(j).l(1) = 0;
else
data(j).l(maxPot(t)) = data(j).right(maxPot(t)) - data(j).left(maxPot(t)) + 1;
end
end
%data(j).l(maxPot) = data(j).right(maxPot) - data(j).left(maxPot) + 1;
maxLength = find(data(j).l == max(data(j).l));
%left and right interval limits of the selected
%interval
leftIntLimit(j) = data(j).left(maxLength(1));
rightIntLimit(j) = data(j).right(maxLength(1));
else
%left and right interval limits of the selected
%interval
leftIntLimit(j) = data(j).left(maxPot);
rightIntLimit(j) = data(j).right(maxPot);
end
%create segment associated by the selected interval
if leftIntLimit(j) ~= 0
segment(j,leftIntLimit(j):rightIntLimit(j)) = 1;
end
end
%write the segment in shape_k
shape_k = segment;
%show the leaf positions
if visBool
seqSubPlots(4) = subplot(2,2,3.5,'parent',seqFig);
imagesc(shape_k,'parent',seqSubPlots(4));
hold(seqSubPlots(4),'on');
set(seqSubPlots(4),'YDir','normal')
xlabel(seqSubPlots(4),'x - direction parallel to leaf motion ')
ylabel(seqSubPlots(4),'z - direction perpendicular to leaf motion ')
title(seqSubPlots(4),['beam # ' num2str(i) ' shape # ' num2str(k) ' d_k = ' num2str(d_k)]);
for j = 1:dimOfFluenceMxZ
leftLeafIx = find(shape_k(j,:)>0,1,'first');
rightLeafIx = find(shape_k(j,:)>0,1,'last');
if leftLeafIx > 1
plot(seqSubPlots(4),[.5 leftLeafIx-.5],j-[.5 .5] ,'w','LineWidth',2)
plot(seqSubPlots(4),[.5 leftLeafIx-.5],j+[.5 .5] ,'w','LineWidth',2)
plot(seqSubPlots(4),[ leftLeafIx-.5 leftLeafIx-.5],j+[.5 -.5] ,'w','LineWidth',2)
end
if rightLeafIx<dimOfFluenceMxX
plot(seqSubPlots(4),[dimOfFluenceMxX+.5 rightLeafIx+.5],j-[.5 .5] ,'w','LineWidth',2)
plot(seqSubPlots(4),[dimOfFluenceMxX+.5 rightLeafIx+.5],j+[.5 .5] ,'w','LineWidth',2)
plot(seqSubPlots(4),[ rightLeafIx+.5 rightLeafIx+.5],j+[.5 -.5] ,'w','LineWidth',2)
end
if isempty(rightLeafIx) && isempty (leftLeafIx)
plot(seqSubPlots(4),[dimOfFluenceMxX+.5 .5],j-[.5 .5] ,'w','LineWidth',2)
plot(seqSubPlots(4),[dimOfFluenceMxX+.5 .5],j+[.5 .5] ,'w','LineWidth',2)
plot(seqSubPlots(4),.5*dimOfFluenceMxX*[1 1]+[0.5],j+[.5 -.5] ,'w','LineWidth',2)
end
end
axis tight
drawnow
pause(1);
end
%save shape_k in container
shapes(:,:,k) = shape_k;
%save the calculated MU
shapesWeight(k) = d_k;
%calculate new matrix, the diference matrix and complexities
D_k = D_k - d_k*shape_k;
%delete variables
clear data;
clear segment;
clear u;
clear leftIntLimit;
clear rightIntLimit;
end
sequencing.beam(i).numOfShapes = k;
sequencing.beam(i).shapes = shapes(:,:,1:k);
sequencing.beam(i).shapesWeight = shapesWeight(1:k)/numOfLevels*calFac;
sequencing.beam(i).bixelIx = 1+offset:numOfRaysPerBeam+offset;
sequencing.beam(i).fluence = D_0;
sequencing.w(1+offset:numOfRaysPerBeam+offset,1) = D_0(indInFluenceMx)/numOfLevels*calFac;
offset = offset + numOfRaysPerBeam;
end
resultGUI.w = sequencing.w;
resultGUI.wSequenced = sequencing.w;
resultGUI.sequencing = sequencing;
resultGUI.apertureInfo = matRad_sequencing2ApertureInfo(sequencing,stf);
doseSequencedDoseGrid = reshape(dij.physicalDose{1} * sequencing.w,dij.doseGrid.dimensions);
% interpolate to ct grid for visualiation & analysis
resultGUI.physicalDose = matRad_interp3(dij.doseGrid.x,dij.doseGrid.y',dij.doseGrid.z, ...
doseSequencedDoseGrid, ...
dij.ctGrid.x,dij.ctGrid.y',dij.ctGrid.z);
% if weights exists from an former DAO remove it
if isfield(resultGUI,'wDao')
resultGUI = rmfield(resultGUI,'wDao');
end
end