Mercurial > hg > camir-aes2014

annotate toolboxes/MIRtoolbox1.3.2/MIRToolbox/@miremotion/miremotion.m @ 0:e9a9cd732c1e tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
first hg version after svn
author wolffd
date Tue, 10 Feb 2015 15:05:51 +0000
parents
children
rev   line source
wolffd@0 1 function varargout = miremotion(orig,varargin)
wolffd@0 2 % Predicts emotion along three dimensions and five basic concepts.
wolffd@0 3 % Optional parameters:
wolffd@0 4 % miremotion(...,'Dimensions',0) excludes all three dimensions.
wolffd@0 5 % miremotion(...,'Dimensions',3) includes all three dimensions (default).
wolffd@0 6 % miremotion(...,'Activity') includes the 'Activity' dimension.
wolffd@0 7 % miremotion(...,'Valence') includes the 'Valence' dimension.
wolffd@0 8 % miremotion(...,'Tension') includes the 'Tension' dimension.
wolffd@0 9 % miremotion(...,'Dimensions',2) includes 'Activity' and 'Valence'.
wolffd@0 10 % miremotion(...,'Arousal') includes 'Activity' and 'Tension'.
wolffd@0 11 % miremotion(...,'Concepts',0) excludes all five concepts.
wolffd@0 12 % miremotion(...,'Concepts') includes all five concepts (default).
wolffd@0 13 % miremotion(...,'Happy') includes the 'Happy' concept.
wolffd@0 14 % miremotion(...,'Sad') includes the 'Sad' concept.
wolffd@0 15 % miremotion(...,'Tender') includes the 'Tender' concept.
wolffd@0 16 % miremotion(...,'Anger') includes the 'Anger' concept.
wolffd@0 17 % miremotion(...,'Fear') includes the 'Fear' concept.
wolffd@0 18 % miremotion(...,'Frame',...) predict emotion frame by frame.
wolffd@0 19 %
wolffd@0 20 % Selection of features and coefficients are taken from a study:
wolffd@0 21 % Eerola, T., Lartillot, O., and Toiviainen, P.
wolffd@0 22 % (2009). Prediction of multidimensional emotional ratings in
wolffd@0 23 % music from audio using multivariate regression models.
wolffd@0 24 % In Proceedings of 10th International Conference on Music Information Retrieval
wolffd@0 25 % (ISMIR 2009), pages 621-626.
wolffd@0 26 %
wolffd@0 27 % The implemented models are based on multiple linear regression with 5 best
wolffd@0 28 % predictors (MLR option in the paper). The box-cox transformations have now been
wolffd@0 29 % removed until the normalization values have been established with a large sample of music.
wolffd@0 30 %
wolffd@0 31 % TODO: Revision of coefficients to (a) force the output range between 0 - 1 and
wolffd@0 32 % (b) to be based on alternative models and materials (training sets).
wolffd@0 33 %
wolffd@0 34 % Updated 03.05.2010 TE
wolffd@0 35 %
wolffd@0 36 frame.key = 'Frame';
wolffd@0 37 frame.type = 'Integer';
wolffd@0 38 frame.number = 2;
wolffd@0 39 frame.default = [0 0];
wolffd@0 40 frame.keydefault = [1 1];
wolffd@0 41 option.frame = frame;
wolffd@0 42
wolffd@0 43 dim.key = 'Dimensions';
wolffd@0 44 dim.type = 'Integer';
wolffd@0 45 dim.default = NaN;
wolffd@0 46 dim.keydefault = 3;
wolffd@0 47 option.dim = dim;
wolffd@0 48
wolffd@0 49 activity.key = 'Activity';
wolffd@0 50 activity.type = 'Boolean';
wolffd@0 51 activity.default = NaN;
wolffd@0 52 option.activity = activity;
wolffd@0 53
wolffd@0 54 valence.key = 'Valence';
wolffd@0 55 valence.type = 'Boolean';
wolffd@0 56 valence.default = NaN;
wolffd@0 57 option.valence = valence;
wolffd@0 58
wolffd@0 59 tension.key = 'Tension';
wolffd@0 60 tension.type = 'Boolean';
wolffd@0 61 tension.default = NaN;
wolffd@0 62 option.tension = tension;
wolffd@0 63
wolffd@0 64 arousal.key = 'Arousal';
wolffd@0 65 arousal.type = 'Boolean';
wolffd@0 66 arousal.default = NaN;
wolffd@0 67 option.arousal = arousal;
wolffd@0 68
wolffd@0 69 concepts.key = 'Concepts';
wolffd@0 70 concepts.type = 'Boolean';
wolffd@0 71 concepts.default = NaN;
wolffd@0 72 option.concepts = concepts;
wolffd@0 73
wolffd@0 74 happy.key = 'Happy';
wolffd@0 75 happy.type = 'Boolean';
wolffd@0 76 happy.default = NaN;
wolffd@0 77 option.happy = happy;
wolffd@0 78
wolffd@0 79 sad.key = 'Sad';
wolffd@0 80 sad.type = 'Boolean';
wolffd@0 81 sad.default = NaN;
wolffd@0 82 option.sad = sad;
wolffd@0 83
wolffd@0 84 tender.key = 'Tender';
wolffd@0 85 tender.type = 'Boolean';
wolffd@0 86 tender.default = NaN;
wolffd@0 87 option.tender = tender;
wolffd@0 88
wolffd@0 89 anger.key = 'Anger';
wolffd@0 90 anger.type = 'Boolean';
wolffd@0 91 anger.default = NaN;
wolffd@0 92 option.anger = anger;
wolffd@0 93
wolffd@0 94 fear.key = 'Fear';
wolffd@0 95 fear.type = 'Boolean';
wolffd@0 96 fear.default = NaN;
wolffd@0 97 option.fear = fear;
wolffd@0 98
wolffd@0 99 specif.option = option;
wolffd@0 100 specif.defaultframelength = 1;
wolffd@0 101 %specif.defaultframehop = .5;
wolffd@0 102
wolffd@0 103 specif.combinechunk = {'Average',@nothing};
wolffd@0 104 specif.extensive = 1;
wolffd@0 105
wolffd@0 106 varargout = mirfunction(@miremotion,orig,varargin,nargout,specif,@init,@main);
wolffd@0 107
wolffd@0 108
wolffd@0 109 %%
wolffd@0 110 function [x type] = init(x,option)
wolffd@0 111
wolffd@0 112 option = process(option);
wolffd@0 113
wolffd@0 114 if option.frame.length.val
wolffd@0 115 hop = option.frame.hop.val;
wolffd@0 116 if strcmpi(option.frame.hop.unit,'Hz')
wolffd@0 117 hop = 1/hop;
wolffd@0 118 option.frame.hop.unit = 's';
wolffd@0 119 end
wolffd@0 120 if strcmpi(option.frame.hop.unit,'s')
wolffd@0 121 hop = hop*get(x,'Sampling');
wolffd@0 122 end
wolffd@0 123 if strcmpi(option.frame.hop.unit,'%')
wolffd@0 124 hop = hop/100;
wolffd@0 125 option.frame.hop.unit = '/1';
wolffd@0 126 end
wolffd@0 127 if strcmpi(option.frame.hop.unit,'/1')
wolffd@0 128 hop = hop*option.frame.length.val;
wolffd@0 129 end
wolffd@0 130 frames = 0:hop:1000000;
wolffd@0 131 x = mirsegment(x,[frames;frames+option.frame.length.val]);
wolffd@0 132 elseif isa(x,'mirdesign')
wolffd@0 133 x = set(x,'NoChunk',1);
wolffd@0 134 end
wolffd@0 135 rm = mirrms(x,'Frame',.046,.5);
wolffd@0 136
wolffd@0 137 le = 0; %mirlowenergy(rm,'ASR');
wolffd@0 138
wolffd@0 139 o = mironsets(x,'Filterbank',15,'Contrast',0.1);
wolffd@0 140 at = mirattacktime(o);
wolffd@0 141 as = 0; %mirattackslope(o);
wolffd@0 142 ed = 0; %mireventdensity(o,'Option1');
wolffd@0 143
wolffd@0 144 fl = mirfluctuation(x,'Summary');
wolffd@0 145 fp = mirpeaks(fl,'Total',1);
wolffd@0 146 fc = 0; %mircentroid(fl);
wolffd@0 147
wolffd@0 148 tp = 0; %mirtempo(x,'Frame',2,.5,'Autocor','Spectrum');
wolffd@0 149 pc = mirpulseclarity(x,'Frame',2,.5); %%%%%%%%%%% Why 'Frame'??
wolffd@0 150
wolffd@0 151 s = mirspectrum(x,'Frame',.046,.5);
wolffd@0 152 sc = mircentroid(s);
wolffd@0 153 ss = mirspread(s);
wolffd@0 154 sr = mirroughness(s);
wolffd@0 155
wolffd@0 156 %ps = mirpitch(x,'Frame',.046,.5,'Tolonen');
wolffd@0 157
wolffd@0 158 c = mirchromagram(x,'Frame','Wrap',0,'Pitch',0); %%%%%%%%%%%%%%%%%%%% Previous frame size was too small.
wolffd@0 159 cp = mirpeaks(c,'Total',1);
wolffd@0 160 ps = 0;%cp;
wolffd@0 161 ks = mirkeystrength(c);
wolffd@0 162 [k kc] = mirkey(ks);
wolffd@0 163 mo = mirmode(ks);
wolffd@0 164 hc = mirhcdf(c);
wolffd@0 165
wolffd@0 166 se = mirentropy(mirspectrum(x,'Collapsed','Min',40,'Smooth',70,'Frame',1.5,.5)); %%%%%%%%% Why 'Frame'??
wolffd@0 167
wolffd@0 168 ns = mirnovelty(mirspectrum(x,'Frame',.1,.5,'Max',5000),'Normal',0);
wolffd@0 169 nt = mirnovelty(mirchromagram(x,'Frame',.2,.25),'Normal',0); %%%%%%%%%%%%%%%%%%%% Previous frame size was too small.
wolffd@0 170 nr = mirnovelty(mirchromagram(x,'Frame',.2,.25,'Wrap',0),'Normal',0); %%%%%%%%%%%%%%%%%%%% Previous frame size was too small.
wolffd@0 171
wolffd@0 172
wolffd@0 173
wolffd@0 174 x = {rm,le, at,as,ed, fp,fc, tp,pc, sc,ss,sr, ps, cp,kc,mo,hc, se, ns,nt,nr};
wolffd@0 175
wolffd@0 176 type = {'miremotion','mirscalar','mirscalar',...
wolffd@0 177 'mirscalar','mirscalar','mirscalar',...
wolffd@0 178 'mirspectrum','mirscalar',...
wolffd@0 179 'mirscalar','mirscalar',...
wolffd@0 180 'mirscalar','mirscalar','mirscalar',...
wolffd@0 181 'mirscalar',...
wolffd@0 182 'mirchromagram','mirscalar','mirscalar','mirscalar',...
wolffd@0 183 'mirscalar',...
wolffd@0 184 'mirscalar','mirscalar','mirscalar'};
wolffd@0 185
wolffd@0 186
wolffd@0 187 %%
wolffd@0 188 function e = main(x,option,postoption)
wolffd@0 189
wolffd@0 190 warning('WARNING IN MIRENOTION: The current model of miremotion is not correctly calibrated with this version of MIRtoolbox (but with version 1.3 only).');
wolffd@0 191
wolffd@0 192 option = process(option);
wolffd@0 193 rm = get(x{1},'Data');
wolffd@0 194 %le = get(x{2},'Data');
wolffd@0 195 at = get(x{3},'Data');
wolffd@0 196 %as = get(x{4},'Data');
wolffd@0 197 %ed = get(x{5},'Data');
wolffd@0 198 %fpp = get(x{6},'PeakPosUnit');
wolffd@0 199 fpv = get(x{6},'PeakVal');
wolffd@0 200 %fc = get(x{7},'Data');
wolffd@0 201 %tp = get(x{8},'Data');
wolffd@0 202 pc = get(x{9},'Data');
wolffd@0 203 sc = get(x{10},'Data');
wolffd@0 204 ss = get(x{11},'Data');
wolffd@0 205 rg = get(x{12},'Data');
wolffd@0 206 %ps = get(x{13},'PeakPosUnit');
wolffd@0 207 cp = get(x{14},'PeakPosUnit');
wolffd@0 208 kc = get(x{15},'Data');
wolffd@0 209 mo = get(x{16},'Data');
wolffd@0 210 hc = get(x{17},'Data');
wolffd@0 211 se = get(x{18},'Data');
wolffd@0 212 ns = get(x{19},'Data');
wolffd@0 213 nt = get(x{20},'Data');
wolffd@0 214 nr = get(x{21},'Data');
wolffd@0 215
wolffd@0 216
wolffd@0 217 e.dim = {};
wolffd@0 218 e.dimdata = mircompute(@initialise,rm);
wolffd@0 219 if option.activity == 1
wolffd@0 220 [e.dimdata e.activity_fact] = mircompute(@activity,e.dimdata,rm,fpv,sc,ss,se);
wolffd@0 221 e.dim = [e.dim,'Activity'];
wolffd@0 222 else
wolffd@0 223 e.activity_fact = NaN;
wolffd@0 224 end
wolffd@0 225 if option.valence == 1
wolffd@0 226 [e.dimdata e.valence_fact] = mircompute(@valence,e.dimdata,rm,fpv,kc,mo,ns);
wolffd@0 227 e.dim = [e.dim,'Valence'];
wolffd@0 228 else
wolffd@0 229 e.valence_fact = NaN;
wolffd@0 230 end
wolffd@0 231 if option.tension == 1
wolffd@0 232 [e.dimdata e.tension_fact] = mircompute(@tension,e.dimdata,rm,fpv,kc,hc,nr);
wolffd@0 233 e.dim = [e.dim,'Tension'];
wolffd@0 234 else
wolffd@0 235 e.tension_fact = NaN;
wolffd@0 236 end
wolffd@0 237
wolffd@0 238 e.class = {};
wolffd@0 239 e.classdata = mircompute(@initialise,rm);
wolffd@0 240 if option.happy == 1
wolffd@0 241 [e.classdata e.happy_fact] = mircompute(@happy,e.classdata,fpv,ss,cp,kc,mo);
wolffd@0 242 e.class = [e.class,'Happy'];
wolffd@0 243 else
wolffd@0 244 e.happy_fact = NaN;
wolffd@0 245 end
wolffd@0 246 if option.sad == 1
wolffd@0 247 [e.classdata e.sad_fact] = mircompute(@sad,e.classdata,ss,cp,mo,hc,nt);
wolffd@0 248 e.class = [e.class,'Sad'];
wolffd@0 249 else
wolffd@0 250 e.sad_fact = NaN;
wolffd@0 251 end
wolffd@0 252 if option.tender == 1
wolffd@0 253 [e.classdata e.tender_fact] = mircompute(@tender,e.classdata,sc,rg,kc,hc,ns);
wolffd@0 254 e.class = [e.class,'Tender'];
wolffd@0 255 else
wolffd@0 256 e.tender_fact = NaN;
wolffd@0 257 end
wolffd@0 258 if option.anger == 1
wolffd@0 259 [e.classdata e.anger_fact] = mircompute(@anger,e.classdata,rg,kc,se,nr);
wolffd@0 260 e.class = [e.class,'Anger'];
wolffd@0 261 else
wolffd@0 262 e.anger_fact = NaN;
wolffd@0 263 end
wolffd@0 264 if option.fear == 1
wolffd@0 265 [e.classdata e.fear_fact] = mircompute(@fear,e.classdata,rm,at,fpv,kc,mo);
wolffd@0 266 e.class = [e.class,'Fear'];
wolffd@0 267 else
wolffd@0 268 e.fear_fact = NaN;
wolffd@0 269 end
wolffd@0 270
wolffd@0 271 e = class(e,'miremotion',mirdata(x{1}));
wolffd@0 272 e = purgedata(e);
wolffd@0 273 fp = mircompute(@noframe,get(x{1},'FramePos'));
wolffd@0 274 e = set(e,'Title','Emotion','Abs','emotions','Ord','magnitude','FramePos',fp);
wolffd@0 275
wolffd@0 276 %%
wolffd@0 277 function option = process(option)
wolffd@0 278 if option.arousal==1
wolffd@0 279 option.activity = 1;
wolffd@0 280 option.tension = 1;
wolffd@0 281 if isnan(option.dim)
wolffd@0 282 option.dim = 0;
wolffd@0 283 end
wolffd@0 284 end
wolffd@0 285 if option.activity==1 || option.valence==1 || option.tension==1
wolffd@0 286 if isnan(option.activity)
wolffd@0 287 option.activity = 0;
wolffd@0 288 end
wolffd@0 289 if isnan(option.valence)
wolffd@0 290 option.valence = 0;
wolffd@0 291 end
wolffd@0 292 if isnan(option.tension)
wolffd@0 293 option.tension = 0;
wolffd@0 294 end
wolffd@0 295 if isnan(option.concepts)
wolffd@0 296 option.concepts = 0;
wolffd@0 297 end
wolffd@0 298 end
wolffd@0 299 if not(isnan(option.dim)) && option.dim
wolffd@0 300 if isnan(option.concepts)
wolffd@0 301 option.concepts = 0;
wolffd@0 302 end
wolffd@0 303 end
wolffd@0 304 if not(isnan(option.concepts)) && option.concepts
wolffd@0 305 if isnan(option.dim)
wolffd@0 306 option.dim = 0;
wolffd@0 307 end
wolffd@0 308 end
wolffd@0 309 if not(isnan(option.dim))
wolffd@0 310 switch option.dim
wolffd@0 311 case 0
wolffd@0 312 if isnan(option.activity)
wolffd@0 313 option.activity = 0;
wolffd@0 314 end
wolffd@0 315 if isnan(option.valence)
wolffd@0 316 option.valence = 0;
wolffd@0 317 end
wolffd@0 318 if isnan(option.tension)
wolffd@0 319 option.tension = 0;
wolffd@0 320 end
wolffd@0 321 case 2
wolffd@0 322 option.activity = 1;
wolffd@0 323 option.valence = 1;
wolffd@0 324 if isnan(option.tension)
wolffd@0 325 option.tension = 0;
wolffd@0 326 end
wolffd@0 327 case 3
wolffd@0 328 option.activity = 1;
wolffd@0 329 option.valence = 1;
wolffd@0 330 option.tension = 1;
wolffd@0 331 end
wolffd@0 332 end
wolffd@0 333 if isnan(option.activity)
wolffd@0 334 option.activity = 1;
wolffd@0 335 end
wolffd@0 336 if isnan(option.valence)
wolffd@0 337 option.valence = 1;
wolffd@0 338 end
wolffd@0 339 if isnan(option.tension)
wolffd@0 340 option.tension = 1;
wolffd@0 341 end
wolffd@0 342 if isnan(option.concepts)
wolffd@0 343 option.concepts = 1;
wolffd@0 344 end
wolffd@0 345 if option.concepts
wolffd@0 346 option.happy = 1;
wolffd@0 347 option.sad = 1;
wolffd@0 348 option.tender = 1;
wolffd@0 349 option.anger = 1;
wolffd@0 350 option.fear = 1;
wolffd@0 351 end
wolffd@0 352 if option.happy==1 || option.sad==1 || option.tender==1 ...
wolffd@0 353 || option.anger==1 || option.fear==1
wolffd@0 354 if isnan(option.happy)
wolffd@0 355 option.happy = 0;
wolffd@0 356 end
wolffd@0 357 if isnan(option.sad)
wolffd@0 358 option.sad = 0;
wolffd@0 359 end
wolffd@0 360 if isnan(option.tender)
wolffd@0 361 option.tender = 0;
wolffd@0 362 end
wolffd@0 363 if isnan(option.anger)
wolffd@0 364 option.anger = 0;
wolffd@0 365 end
wolffd@0 366 if isnan(option.fear)
wolffd@0 367 option.fear = 0;
wolffd@0 368 end
wolffd@0 369 end
wolffd@0 370
wolffd@0 371
wolffd@0 372 %%
wolffd@0 373 function e = initialise(rm)
wolffd@0 374 e = [];
wolffd@0 375
wolffd@0 376
wolffd@0 377 function e = activity(e,rm,fpv,sc,ss,se) % without the box-cox transformation, revised coefficients
wolffd@0 378 af = zeros(5,1);
wolffd@0 379 af(1) = 0.6664* ((mean(rm) - 0.0559)/0.0337); %
wolffd@0 380 af(2) = 0.6099 * ((mean(fpv{1}) - 13270.1836)/10790.655);
wolffd@0 381 af(3) = 0.4486*((mean(cell2mat(sc)) - 1677.7)./570.34);
wolffd@0 382 af(4) = -0.4639*((mean(cell2mat(ss)) - 250.5574)./205.3147);
wolffd@0 383 af(5) = 0.7056*((mean(se) - 0.954)./0.0258);
wolffd@0 384 af(isnan(af)) = [];
wolffd@0 385 e(end+1,:) = sum(af)+5.4861;
wolffd@0 386 e = {e af};
wolffd@0 387
wolffd@0 388 function e = valence(e,rm,fpv,kc,mo,ns) % without the box-cox transformation, revised coefficients
wolffd@0 389 vf = zeros(5,1);
wolffd@0 390 vf(1) = -0.3161 * ((std(rm) - 0.024254)./0.015667);
wolffd@0 391 vf(2) = 0.6099 * ((mean(fpv{1}) - 13270.1836)/10790.655);
wolffd@0 392 vf(3) = 0.8802 * ((mean(kc) - 0.5123)./0.091953);
wolffd@0 393 vf(4) = 0.4565 * ((mean(mo) - -0.0019958)./0.048664);
wolffd@0 394 ns(isnan(ns)) = [];
wolffd@0 395 vf(5) = 0.4015 * ((mean(ns) - 131.9503)./47.6463);
wolffd@0 396 vf(isnan(vf)) = [];
wolffd@0 397 e(end+1,:) = sum(vf)+5.2749;
wolffd@0 398 e = {e vf};
wolffd@0 399
wolffd@0 400 function e = tension(e,rm,fpv,kc,hc,nr)
wolffd@0 401 tf = zeros(5,1);
wolffd@0 402 tf(1) = 0.5382 * ((std(rm) - 0.024254)./0.015667);
wolffd@0 403 tf(2) = -0.5406 * ((mean(fpv{1}) - 13270.1836)/10790.655);
wolffd@0 404 tf(3) = -0.6808 * ((mean(kc) - 0.5124)./0.092);
wolffd@0 405 tf(4) = 0.8629 * ((mean(hc) - 0.2962)./0.0459);
wolffd@0 406 tf(5) = -0.5958 * ((mean(nr) - 71.8426)./46.9246);
wolffd@0 407 tf(isnan(tf)) = [];
wolffd@0 408 e(end+1,:) = sum(tf)+5.4679;
wolffd@0 409 e = {e tf};
wolffd@0 410
wolffd@0 411
wolffd@0 412 % BASIC EMOTION PREDICTORS
wolffd@0 413
wolffd@0 414 function e = happy(e,fpv,ss,cp,kc,mo)
wolffd@0 415 ha_f = zeros(5,1);
wolffd@0 416 ha_f(1) = 0.7438*((mean(cell2mat(fpv)) - 13270.1836)./10790.655);
wolffd@0 417 ha_f(2) = -0.3965*((mean(cell2mat(ss)) - 250.5574)./205.3147);
wolffd@0 418 ha_f(3) = 0.4047*((std(cell2mat(cp)) - 8.5321)./2.5899);
wolffd@0 419 ha_f(4) = 0.7780*((mean(kc) - 0.5124)./0.092);
wolffd@0 420 ha_f(5) = 0.6220*((mean(mo) - -0.002)./0.0487);
wolffd@0 421 ha_f(isnan(ha_f)) = [];
wolffd@0 422 e(end+1,:) = sum(ha_f)+2.6166;
wolffd@0 423 e = {e ha_f};
wolffd@0 424
wolffd@0 425 function e = sad(e,ss,cp,mo,hc,nt)
wolffd@0 426 sa_f = zeros(5,1);
wolffd@0 427 sa_f(1) = 0.4324*((mean(cell2mat(ss)) - 250.5574)./205.3147);
wolffd@0 428 sa_f(2) = -0.3137*((std(cell2mat(cp)) - 8.5321)./2.5899);
wolffd@0 429 sa_f(3) = -0.5201*((mean(mo) - -0.0020)./0.0487);
wolffd@0 430 sa_f(4) = -0.6017*((mean(hc) - 0.2962)./0.0459);
wolffd@0 431 sa_f(5) = 0.4493*((mean(nt) - 42.2022)./36.7782);
wolffd@0 432 sa_f(isnan(sa_f)) = [];
wolffd@0 433 e(end+1,:) = sum(sa_f)+2.9756;
wolffd@0 434 e = {e sa_f};
wolffd@0 435
wolffd@0 436 function e = tender(e,sc,rg,kc,hc,ns)
wolffd@0 437 te_f = zeros(5,1);
wolffd@0 438 te_f(1) = -0.2709*((mean(cell2mat(sc)) - 1677.7106)./570.3432);
wolffd@0 439 te_f(2) = -0.4904*((std(rg) - 85.9387)./106.0767);
wolffd@0 440 te_f(3) = 0.5192*((mean(kc) - 0.5124)./0.0920);
wolffd@0 441 te_f(4) = -0.3995*((mean(hc) - 0.2962)./0.0459);
wolffd@0 442 te_f(5) = 0.3391*((mean(ns) - 131.9503)./47.6463);
wolffd@0 443 te_f(isnan(te_f)) = [];
wolffd@0 444 e(end+1,:) = sum(te_f)+2.9756;
wolffd@0 445 e = {e te_f};
wolffd@0 446
wolffd@0 447 function e = anger(e,rg,kc,se,nr) %
wolffd@0 448 an_f = zeros(5,1);
wolffd@0 449 %an_f(1) = -0.2353*((mean(pc) - 0.1462)./.1113);
wolffd@0 450 an_f(2) = 0.5517*((mean(rg) - 85.9387)./106.0767);
wolffd@0 451 an_f(3) = -.5802*((mean(kc) - 0.5124)./0.092);
wolffd@0 452 an_f(4) = .2821*((mean(se) - 0.954)./0.0258);
wolffd@0 453 an_f(5) = -.2971*((mean(nr) - 71.8426)./46.9246);
wolffd@0 454 an_f(isnan(an_f)) = [];
wolffd@0 455 e(end+1,:) = sum(an_f)+1.9767;
wolffd@0 456 e = {e an_f};
wolffd@0 457
wolffd@0 458 function e = fear(e,rm,at,fpv,kc,mo)
wolffd@0 459 fe_f = zeros(5,1);
wolffd@0 460 fe_f(1) = 0.4069*((std(rm) - 0.0243)./0.0157);
wolffd@0 461 fe_f(2) = -0.6388*((mean(at) - 0.0707)./0.015689218536423);
wolffd@0 462 fe_f(3) = -0.2538*((mean(cell2mat(fpv)) - 13270.1836)./10790.655);
wolffd@0 463 fe_f(4) = -0.9860*((mean(kc) - 0.5124)./0.0920);
wolffd@0 464 fe_f(5) = -0.3144*((mean(mo) - -0.0019958)./0.048663550639094);
wolffd@0 465 fe_f(isnan(fe_f)) = [];
wolffd@0 466 e(end+1,:) = sum(fe_f)+2.7847;
wolffd@0 467 e = {e fe_f};
wolffd@0 468
wolffd@0 469 function fp = noframe(fp)
wolffd@0 470 fp = [fp(1);fp(end)];