-
Notifications
You must be signed in to change notification settings - Fork 28
/
Copy pathtest_MFCC.m
170 lines (160 loc) · 5.05 KB
/
test_MFCC.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
% This script test under 10-fold Cross Validation, divided by ID, how dtfv works
% Using MFCC feature
function prob = test_MFCC(part, method)
run('../../../toolbox/vlfeat-0.9.20/toolbox/vl_setup')
addpath('../../../toolbox/libsvm/matlab');
%part = 1;
trainfile = ['../Scripts_by_ID/trainVideo',num2str(part),'.txt'];
testfile = ['../Scripts_by_ID/testVideo',num2str(part),'.txt'];
feapath = '../dataset_trial/MFCCfeats.mat';
load(feapath); %load into mapMFCC, with each video name as key
%Build dictionary
if ~exist('MFCC_words.mat', 'file')
traindic = ['../Scripts_by_ID/trainVideo0.txt'];
testdic = ['../Scripts_by_ID/testVideo0.txt'];
fid = fopen(traindic);
Ctr = textscan(fid, '%s');
fclose(fid);
fid = fopen(testdic);
Cte = textscan(fid, '%s');
fclose(fid);
data = [];
num_tr = length(Ctr{1})/2;
for i = 1:num_tr
[pathstr,name,ext] = fileparts(Ctr{1}{2*i-1});
fea = mapMFCC([name,ext]); %dim-by-length feature
fea = fea(:, sum(isnan(fea),1)==0);
data = [data, fea];
end
num_te = length(Cte{1})/2;
for i = 1:num_te
[pathstr,name,ext] = fileparts(Cte{1}{2*i-1});
fea = mapMFCC([name,ext]); %dim-by-length feature
fea = fea(:, sum(isnan(fea),1)==0);
data = [data, fea];
end
fprintf('Total word number is %d.\n', size(data,2));
save('MFCC_words.mat', 'data');
else
load('MFCC_words.mat');
fprintf('Load words finished.\n');
end
% Dictionary
numClusters = 64;
if ~exist('MFCC_dict.mat', 'file')
fprintf('Start building dictionary.\n');
[means, covariances, priors] = vl_gmm(data, numClusters);
fprintf('Build dictionary finished.\n');
save('MFCC_dict.mat','means', 'covariances', 'priors');
else
load('MFCC_dict.mat');
end
% Load data
fid = fopen(trainfile);
C = textscan(fid, '%s');
fclose(fid);
fea_dim = 2*size(data,1)*numClusters;%+24576+27648;
num_v = length(C{1})/2;
train_fea = zeros(num_v, fea_dim);
train_lab = zeros(num_v, 1);
for i = 1:num_v
[pathstr,name,ext] = fileparts(C{1}{2*i-1});
s = findstr(name, 'lie');
lab = isempty(s);
tmpdata = mapMFCC([name,ext]); %dim-by-length feature
tmpdata = tmpdata(:, sum(isnan(tmpdata),1)==0);
encoding = vl_fisher(tmpdata, means, covariances, priors);
train_fea(i,:) = encoding';
train_lab(i) = lab;
end
%% test phase
fid = fopen(testfile);
C = textscan(fid, '%s');
fclose(fid);
num_v = length(C{1})/2;
test_fea = zeros(num_v, fea_dim);
test_lab = zeros(num_v, 1);
for i = 1:num_v
[pathstr,name,ext] = fileparts(C{1}{2*i-1});
s = findstr(name, 'lie');
lab = isempty(s);
tmpdata = mapMFCC([name,ext]); %dim-by-length feature
tmpdata = tmpdata(:, sum(isnan(tmpdata),1)==0);
encoding = vl_fisher(tmpdata, means, covariances, priors);
test_fea(i,:) = encoding';
test_lab(i) = lab;
end
switch(method)
case 'NN'
net = feedforwardnet(10);
net.trainFcn = 'trainscg';
net = configure(net, train_fea', train_lab');
net = train(net, train_fea', train_lab');
prob = net(test_fea');
case 'tree'
tc = fitctree(train_fea, train_lab);
[label,score,node,cnum] = predict(tc, test_fea);
prob = score(:,1);
case 'randforest'
BaggedEnsemble = TreeBagger(50,train_fea,train_lab,'OOBPred','On');
[label,scores] = predict(BaggedEnsemble, test_fea);
prob = scores(:,1);
case 'bayes'
flag = bitand(var(train_fea(train_lab==1,:))>1e-10,var(train_fea(train_lab==0,:))>1e-10); %clear 0 variance features
O1 = fitNaiveBayes(train_fea(:,flag), train_lab);
C1 = posterior(O1, test_fea(:,flag));
prob = C1(:,1);
case 'log'
B = glmfit(train_fea, [train_lab ones(size(train_lab,1),1)], 'binomial', 'link', 'logit');
Z = repmat(B(1), size(test_lab,1),1) + test_fea*B(2:end);
prob = 1 ./ (1 + exp(-Z));
prob = 1-prob;
case 'boost'
ens = fitensemble(train_fea,train_lab,'AdaBoostM1',100,'Tree')
[~, prob] = predict(ens,test_fea)
prob = prob(:,1);
case 'linearsvm'
%model = svmtrain(train_lab, tmptrain_fea, '-t 0 -q -b 1');
%fprintf('Finished training.\n');
%[pred, acc, prob] = svmpredict(test_lab, tmptest_fea, model, '-q -b 1');
%prob = prob(:,2);
model = svmtrain(train_lab, train_fea, '-t 0 -q');
[pred, acc, prob] = svmpredict(test_lab, test_fea, model, '-q');
lie_id = find(prob<0);
if ~isempty(lie_id)
if pred(lie_id(1)) == 0
isign = -1;
else
isign = 1;
end
else
if pred(1) == 1
isign = -1;
else
isign = 1;
end
end
prob = isign*prob;
case 'kernelsvm'
%model = svmtrain(train_lab, tmptrain_fea, '-t 0 -q -b 1');
%fprintf('Finished training.\n');
%[pred, acc, prob] = svmpredict(test_lab, tmptest_fea, model, '-q -b 1');
%prob = prob(:,2);
model = svmtrain(train_lab, train_fea, '-t 1 -c 1 -g 1 -q');
[pred, acc, prob] = svmpredict(test_lab, test_fea, model, '-q');
lie_id = find(prob<0);
if ~isempty(lie_id)
if pred(lie_id(1)) == 0
isign = -1;
else
isign = 1;
end
else
if pred(1) == 1
isign = -1;
else
isign = 1;
end
end
prob = isign*prob;
end