forked from cvjena/knfst
-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_oneClassNovelty_knfst_artificialClass.m
36 lines (32 loc) · 1.44 KB
/
test_oneClassNovelty_knfst_artificialClass.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
% Test method for one-class classification with KNFST according to the work:
%
% Paul Bodesheim and Alexander Freytag and Erik Rodner and Michael Kemmler and Joachim Denzler:
% "Kernel Null Space Methods for Novelty Detection".
% Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2013.
%
% Please cite that paper if you are using this code!
%
%
% function scores = test_oneClassNovelty_knfst_artificialClass(model, Ks_artificial)
%
% compute novelty scores using the one-class KNFST model obtained from learn_oneClassNovelty_knfst_artificialClass
%
% INPUT:
% model -- model obtained from learn_oneClassNovelty_knfst_artificialClass
% Ks_artificial -- (2n x m) kernel matrix containing similarities of m test samples to n training samples X in the first n rows and
% the similarities to the negative replicates -X in the last n rows
%
% OUTPUT:
% scores -- novelty scores for the test samples (distances in the null space)
%
%
% (LGPL) copyright by Paul Bodesheim and Alexander Freytag and Erik Rodner and Michael Kemmler and Joachim Denzler
%
function scores = test_oneClassNovelty_knfst_artificialClass(model, Ks_artificial)
% projected test samples:
projectionVectors = Ks_artificial'*model.proj;
% differences to the target value:
diff = projectionVectors-ones(size(Ks_artificial,2),1)*model.targetValue;
% distances to the target value:
scores = sqrt(sum(diff.*diff,2));
end