Signal Processing - July 2016 - 55

a face image sensed by the front-facing camera, pretrained
attribute classifiers are used to extract a 44-dimensional attribute feature. The binary attribute classifiers are trained using
the PubFig data set [30] and provide compact visual descriptions of faces. The score is determined by comparing the
extracted attribute features with the features corresponding to
the enrolled user. These score values are essentially used to
continuously authenticate a mobile-device user. Furthermore,
it was shown that the attribute-based method can be fused
with an LBP-based method such as in [21] to obtain improved
matching performance.
Table 2 summarizes key face-based continuous authentication methods. Here, the recognition rate (RR), true accept rate
(TAR), and average authentication rate (AAR) are noted for
some of the studies.

Gait dynamics

Accelerometer
Measurements

Gait dynamics-based continuous authentication systems
identify users based on how they walk. The data needed for
gait-based authentication are often measured by the built-in
accelerometer and gyroscope sensors. Once the raw data are
measured, discriminative features are extracted, which are
then fed into a classifier to distinguish users. In recent years,
several methods have been developed for gait-based recognition on mobile devices [32]-[38]. These methods differ essentially in the types of features extracted from the raw data for
classification or the types of classification methods used for
authentication. For instance, methods based on correlation, frequency domain analysis, and data distribution statics are used
in [32], while methods based on dynamic time warping are
used in [36] and [37]. Rather than using the gait cycles for
extracting features, [35] proposes an application of hidden
Markov models (HMMs) for gait recognition. In particular,
a sensor orientation invariant gait representation called gait
dynamic images (GDIs) was proposed in [39]. Given a 3-D
time series captured by a three-axis accelerometer, its GDI is
calculated by the cosine similarity of the motion measurement
at time t with the time-lagged signal of lag l. Figure 8 shows
an example of raw three-axis accelerometer data and their

Table 2. A summary of key face-based continuous authentication methods.
Study
Abeni et al. [28]
Hadid et al. [21]
Fathy et al. [26]
Crouse et al. [12]
Samangouei
et al. [31]

Performance (%)
EER: 3.95-7.92
AAR: 82-96
RR: ~95
TAR: ~40-50
@FAR 0.1
EER: 13-30

MEEN: mouth, left eye, right eye, nose.

corresponding GDI. As can be seen from this figure, since
GDI is invariant in regard to sensor orientation, it shows much
better consistency before and after sensor rotation. Also, paceindependent gait recognition approaches have been proposed
in [34] and [38]. In [38], GDIs are used, while in [34] cyclostationarity and continuous wavelet transform spectrogram analysis are used for gait-based recognition. Table 3 summarizes
key gait dynamics-based continuous authentication methods in
terms of their performance on various data sets. In this table,
the verification rate (VR), false nonmatch rate (FNMR), and
false match rate (FMR) are noted in two of the studies.

Behavior-based profiling
Behavior profiling techniques verify the user's identity based
on the applications and services they use. The research into
mobile behavior profiling started in late 1990s, focusing
mainly on developing intrusion detection systems (IDSs) to
detect telephony service fraud by monitoring user calling and
migration behavior [40]-[42]. In these systems, user profiles
are created by monitoring user activities for a period of time
and are compared against the current activity profiles of the
user. If a significant deviation is observed, a possible intrusion
is detected.

15
10
5
0
-5
125

Number
of Users Method/Features
32
1-class SVM/Fourier
transform
12
Histogram intersection distance/LBP
50
Nine different
classifiers/MEEN
10
SVM/biologically
inspired model
50
Attributes

xx-yy-zz-130

135

140
Time (s)
(a)

145

150

155

(b)

FIGURE 8. (a) Data measurements from a three-axis accelerometer embedded in a mobile phone carried by a walking user. (b) The corresponding GDI.
IEEE Signal Processing Magazine

|

July 2016

|

55



Table of Contents for the Digital Edition of Signal Processing - July 2016

Signal Processing - July 2016 - Cover1
Signal Processing - July 2016 - Cover2
Signal Processing - July 2016 - 1
Signal Processing - July 2016 - 2
Signal Processing - July 2016 - 3
Signal Processing - July 2016 - 4
Signal Processing - July 2016 - 5
Signal Processing - July 2016 - 6
Signal Processing - July 2016 - 7
Signal Processing - July 2016 - 8
Signal Processing - July 2016 - 9
Signal Processing - July 2016 - 10
Signal Processing - July 2016 - 11
Signal Processing - July 2016 - 12
Signal Processing - July 2016 - 13
Signal Processing - July 2016 - 14
Signal Processing - July 2016 - 15
Signal Processing - July 2016 - 16
Signal Processing - July 2016 - 17
Signal Processing - July 2016 - 18
Signal Processing - July 2016 - 19
Signal Processing - July 2016 - 20
Signal Processing - July 2016 - 21
Signal Processing - July 2016 - 22
Signal Processing - July 2016 - 23
Signal Processing - July 2016 - 24
Signal Processing - July 2016 - 25
Signal Processing - July 2016 - 26
Signal Processing - July 2016 - 27
Signal Processing - July 2016 - 28
Signal Processing - July 2016 - 29
Signal Processing - July 2016 - 30
Signal Processing - July 2016 - 31
Signal Processing - July 2016 - 32
Signal Processing - July 2016 - 33
Signal Processing - July 2016 - 34
Signal Processing - July 2016 - 35
Signal Processing - July 2016 - 36
Signal Processing - July 2016 - 37
Signal Processing - July 2016 - 38
Signal Processing - July 2016 - 39
Signal Processing - July 2016 - 40
Signal Processing - July 2016 - 41
Signal Processing - July 2016 - 42
Signal Processing - July 2016 - 43
Signal Processing - July 2016 - 44
Signal Processing - July 2016 - 45
Signal Processing - July 2016 - 46
Signal Processing - July 2016 - 47
Signal Processing - July 2016 - 48
Signal Processing - July 2016 - 49
Signal Processing - July 2016 - 50
Signal Processing - July 2016 - 51
Signal Processing - July 2016 - 52
Signal Processing - July 2016 - 53
Signal Processing - July 2016 - 54
Signal Processing - July 2016 - 55
Signal Processing - July 2016 - 56
Signal Processing - July 2016 - 57
Signal Processing - July 2016 - 58
Signal Processing - July 2016 - 59
Signal Processing - July 2016 - 60
Signal Processing - July 2016 - 61
Signal Processing - July 2016 - 62
Signal Processing - July 2016 - 63
Signal Processing - July 2016 - 64
Signal Processing - July 2016 - 65
Signal Processing - July 2016 - 66
Signal Processing - July 2016 - 67
Signal Processing - July 2016 - 68
Signal Processing - July 2016 - 69
Signal Processing - July 2016 - 70
Signal Processing - July 2016 - 71
Signal Processing - July 2016 - 72
Signal Processing - July 2016 - 73
Signal Processing - July 2016 - 74
Signal Processing - July 2016 - 75
Signal Processing - July 2016 - 76
Signal Processing - July 2016 - 77
Signal Processing - July 2016 - 78
Signal Processing - July 2016 - 79
Signal Processing - July 2016 - 80
Signal Processing - July 2016 - 81
Signal Processing - July 2016 - 82
Signal Processing - July 2016 - 83
Signal Processing - July 2016 - 84
Signal Processing - July 2016 - 85
Signal Processing - July 2016 - 86
Signal Processing - July 2016 - 87
Signal Processing - July 2016 - 88
Signal Processing - July 2016 - 89
Signal Processing - July 2016 - 90
Signal Processing - July 2016 - 91
Signal Processing - July 2016 - 92
Signal Processing - July 2016 - 93
Signal Processing - July 2016 - 94
Signal Processing - July 2016 - 95
Signal Processing - July 2016 - 96
Signal Processing - July 2016 - 97
Signal Processing - July 2016 - 98
Signal Processing - July 2016 - 99
Signal Processing - July 2016 - 100
Signal Processing - July 2016 - 101
Signal Processing - July 2016 - 102
Signal Processing - July 2016 - 103
Signal Processing - July 2016 - 104
Signal Processing - July 2016 - Cover3
Signal Processing - July 2016 - Cover4
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_201809
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_201807
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_201805
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_201803
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_201801
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_1117
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0917
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0717
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0517
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0317
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0117
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_1116
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0916
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0716
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0516
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0316
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0116
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_1115
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0915
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0715
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0515
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0315
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0115
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_1114
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0914
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0714
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0514
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0314
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0114
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_1113
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0913
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0713
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0513
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0313
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0113
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_1112
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0912
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0712
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0512
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0312
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0112
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_1111
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0911
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0711
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0511
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0311
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0111
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_1110
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0910
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0710
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0510
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0310
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0110
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_1109
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0909
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0709
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0509
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0309
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0109
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_1108
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0908
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0708
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0508
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0308
https://www.nxtbook.com/nxtbooks/ieee/signalprocessing_0108
https://www.nxtbookmedia.com