-
Notifications
You must be signed in to change notification settings - Fork 0
/
Produce1.m
329 lines (279 loc) · 12.4 KB
/
Produce1.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
function [bpm, RR_interval] = Produce1(Dir,filename,vidFile,fl,fh,fps)
% Author : BoYu Huang
% Date : 2018.5.15
% This is the produce.m version 2.0. Aftering using Transform_face.m
% to transform video into a fewer pixel one in order to reduce the computation
% time significantly. However, ROI still include many irrevelent background
% image part.
% Date : 2018.5.25
% Define the region of interest by proportionally selecting the forehead and cheek
% after face detection
% Date : 2018.5.27
% Add peak detection (FFT)
% https://nl.mathworks.com/help/signal/ug/find-peaks-in-data.html
% Date : 2018.7.14
% Add normalization to the signal extracted from the RGB channels
% x'(t) = [x(t) - mean] / sigma
% Date : 2018.12.3
% https://www.mathworks.com/help/matlab/ref/spline.html#bvjdpi3-xq
% Testing cubic interploation in PPG data after EVM processing
% Date : 2019.1.22
% Fix face not found problem
% Date : 2018.
% Use HSV or LAB to detect the face region
resultsDir = Dir;
mkdir(resultsDir);
global samplingRate
%% face
outDir = resultsDir;
alpha = 50;
level = 4;
samplingRate = fps; % Depend on video fps
chromAttenuation = 1;
[~,vidName] = fileparts(vidFile);
% Build the full file name frome parts
outName = fullfile(outDir,[vidName '-ideal-from-' num2str(fl) ...
'-to-' num2str(fh) ...
'-alpha-' num2str(alpha) ...
'-level-' num2str(level) ...
'-chromAtn-' num2str(chromAttenuation) '.avi']);
% Read video
vid = VideoReader(vidFile);
% Extract video info
vidHeight = vid.Height; % set the range to reduce time
vidWidth = vid.Width; % set the range to reduce time
nChannels = 3;
fr = vid.FrameRate;
len = vid.NumberOfFrames;
% Create a cascade detector object.
faceDetector = vision.CascadeObjectDetector();
% Read a video frame and run the face detector.
videoFileReader = vision.VideoFileReader(vidFile);
videoFrame = step(videoFileReader);
bbox = step(faceDetector, videoFrame); % bounding box [x,y,width,height]
if (isempty(bbox) == 1 || (bbox(3) < 200 && bbox(4) < 200))
forehead_box = [int64(0.3*vidWidth),int64(0.05*vidHeight),int64(0.4*vidWidth),int64(0.1*vidHeight)];
rcheek_box = [int64(0.15*vidWidth),int64(0.55*vidHeight),int64(0.15*vidWidth),int64(0.1*vidHeight)];
lcheek_box = [int64(0.7*vidWidth),int64(0.55*vidHeight),int64(0.15*vidWidth),int64(0.1*vidHeight)];
bbox = [0 0 vidWidth vidHeight];
else
forehead_box = [int64(bbox(1)+0.3*bbox(3)),int64(bbox(2)+0.05*bbox(4)),int64(0.4*bbox(3)),int64(0.1*bbox(4))];
rcheek_box = [int64(bbox(1)+0.15*bbox(3)),int64(bbox(2)+0.55*bbox(4)),int64(0.15*bbox(3)),int64(0.1*bbox(4))];
lcheek_box = [int64(bbox(1)+0.7*bbox(3)),int64(bbox(2)+0.55*bbox(4)),int64(0.15*bbox(3)),int64(0.1*bbox(4))];
end
% Draw the returned bounding box around the detected face.
videoFrame = insertShape(videoFrame, 'Rectangle', bbox);
videoFrame = insertShape(videoFrame, 'Rectangle',forehead_box);
videoFrame = insertShape(videoFrame, 'Rectangle',rcheek_box);
videoFrame = insertShape(videoFrame, 'Rectangle',lcheek_box);
figure; imshow(videoFrame); title('Detected face');
% create a 3d array with elements all zero(unsigned int 8 bit)
temp = struct('cdata', zeros(vidHeight, vidWidth, nChannels, 'uint8'), 'colormap', []);
startIndex = 1;
endIndex = len-10;
vidOut = VideoWriter(outName);
vidOut.FrameRate = fr;
open(vidOut) % open the video to write
% compute Gaussian blur stack
disp('Spatial filtering...')
Gdown_stack = build_GDown_stack(vidFile, startIndex, endIndex, level);
disp('Finished')
% Properties of Gdown_stack:
% the first dimension is the time axis
% the second dimension is the y axis of the video
% the third dimension is the x axis of the video
% the forth dimension is the color channel
% Temporal filtering
disp('Temporal filtering...')
filtered_stack = ideal_bandpassing(Gdown_stack, 1, fl, fh, samplingRate);
disp('Finished')
%% amplify
filtered_stack(:,:,:,1) = filtered_stack(:,:,:,1) .* alpha;
filtered_stack(:,:,:,2) = filtered_stack(:,:,:,2) .* alpha .* chromAttenuation;
filtered_stack(:,:,:,3) = filtered_stack(:,:,:,3) .* alpha .* chromAttenuation;
%% Render on the input video
disp('Rendering...')
% output video
k = samplingRate;
%% create multiple array
for i = startIndex + samplingRate : endIndex - samplingRate
k = k+1
temp.cdata = read(vid, i);
[rgbframe,~] = frame2im(temp);
rgbframe = im2double(rgbframe);
frame = rgb2ntsc(rgbframe);
filtered = squeeze(filtered_stack(k,:,:,:));
filtered = imresize(filtered,[vidHeight vidWidth]);
%% Capture the RGB data (y,x,[RGB])
% Focusing on:
% 1. forehead :
% (bbox(2)+0.05*bbox(4):bbox(2)+0.15*bbox(4),bbox(1)+0.3*bbox(1):bbox(1)+0.7*bbox(3))
% 2. right cheek :
% (bbox(2)+0.55*bbox(2):bbox(2)+0.65*bbox(4),bbox(1)+0.15*bbox(3):bbox(1)+0.3*bbox(3))
% 3. left cheek :
% (bbox(2)+0.55*bbox(2):bbox(2)+0.65*bbox(4),bbox(1)+0.7*bbox(3):bbox(1)+0.85*bbox(3))
forehead_r(1,k-samplingRate) = mean(mean(filtered(forehead_box(2):forehead_box(2)+forehead_box(4),...
forehead_box(1):forehead_box(1)+forehead_box(3),1)));
forehead_g(1,k-samplingRate) = mean(mean(filtered(forehead_box(2):forehead_box(2)+forehead_box(4),...
forehead_box(1):forehead_box(1)+forehead_box(3),2)));
%forehead_b(1,k-samplingRate) = mean(mean(filtered(forehead_box(2):forehead_box(2)+forehead_box(4),...
%forehead_box(1):forehead_box(1)+forehead_box(3),3)));
rcheek_r(1,k-samplingRate) = mean(mean(filtered(rcheek_box(2):rcheek_box(2)+rcheek_box(4),...
rcheek_box(1):rcheek_box(1)+rcheek_box(3),1)));
rcheek_g(1,k-samplingRate) = mean(mean(filtered(rcheek_box(2):rcheek_box(2)+rcheek_box(4),...
rcheek_box(1):rcheek_box(1)+rcheek_box(3),2)));
%rcheek_b(1,k-samplingRate) = mean(mean(filtered(rcheek_box(2):rcheek_box(2)+rcheek_box(4),...
%rcheek_box(1):rcheek_box(1)+rcheek_box(3),3)));
lcheek_r(1,k-samplingRate) = mean(mean(filtered(lcheek_box(2):lcheek_box(2)+lcheek_box(4),...
lcheek_box(1):lcheek_box(1)+lcheek_box(3),1)));
lcheek_g(1,k-samplingRate) = mean(mean(filtered(lcheek_box(2):lcheek_box(2)+lcheek_box(4),...
lcheek_box(1):lcheek_box(1)+lcheek_box(3),2)));
%lcheek_b(1,k-samplingRate) = mean(mean(filtered(lcheek_box(2):lcheek_box(2)+lcheek_box(4),...
%lcheek_box(1):lcheek_box(1)+lcheek_box(3),3)));
whole_r(1,k-samplingRate) = mean([forehead_r(1,k-samplingRate),rcheek_r(1,k-samplingRate),lcheek_r(1,k-samplingRate)]);
whole_g(1,k-samplingRate) = mean([forehead_g(1,k-samplingRate),rcheek_g(1,k-samplingRate),lcheek_g(1,k-samplingRate)]);
% convert to HSV channel
% h = hsv(:, :, 1); % Hue image
% green channel corresponding to 100~150 Hue (0.2778~0.4167)
%hsv = rgb2hsv(filtered);
%whole_h(1,k-samplingRate) = mean(mean(hsv(:,:,1)));
%% fusing the image and vidoutput
filtered = filtered+frame;
frame = ntsc2rgb(filtered);
frame(frame > 1) = 1;
frame(frame < 0) = 0;
writeVideo(vidOut,im2uint8(frame));
end
disp('Finished')
close(vidOut); % close the video to finish writing
%% Normalization
forehead_r = Normalization(forehead_r);
forehead_g = Normalization(forehead_g);
%forehead_b = Normalization(forehead_b);
rcheek_r = Normalization(rcheek_r);
rcheek_g = Normalization(rcheek_g);
%rcheek_b = Normalization(rcheek_b);
lcheek_r = Normalization(lcheek_r);
lcheek_g = Normalization(lcheek_g);
%lcheek_b = Normalization(lcheek_b);
whole_r = Normalization(whole_r);
whole_g = Normalization(whole_g);
%% plot the RGB signal
figure
plot(forehead_r,'r');hold on
plot(forehead_g,'g');
xlabel('Number of frame','FontWeight','bold','FontSize',16);
ylabel('channel value (Normalization)','FontWeight','bold','FontSize',16);title('Forehead');
h = legend('red channel','green channel');
set(h,'Fontsize',16);
saveas(gca,[outDir '/' filename ' ' num2str(fl) '-to-' num2str(fh) '_forehead_RG.png']);
hold off
figure
plot(rcheek_r,'r');hold on
plot(rcheek_g,'g');
xlabel('Number of frame','FontWeight','bold','FontSize',16);
ylabel('channel value (Normalization)','FontWeight','bold','FontSize',16);title('Right cheek');
h = legend('red channel','green channel');
set(h,'Fontsize',16);
saveas(gca,[outDir '/' filename ' ' num2str(fl) '-to-' num2str(fh) '_rcheek_RG.png']);
hold off
figure
plot(lcheek_r,'r');hold on
plot(lcheek_g,'g');
xlabel('Number of frame','FontWeight','bold','FontSize',16);
ylabel('channel value (Normalization)','FontWeight','bold','FontSize',16);title('Left cheek');
h = legend('red channel','green channel');
set(h,'Fontsize',16);
saveas(gca,[outDir '/' filename ' ' num2str(fl) '-to-' num2str(fh) '_lcheek_RG.png']);
hold off
figure
plot(whole_r,'r');hold on
plot(whole_g,'g');
xlabel('Number of frame','FontWeight','bold','FontSize',16);
ylabel('channel value (Normalization)','FontWeight','bold','FontSize',16);title('Combined estimation');
h = legend('red channel','green channel');
set(h,'Fontsize',16);
saveas(gca,[outDir '/' filename ' ' num2str(fl) '-to-' num2str(fh) '_Combined_RG.png']);
hold off
% figure
% plot(whole_h,'r');
% xlabel('Number of frame','FontWeight','bold','FontSize',16);title('Whole image');
% legend('Hue');
% saveas(gca,[outDir '/' filename ' ' num2str(fl) '-to-' num2str(fh) '_HSV.jpg']);
%% FFT analysis
FFT_series(forehead_r,outDir,filename,'forehead_r');
FFT_series(forehead_g,outDir,filename,'forehead_g');
FFT_series(rcheek_r,outDir,filename,'rcheek_r');
FFT_series(rcheek_g,outDir,filename,'rcheek_g');
FFT_series(lcheek_r,outDir,filename,'lcheek_r');
FFT_series(lcheek_g,outDir,filename,'lcheek_g');
FFT_series(whole_r,outDir,filename,'whole_r');
FFT_series(whole_g,outDir,filename,'whole_g');
%% Do the peak detection examination
% [pks,locs] = findpeaks(rcheek_g,'MinPeakDistance',samplingRate/2);
% cycles = diff(locs);
% meanCycle = mean(cycles)
%
% Fs = 1;
% Nf = 512;
% df = Fs/Nf;
% f = 0:df:Fs/2-df;
%
% trSpots = fftshift(fft(rcheek_g-mean(rcheek_g),Nf));
%
% dBspots = 20*log10(abs(trSpots(Nf/2+1:Nf)));
%
% figure
% yaxis = [10 50];hold on
% plot(f,dBspots,1./[meanCycle meanCycle],yaxis)
% xlabel('Frequency (frame^{-1})')
% ylabel('| FFT | (dB)')
% axis([0 1/2 yaxis])
% text(1/meanCycle + .002,25,['<== 1/' num2str(meanCycle)])
% saveas(gca,[outDir '/' filename '_rcheek_g.jpg']);
% hold off
%% Finding the BPM
[RR_forehead_r, RR_forehead_g] = compareRGB(forehead_r,forehead_g);
[RR_rcheek_r, RR_rcheek_g] = compareRGB(rcheek_r,rcheek_g);
[RR_lcheek_r, RR_lcheek_g] = compareRGB(lcheek_r,lcheek_g);
[RR_whole_r, RR_whole_g] = compareRGB(whole_r,whole_g);
T_head = Time_domain(forehead_r,forehead_g);
T_rc = Time_domain(rcheek_r,rcheek_g);
T_lc = Time_domain(lcheek_r,lcheek_g);
T_wh = Time_domain(whole_r,whole_g);
bpm(1,1) = samplingRate/mean(RR_forehead_r)*60;bpm(1,2:3) = T_head(1,:); % red forehead
bpm(2,1) = samplingRate/mean(RR_rcheek_r)*60;bpm(2,2:3) = T_rc(1,:); % red right cheek
bpm(3,1) = samplingRate/mean(RR_lcheek_r)*60;bpm(3,2:3) = T_lc(1,:); % red left cheek
bpm(4,1) = samplingRate/mean(RR_whole_r)*60;bpm(4,2:3) = T_wh(1,:); % red - combined region
bpm(5,1) = samplingRate/mean(RR_forehead_g)*60;bpm(5,2:3) = T_head(2,:); % green forehead
bpm(6,1) = samplingRate/mean(RR_rcheek_g)*60;bpm(6,2:3) = T_rc(2,:); % green right cheek
bpm(7,1) = samplingRate/mean(RR_lcheek_g)*60;bpm(7,2:3) = T_lc(2,:); % green left cheek
bpm(8,1) = samplingRate/mean(RR_whole_g)*60;bpm(8,2:3) = T_wh(2,:); % green - combined region
% return RR interval of red channel with lowest standard deviation in all region
[value,index] = min([std(RR_forehead_r) std(RR_rcheek_r) std(RR_lcheek_r) std(RR_whole_r)]);
if index == 1
%RR_interval = RR_forehead_r;
signal = forehead_r;
elseif index == 2
%RR_interval = RR_rcheek_r;
signal = rcheek_r;
elseif index == 3
%RR_interval = RR_lcheek_r;
signal = lcheek_r;
else
%RR_interval = RR_whole_r;
signal = whole_r;
end
%% Cubic spline interploation
x = 1:4:length(signal)*4;
xx = 1:length(signal)*4;
yy = spline(x,signal,xx);
figure
plot(x,signal,'o',xx,yy)
saveas(gca,[outDir '/' 'spline interpolation.png']);
samplingRate = 120;
RR_interval = captureinterval(yy);
RR_interval = IntervalConstraint(RR_interval);
RR_interval(RR_interval == 0) = [];
RR_interval(1) = [];
end