J. Software Engineering & Applications, 2010, 3: 119-124
doi:10.4236/jsea.2010.32015 Published Online February 2010 (http://www.SciRP.org/journal/jsea)
Copyright © 2010 SciRes JSEA
119
A Codebook Design Method for Robust VQ-Based
Face Recognition Algorithm
Qiu Chen1, Koji Kotani2, Feifei Lee1, Tadahiro Ohmi1
1New Industry Creatio n Hatchery Cen ter, Tohoku University; 2Department of Electronics, Graduate School of Engineering, Tohoku Univer-
sity, Japan..
Email: qiu@fff.niche.tohoku.ac.jp
Received September 5th, 2009; revised November 2nd, 2009; accepted November 6th, 2009.
ABSTRACT
In this paper, we present a theoretical codebook design method for VQ-based fast face recognition algorithm to im-
prove recognition accuracy. Based on the systema tic analysis and classification of code patterns, firstly we th eoretically
create a systematically organized codebook. Combined with another codebook created by Kohonen’s Self-Organizing
Maps (SOM) method, an optimized codebook consisted of 2×2 codevectors for facial images is generated. Experimental
results show face recognition using such a codebook is more efficient than the codebook consisted of 4×4 codevector
used in conventional algorithm. The highest average recognition rate of 98.6% is obtained for 40 persons’ 400 images
of publicly available face database of AT&T Laboratories Cambridge containing variations in lighting, posing, and
expressions. A table look-up (TLU) method is also proposed for the speed up of the recogn ition processin g. By applying
this method in the quantization step, the total recognition processing time achieves only 28 msec, enabling real-time
face recognition.
Keywords: Face Recognition, Vector Quantization (VQ), Codebook Design, Code Classification, Histogram Method
1. Introduction
After September 11th, security systems utilizing per-
sonal biometric features, such as, face, voice, finger-
print, iris pattern, etc. are attracting a lot of attention.
Among them, face recognition have become the subject
of increased interest [1], which seems to be the most
natural and effective method to identify a person since
it is the same as the way human does and there is no
need to use special equipments. In face recognition,
personal facial feature extraction is the key to creating
more robust systems.
A lot of algorithms have been proposed for solving
face recognition problem. Based on the use of the Kar-
hunen-Loeve transform, PCA [2] is used to represent a
face in terms of an optimal coordinate system which
contains the most significant eigenfaces and the mean
square error is minimal. However, it is highly compli-
cated and computational-power hungry, making it diffi-
cult to implement them into real-time face recognition
applications. Feature-based approach [3,4] uses the rela-
tionship between facial features, such as the locations of
eye, mouth and nose. It can implement very fast, but
recognition rate usually depends on the location accuracy
of facial features, so it can not give a satisfied recogni-
tion result. There are many other algorithms have been
used for face recognition. Such as Local Feature Analysis
(LFA) [5], neural network [6], local autocorrelations and
multi-scale integration technique [7], and other tech-
niques [8–14] have been pro p o sed.
Kotani et al. [15] have proposed a novel informa-
tion-processing algorithm called Vector Quantization
(VQ) codebook space information processing which dif-
fers from the traditional ways of processing algorithm.
Based on this algorithm, we have developed a very sim-
ple yet highly reliable face recognition method called VQ
histogram method by using a systematically organized
Codebook for 4×4 blocks with 33 codevectors having
monotonic intensity variation without DC component.
VQ algorithm [16] is well known in the field of image
coding (compression). Input image is first divided into
small blocks, which are taken as input vectors in VQ
operation. Each input vector is then matched with code-
vectors in a codebook by calculating distances between
them. The codevector having the maximum similarity to
the input vector is selected by searching the minimum
distance and the index number of the selected codevector
is output.
This index number information was paid attention to.
It was found that a codevector histogram, which is ob-
tained by counting the matching frequency of individual
A Codebook Design Method for Robust VQ-Based Face Recognition Algorithm
120
codevector, contains very effective facial feature infor-
mation. By utilizing this technique, a novel face recogni-
tion algorithm called VQ histogram method has been
developed.
A codebook is very important since it directly affects
the quality of VQ processing. In [15], a special codebook
was used, which is systematically organized for 4x4
blocks with 33 codevectors having monotonic intensity
variation without DC component.
In this paper, a theoretical codebook design method is
proposed. At first, a systematically organized codebook
is created based on the distribution of code patterns ab-
stracted from facial images [21], and then another code-
book with the same size is created using Kohonen’s
Self-Organizing Maps (SOM) [20]. Combining the two
codebooks obtained above, final optimized codebook
consisted of 2×2 codevectors for facial images will be
generated [22,23]. It can represent the features of the
facial images more adequately. Furthermore, a table
look-up (TLU) method is also proposed for the speed up
of the recognition processing.
This paper is organized as follows. First, VQ histo-
gram method will be introduced in detail in Section 2.
Proposed codebook design method combining classifica-
tion of code patterns and Kohonen’s Self-Organizing
Maps (SOM) will be described in Section 3. Experimen-
tal results compared with the algorithms employing
original codebook or SOM codebook separately will be
discussed in Section 4. Finally, we make a conclusion in
Section 5.
2. Vector Quantization Histogram Method
In this section, we will describe the face recognition
algorithm using Vector Quantization (VQ) histogram
method [15]. Figure 1 shows face recognition process
steps. First, low-pass filtering is carried out using sim-
ple 2-D moving average filter. This low-pass filtering is
essential for reducing high-frequency noise and ex-
tracting most effective low frequency component for
recognition. Block segmentation step, in which facial
image is divided into small image blocks (for example,
2×2) with overlap, namely, by sliding dividing-partition
one pixel by one pixel, is the following. Next, minimum
intensity in the individual block is searched, and found
minimum intensity is subtracted from each pixel in the
block. Only the intensity variation in the block is ex-
tracted by this process. This is very effective for mini-
mizing the effect of overall brightness variations. Vec-
tor quantization is then applied to intensity-variation
blocks (vectors) by using a codebook which prepared in
advance. The most similar (matched) codevector to the
input block is selected.
After performing VQ for all blocks divided from a fa-
cial image, matched frequencies for each codevector are
counted and histogram is generated. This histogram be-
Facial image
Low Pass Filtering
Block Division
Min. Intensity Subtraction
Vector Quantization
Histogram Generation
Database Matching
Recognition result
Codebook
Database
Figure 1. Face recognition process steps
comes the feature vector of human face. In the registra-
tion procedure, this histogram is saved in a database as
personal identification information. In the recognition
procedure, the histogram made from an input facial im-
age is compared with registered individual histograms
and the best match is output as a recognition result.
Manhattan distance between histograms is utilized as a
matching measure.
Codebook which consists of typical feature patterns
for representing the features of face image is very im-
portant. In [15], 32 codevectors of 4x4 codebook ar e cre-
ated by changing the direction (8 different directions)
and the range of intensity variation (Step values are 2, 6,
10, and 20). By adding one codevector having no inten-
sity variation, complete codebook is organized.
In the next section, we will propose a novel codebook
design method for 2×2 codebook, which can represent
the features of the facial image more adequately.
3. Codebook Design
Because of the characteristic of face image, the codevec-
tors of 4x4 codebook are all low-frequency patterns. Al-
though the average face recognition rate of 95.6% has
been obtained using such a codebook, it is still difficult
to say this codebook is the most suitable because the
number of categories for the 4x4 block patterns is too
large to be classified by only 33 categories, and it has not
been proved in theory that these 33 codevectors most
adequately represent the face patterns. But in the case of
2×2 code patterns, the condition is considerably different.
Nakayama et al. [17] have developed a complete classi-
fication method for 2×2 codebook design for image
compression. By the similar consideration, we classify
and analyze the code patterns in the face images, and
then theoretically create a new codebook of 2×2 code
patterns for face recognition algorithm.
Copyright © 2010 SciRes JSEA
A Codebook Design Method for Robust VQ-Based Face Recognition Algorithm 121
3.1 Previous Work
Nakayama et al. [17] proposed a complete classification
method for 2×2 codebook design in image compression.
Figure 2 shows all categories for the 2×2 image block
patterns without considering the location of pixels. In a
2×2 block, pixel intensities are marked by alphabet ‘a’,
‘b’, ‘c’, ‘d’, and a>b>c>d is prescribed. In [17], it was
found that the number of typical patterns for all 2×2 im-
age block is only 11. The number of varieties in pixel
arrangement of each 2×2 typical pattern is also shown in
Figure 2. That means the total number of image patterns
for 2×2 pixel blocks is theoretically only 75.
By the similar consideration, we classified and ana-
lyzed the code patterns in the face images [21]. We
found that in all filter size, the numbers of code patterns
belong to categories 7, 10, and 11 are very few. It
means such code patterns are almost not used in face
images. Based on this result, we created a new code-
book for 2×2 code patterns, and the rules of codebook
creation are as follows.
1) Create very small intensity-variation (intensity dif-
ference among the block is only 1 or 2) code patterns
having monotonic intensity variation, the number of
code pattern is 16 by changing the direction of intensity
variation.
2) Create code patterns of category No.2, 3, 4, 5, 6, 8
and 9, and intensity differences among the blocks are set
to be 3, 6, and 10.
3) Do not make code patterns of category No. 7, 10,
and 11.
4) Add one code pattern having no intensity variation.
Thus, complete codebook is systematically organized
with 169 code pat t e rns.
By using publicly available face database of AT&T
Laboratories Cambridge [19], highest average recogni-
tion rate of 97.4% is obtained. Compared to the results of
4x4 codebook which the highest averag e recognition rate
is 95.6%, recognition rate increases by about 2 % .
But in [22], it was found that such distribution of
codevectors appears non-uniform and concentrated only
in some of the regions. According to Maximum Entropy
Principle (MEP), the maximum entropy distribution
⑨⑧⑦⑥⑤④③
x1 x8x8x8x4x8x16x2x4x8x8
abcd
>>
>
Figure 2. Categories of 2×2 code patterns. Pixel intensities
are marked by alphabet ‘a’, ‘b’, ‘c’, ‘d’, and a>b>c>d is
prescribed. The number of typical patterns for all 2×2 im-
age block is only 11, thus total number of image patterns
for 2×2 pixel blocks is theoretically only 75
will be achieved when the value of a random variable
(counts) equals the average. Such a non-uniform distri-
bution can not satisfy the MEP, so the recognition per-
formance can not be expected be best because the aver-
age information conten t will not be maximum.
As a solution, Chen et al. [22] optimized the cod ebook
by sorting the frequencies of all individual codevectors
abstracted from facial images and excluding the code-
vectors in the codebook with lo w frequen cy. This method
improved recognition performance and highest average
recognition rate of 98.2% is obtained by using the same
face database of AT&T Laboratories Cambridge [19].
3.2 Proposed Codebook Design Method
The essence of the method in [22] is to abstract the code
patterns which are most frequently used in facial images.
The consideration is correct, but it ignored the differ-
ences between different persons. The frequencies of
some code patterns used in facial images may vary
greatly for different persons and the values appear small.
The code patterns selected to be used in codebook should
not only present the most common features of faces, but
also discriminate the difference of persons. The latter
characteristic is very important in recognition task and
can be evaluated by mean square error (MSE) of code
patterns. Figure 3(a) shows the average histogram of 40
facial images in the database of AT&T Laboratories
Cambridge [19] by using the 2×2 codebook generated
according to the method proposed in [22]. In this case,
the codebook size is set to be 80. The distribution is
sorted in order of frequency. But the order of respective
MSE values is out of accord as shown in Figure 3(b).
The code patterns with high frequencies but low MSE
values are useful to generate feature vectors but poor
benefit recognition result. So we should consider both
factors of freq uen cie s and MSE values.
As a solution, our strategy is as follows.
Step 1: Create a systematically organized codebook by
applying an improved method based on the code classi-
fication including frequencies and MSE values.
Step 2: Create another codebook with the same size
using Kohonen’ s Self-Organizing Maps (SOM) [20].
Step 3: Combine the two codebook s obtained above to
generate the final optimized codebook consisted of 2×2
codevectors.
3.3 Data Set for Codebook Design
For covering the variations of the photo-taking condi-
tions, two different face databases which are publicly
available FERET database [18] and face database of
AT&T Laboratories Cambridge [19], are utilized to ana-
lyze the code patterns of facial images. 40 facial images
from different persons are selected from each database,
and the face regions are abstracted with the sizes of
146×200 and 92×112, respectively. The typical examples
of facial images are shown in Figure 4.
Copyright © 2010 SciRes JSEA
A Codebook Design Method for Robust VQ-Based Face Recognition Algorithm
122
0
20
40
60
80
100
0 10203040506070
Code number
Norm alized Counts
(a)
0
20
40
60
80
100
0 10203040506070
Code number
Normalized MSE
(b)
Figure 3. (a) Average histogram of 40 facial images using
codebook of size 80. (b) MSE of code patterns. The distri-
bution is sorted in order of frequency in (a). But the order
of respective MSE values is out of accord as shown in (b)
(a) (b)
Figure 4. Data set for analysis of code patterns: (a) FERET
database, (b) AT&T database
3.4 Codebook Generated by Code Classification
In the rules No. 2 of cod ebook creation in Subsection 3.1,
the intensity variation steps are set to be 3, 6, and 10.
From the distribution of the average histogram, we can
see the steps for creating code patterns are not suitable
obviously. We modify the rules and propose our im-
proved codebook design method.
Figure 5 shows the processing steps of the codebook
generation. At first, we change intensity differences
among the blocks to be from 1 to 10, and implement the
rules No.1–4 in Subsection 3.1 to create an initial large
codebook of size 517. Thus the code patterns can almost
cover all intensity variation. Utilizing this initial code-
book, VQ processing is performed for all intensity varia-
tion blocks divided from the facial image in dataset de-
scribed above, matched frequencies for each codevector
are counted and histogram of each facial image is gener-
ated. Then average histogram and the normalized fre-
quencies (i
f) are calculated, where M is the number of
facial images. MSE () (using RMSE, the square root
of MSE) of individual codevectors is calculated by For-
mula 1 and then the scores () are computed by the
weighted average between the normalized frequencies
(
i
e
i
s
i
f) and MSE () as shown in Formula (2).
i
e
2
1
1(()
M
iii
j
eff
M

)j
(1)
2
1
21
jj
ii
ik
ekfk
s (2)
where ki (i=1, 2) is a weighting coefficient of respec-
tive component. The values of k1, k2 are 1, 1 respec-
tively for all images used in our experiments, which
determined by actual experiments.
Next, the scores () of individual codevectors are
sorted, and the codevectors with high scores will be ex-
tracted. In this way, a systematically organized codebook
is generated.
i
s
3.5 Codebook Genera ted by Kohonen’s SOM
As a neural unsupervised learning algorithm, Koho-
nen’s Self-Organizing Maps (SOM) [20] is one of the
Small intensity-variat ion cod e s
One large codebook
(size 517)
system atical ly
organized codebook
(size N)
Input image
Histogram of each
image
Average histogram
VQ
Extract codes with high
frequency and high
mean squa re error
(MSE)by sorting
Intensity-variation codes of No.
2,3,4,5,6,8 a nd 9 (step 1 )
No intens ity- variation code
Small intensity-variat ion cod e s
One large codebook
(size 517)
system atical ly
organized codebook
(size N)
Input image
Histogram of each
image
Average histogram
VQ
Extract codes with high
frequency and high
mean squa re error
(MSE)by sorting
Intensity-variation codes of No.
2,3,4,5,6,8 a nd 9 (step 1 )
No intens ity- variation code
Figure 5. Processing steps of codebook generation based on
code classification
Copyright © 2010 SciRes JSEA
A Codebook Design Method for Robust VQ-Based Face Recognition Algorithm 123
standard algorithms used for codebook design in image
coding. In this paper, SOM algorithm will also be em-
ployed to generate a SOM codebook with the same size
as in Subsection 3.4.
Step 1: Transform the facial images in dataset to in-
tensity variation vectors, and combine all data together
into one training set.
Step 2: Specify the size of the codebook to N and ini-
tialize the codevectors by using continuous intensity
variation vect o rs.
Step 3: Sel ect a new traini ng vecto r from the traini ng set.
Step 4: Find the best-matching codevector closest to
the training vector.
Step 5: Move the best-matching and its neighborhood
codevectors towards the training vector.
Step 6: Repeat from Step 3 until the map converges.
3.6 Generation of Optimized Codebook
After the two codebooks described above are generated,
they will be combined into one codebook of size 2N.
Overlapped codevectors in codebook will be excluded.
Next, like the processing in Subsection 3.4, VQ process-
ing is also performed for all intensity variation blocks
divided from the facial image in dataset, and average
histogram of all images is calculated. The scores of all
individual codevectors are calculated and sorted, and the
codevectors in this codebook with low scores will be
excluded. Thus, the size of codebook will be decreased
from 2N to N, and the final optimized codebook con-
sisted of 2×2 codevectors is generated.
4. Experiments and Discussions
4.1 Database
Face database of AT&T Laboratories Cambridge [19]
is used for recognition experiments. In the database, 10
facial images for each of 40 persons (totally 400 im-
ages) with variations in face angles, face sizes, facial
expressions, and lighting conditions are included. Each
image has a resolution of 92×112. Five images were
selected from each person’s 10 images as probe images
and remaining five images are registered as album im-
ages. Recognition experiment is carried out for 252
(10C5) probe-album combinations by rotation method.
The algorithm is programmed by ANSI C and run on
PC (Pentium(R)D processor 840 3.2GHz).
4.2 Experimental Results and Discussions
Firstly, we discuss the codebook size N. It is necessary
to choose a suitable size of codebook. As the codebook
size is too large, number of codevectors increases, the
resolution of the histogram may become so sensitive
that noise-corrupted codevectors may significantly
distort the histogram. On the contrary, if the number of
codevectors is too small, the histogram can not suffi-
ciently discriminate between different faces.
Figure 6 shows the comparison of the recognition
results using codebooks with different sizes from 30 to
200. The highest average recognition rates obtained in
each codebook are shown here. The best performance
is obtained at codebook size of 80 which is the same as
in [22]. Maximum of the average rate 98.6% is
achieved, which is 0.4% higher than that in [22].
We also compare the proposed algorithm with con-
ventional algorithms. Figure 7 shows the recognition
results. Recognition success rates are shown as a
function of filter size which changes from none fil-
tering to filter size of 23×23. The curve with rhombus
marks stand for the average results in 252 (10C5)
probe-album combinations using the 2×2 codebook
created by proposed design method. The codebook size
of 80 is used here. The curve with triangle marks and
circular marks refer to the results reported in [21,22]
which use original 2×2 codebook. The curve with open
square marks refers to the results of 4×4 codebook [15].
In the curve with rhombus marks, recognition rate
first increases with increase in filter size, and then,
saturated or gradually decreases. The highest average
recognition rate of 98.6% is obtained at the filter size
of 13×13 while that of 98.2%, 97.4% and 95.6% using
codebooks in [15,21,22], respectively. It can be said
96
96.5
97
97.5
98
98.5
99
99.5
100
305080100120 140 160 180 200
Codebook size
Max. ave. rec ognition rate (%)
Figure 6. Comparison of recognition results in different
codebook size
86
88
90
92
94
96
98
100
none3x35x57x79x911x11 13x13 15x15 17x1719x19 21x21 23x23
Filter size
Recog nit ion Rate (%)
CB2x2_80new
CB2x2_80
CB2x2_169
CB4x4_original
Figure 7. Comparison of the recognition results
Copyright © 2010 SciRes JSEA
A Codebook Design Method for Robust VQ-Based Face Recognition Algorithm
Copyright © 2010 SciRes JSEA
124
that the 2×2 codebook created by proposed method
represent the features of the facial images more ade-
quately than conventional codebooks.
4.3 Processing Time
Processing time for single image in the face database of
AT&T Laboratories Cambridge [19] is about 57 msec
using a codebook of size 80, which is composed of 15
msec for pretreatment including filtering, block divi-
sion, and minimum intensity subtraction, 30 msec for
VQ processing, and 12 msec for database matching.
Furthermore, because a 2×2 codevector can be rep-
resented by an array of 4 dimensions, by utilizing the
table look-up (TLU) method in the VQ processing step,
the VQ processing time can be shorten to be about 1
msec, and the total running time will be 28 msec. It
means our fast recognition algorithm achieves real-
time face recognition.
5. Conclusions
In this paper, a theoretical codebook design method for
robust VQ-based face recognition algorithm is pro-
posed. Combining a systematically organized codebook
based on the classification of code patterns and another
codebook created by Kohonen’s Self-Organizing Maps
(SOM), an optimized codebook consisted of 2×2 code-
vectors for facial images is generated. Utilizing such a
codebook of size 80, the highest average recognition
rate of 98.6% is obtained for 40 persons’ 400 images of
the database of AT&T Laboratories Cambridge.
REFERENCES
[1] K. W. Bowyer, “Face recognition technology and the
security versus privacy tradeoff,” IEEE Technology and
Society, pp. 9–20, Spring 2004.
[2] M. Turk and A. Pentland, “Eigenfaces for recognition,”
Journal of Cognitive Neuroscience, Vol. 3, No. 1, pp.
71–86, March 1991.
[3] L. Wiskott, J. M. Fellous, N. Kruger, and C. von der Mal s-
burg, “Face recognition by elastic bunch graph matching,”
IEEE Transactions Pattern Analysis and Machine Intelli-
gence, Vol. 19, No. 7, pp. 775–779, July 1997.
[4] R. Brunelli and T. Poggio, “Face recognition: Features
versus templates,” IEEE Transactions Pattern Analysis
and Machine Intelligence, Vol. 15, No. 10, pp.
1042–1052, October 1993.
[5] P. S. Penev and J. J. Atick, “Local feature analysis: A
general statistical theory for object representation,” Net-
work: Computation in Neural Systems, Vol. 7, No. 3, pp.
477–500, 1996.
[6] R. Chellappa, C. L. Wilson, and S. Sirohey, “Human and
machine recognition of faces: A survey,” Proceedings
IEEE, Vol. 83, No. 5, pp. 705–740, May 1995.
[7] S. Z. Li and A. K. Jain, “Handbook of face recognition,”
Springer, New York, 2005.
[8] F. Goudail, E. Lange, T. Iwamoto, K. Kyuma, and N.
Otsu, “Face recognition system using local autocorrela-
tions and multi-scale integration,” IEEE Transactions Pat-
tern Analysis and Machine Intelligence, Vol. 18, No. 10,
pp. 1024–1028, 1996.
[9] K. M. Lam and H. Yan, “An analytic-to-holistic approach
for face recognition based on a single frontal view,” IEEE
Transactions Pattern Analysis and Machine Intelligence,
Vol. 20, No. 7, pp. 673–686, 1998.
[10] B. Moghaddam and A. Pentland, “Probabilistic visual
learning for object representation,” IEEE Transactions
Pattern Analysis and Machine Intelligence, Vol. 19, No. 7,
pp. 696–710, 1997.
[11] W. Zhao, “Discriminant component analysis for face recog-
nition,” Proceedings ICPR’00, Track 2, pp. 822–825, 2000.
[12] M. S. Bartlett, J. R. Movellan, a nd T. J. Sejnowski, “Face
recognition by independent component analysis,” IEEE
Transactions on Neural Networks, Vol. 13, No. 6, pp.
1450–1464, 2002.
[13] S. G. Karungaru, M. Fukumi, and N. Akamatsu, “Face
recognition in color images using neural networks and
genetic algorithms,” International Journal of Computa-
tional Intelligence and Applications, Vol. 5, No. 1, pp.
55–67, 2005.
[14] S. Aly, N. Tsuruta, and R. Taniguchi, “Face recognition
under varying illumination using Mahalanobis self-orga-
nizing map,” Artificial Life and Robotics, Vol. 13, No. 1,
pp. 298–301, 2008.
[15] K. Kotani, Q. Chen, and T. Ohmi, “Face recognition using
vector quantization histogram method,” IEEE 2002 Interna-
tional Conference o n Image Pr ocessing , II–105– 108, 2002 .
[16] A. Gersho and R. M. Gray, “Vector quantization and
signal compression,” Kluwer Academic, 1992.
[17] T. Nakayama, M. Konda, K. Takeuchi, K. Kotani, and T.
Ohmi, “Still image compression with adaptive resolution
vector quantization technique,” International Journal of
Intelligent Automation and Soft Computing, Vol. 10, No.
2, pp. 155–166, 2004.
[18] P. J. Phillips, H. Wechsler, J. Huang, and P. Rauss, “The
FERET database and evaluation procedure for face rec-
ognition algorithms,” Image and Vision Computing
Journal, Vol. 16, No. 5, pp. 295–306, 1998.
[19] AT&T Laboratories Cambridge, “The database of faces,”
at: http://www.cl.cam.ac.uk/research/dtg/attarchive/face
database.html
[20] T. Kohonen, “Self-Organizing Maps,” Springer, USA,
1995.
[21] Q. Chen, K. Kotani, F. Lee, and T. Ohmi, “Face recogni-
tion using codebook designed by code classification,”
2006 IEEE International Conference on Signal and Image
Processing, Hubli, India, pp. 397–401, December 2006.
[22] Q. Chen, K. Kotani, F. Lee, and T. Ohmi, “A VQ-based
fast face recognition algorithm using optimized code-
book,” Proceedings of the 2008 International Conference
on Wavelet Analysis and Pattern Recognition, Hong
Kong, pp. 298–303, August 2008.
[23] Q. Chen, K. Kotani, F. Lee, and T. Ohmi, “VQ-based face
recognition algorithm using code pattern classification
and Self-Organizing Maps,” Proceedings of 9th Interna-
tional Conference on Signal Processing (ICSP’08), Bei-
jing, pp. 2059–2064, October 2008.