-
Notifications
You must be signed in to change notification settings - Fork 0
/
pca_face_demo.m
51 lines (46 loc) · 1.68 KB
/
pca_face_demo.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
function pca_face_demo(traindir, facedir, nonfacedir)
% Demonstrates the PCA face recognition/detection algorithm.
%
% Arguments:
% traindir: a string with the path to the directory that contains the
% training images.
%
% facedir: a string with the path to the directory that contains the
% face test images.
%
% nonfacedir: a string with the path to the directory that contains the
% non-face test images.
%
% (c) Jaime Soto
% CAP 6411 - Computer Vision Systems
% University of Central Florida
% 7 December 2010
%
pca_face_db = 'pca_face_db.mat';
tic();
pca_face_train(traindir, 'face0*.pgm', [19 19], pca_face_db, 200, 200);
fprintf(1, 'training took %f sec\n', toc());
tic();
[fdifs fdffs] = pca_face_detect(facedir, 'cmu_0*.pgm', pca_face_db, 1000);
fprintf(1, 'detection on %d faces took %f sec\n', length(fdifs), toc());
tic();
[ndifs ndffs] = pca_face_detect(nonfacedir, 'cmu_0*.pgm', pca_face_db, 1000);
fprintf(1, 'detection on %d non-faces took %f sec\n', length(ndifs), toc());
threshold = 7.75;
face = fdifs < threshold;
nface = ndifs < threshold;
fprintf(1, 'True positives: %d\n', nnz(face));
fprintf(1, 'True negatives: %d\n', nnz(nface));
fprintf(1, 'False positives: %d\n', nnz(~nface));
fprintf(1, 'False negatives: %d\n', nnz(~face));
clf
hold on
title('Feature Space Distances for Faces and Non-Faces');
plot(fdifs, fdffs, 'ob');
plot(ndifs, ndffs, 'xr');
legend('Faces', 'Non-Faces');
xlabel('[log] Distance in face space (DIFS): Mahalanobis difference');
ylabel('[log] Distance from face space (DFFS): Sum of Difference of Squares');
print('facespace_distances.png', '-dpng');
hold off
end