Journal of Intelligent Learning Systems and Applications, 2010, 2, 221-228
doi:10.4236/jilsa.2010.24025 Published Online November 2010 (http://www.scirp.org/journal/jilsa)
A Feasible Approach for Automatic Detection and
Recognition of the Bengalese Finch Songnotes and
Their Sequences
Khan Md. Mahfuzus Salam1,*, Tetsuro Nishino1, Kazutoshi Sasahara2, Miki Takahasi2,
Kazuo Okanoya2
1Department of Information and Communication Engineering, the University of Electro-Communications, Tokyo, Japan; 2Laboratory
for Bioliguistics, RIKEN Brain Science Institute (BSI), Saitama, Japan. *corresponding author
Received March 23rd, 2010; revised September 14th, 2010; accepted September 18th, 2010.
ABSTRACT
The Bengalese finch song has been widely studied for its unique features and similarity to human language. For com-
putational analysis the songs must be represented in songnote sequences. An automated approach for this purpose is
highly desired since manual processing makes human annotation cumbersome, and human annotation is very heuristic
and easily lacks objectivity. In this paper, we propose a new approach for automatic detection and recognition of the
songnote sequences via image processing. The proposed method is based on human recognition process to visually
identify the patterns in a sonogram image. The songnotes of the Bengalese finch are dependent on the birds and similar
pattern does not exist in two different birds. Considering this constraint, our experiments on real birdsong data of dif-
ferent Bengalese finch show high accuracy rates for automatic detection and recognition of the songnotes. These results
indicate that the proposed approach is feasible and generalized for any Bengalese finch songs.
Keywords: Birdsong Analysis, Bengalese Finch Song, Songnote Detection and Recognition, Pattern Recognition
1. Introduction
Birdsong has been actively studied via analysis of song-
note sequences to understand the language model of
birds. The songs of the Bengalese finch (Lonchura stri-
ata var. domestica) – a popular fowl in Japan, is widely
employed for this purpose. The song of the Bengalese
finch has a complex structure as compared with those of
other songbirds such as zebra finches (Taeniopygia gut-
tata) [1]. Thus, Bengalese finch songs have been studied
as a model of human language. According to the recent
studies, the courtship songs of Bengalese finches have
unique features and similarity with a human language [2].
In birdsong research, acoustic song analysis is necessary
to find the song elements and their sequence for carrying
out an analysis to understand the song syntax [3] and the
learning process of the song. The current research is fo-
cused on automatic detection and recognition of the
songnote and its sequence. Previous studies that em-
ployed sound processing had drawbacks as an automated
approach. This paper introduces a new generalized ap-
proach that employs image processing to overcome the
drawbacks.
2. Preliminaries
This section briefly introduces the theoretical founda-
tions of a birdsong, its representations, image basics, and
the recognition process by humans as we focused on the
recognition process that is manually carried out by hu-
mans.
2.1. Birdsong Representation
In birdsong analysis, the song data is recorded in an ap-
propriate environment – special cage equipped with au-
tomated recording system and also to avoid noise. From
the recorded sound data, we obtain the sonogram image
of the song. For further computational analysis, the ob-
tained sonogram image is used as the standard represen-
tation of the song [4].
The following of this section briefly explains some
general terms that are used in birdsong research.
Songnote: An independent pattern appearing in sono-

A Feasible Approach for Automatic Detection and Recognition of the Bengalese Finch Songnotes and Their Sequences
222
gram which is assigned a symbol is called a songnote. It
is also referred as a song element or a behavioral element.
From the definition, we can say the text data consisting
of symbols (such as a, b, c, and so on) are called song-
note sequence. Songnotes are analogous to phonemes in
human language.
Chunk: A fixed sequence of song notes is called a
chunk. In Figure 1, for example, the chunks are ab, cde
and fg. Chunks are analogous to words in human lan-
guage.
Song unit: A song unit consists of chunks. Song units
are analogous to sentences in human language.
Sonogram: A sonogram is an image that shows how
the spectral density of a signal varies with time. It is also
known as a spectrogram, voiceprint, or voicegram. So-
nogram are used to identify phonetic sounds to analyze
the animal cries and also in the fields of speech process-
ing, music, sonar/radar, seismology, etc.
There are many variations in the format of the sono-
gram. Sometimes, the vertical and horizontal axes are
switched; sometimes, the amplitude is represented as the
height of a 3D surface instead of color or intensity. The
frequency and amplitude axes can be either linear or lo-
garithmic, depending on what the graph is being used for.
For instance, audio would usually be represented with a
logarithmic amplitude axis, and frequency would be lin-
ear in order to emphasize harmonic relationships, or
logarithmic to emphasize musical, tonal relationships.
The most common format is a graph with two geometric
dimensions: the horizontal axis represents time, and the
vertical axis is frequency; a third dimension indicating
the amplitude of a particular frequency at a particular
time is represented by the intensity or color of each point
in the image. For the birdsong research this common
format is used. Figure 1 shows a sample grayscale so-
nogram image of a Bengalese finch courtship song.
2.2. Bengalese Finch Song
Recent studies on Bengalese finches show that the songs
of male Bengalese finches are neither monotonous nor
random; they consist of chunks, each of which is a fixed
Figure 1. Grayscale sonogram image of a Bengalese finch
song.
Figure 2. Courtship song syntax represented by an auto-
maton
sequence of a few song notes. The song of each individ-
ual can be represented by a finite automaton, which is
called song syntax (see Figure 2) [2]. The songs of
Bengalese finches have double articulation – a sentence
consists of words, and each word consists of phonemes,
which is also one of the important faculties of human
language.
The song syntax is manipulated by the song control
nuclei in the brain. The hierarchy of the song control
nuclei directly corresponds to the song hierarchy [5].
Because of the structural and functional similarities of
vocal leaning between songbirds and humans, the former
have been actively studied as a good model of a human
language [6]. In particular, the song syntax of Bengalese
finches sheds light on the biological foundations of syn-
tax.
2.3. Detection and Recognition
Human vision is one of the most important and percep-
tive mechanisms. It provides information required for the
relatively simple tasks (e.g., object recognition) and for
very complex tasks as well. In bird song research, the
songnote recognition is carried out by humans by in-
specting the patterns visually represented in a sonogram
image [4].
2.3.1. Image Feature Extraction
Digital image processing denotes the analysis carried out
on the basis of the pixel property of the image irrespec-
tive of the image type. A digital image has a finite set of
digital values called picture elements or pixels. The im-
age contains a fixed number of rows and columns of pix-
els. Pixels are the smallest individual elements in an im-
age, holding quantized values that represent the bright-
ness of a given color at any specific point. Typically, the
pixels are stored in computer memory as a raster image
or raster map, a two-dimensional array of small integers.
These values are often transmitted or stored in a com-
pressed form.
Each pixel of a raster image is typically associated
with a specific position in some 2D region and has a
value of one or more quantities related to that position.
Digital images can be classified according to the number
and nature of such samples into different categories like
Binary, Grayscale, Color and False-color. In our re-
search, we use a grayscale sonogram image.

A Feasible Approach for Automatic Detection and Recognition of the Bengalese Finch Songnotes and Their Sequences 223
Grayscale Image: A grayscale digital image is an im-
age in which the value of each pixel is a single sample,
that is, it carries only intensity information. In fact a gray
color is one in which the red, green, and blue compo-
nents all have equal intensity in the RGB space, and
hence, it is only necessary to specify a single intensity
value for each pixel, as opposed to the three intensities
needed to specify each pixel in a full color image.
Pixel Values: For a grayscale image, the pixel value is
a single number that represents the brightness of the pix-
el. The intensity of a pixel is expressed within a given
range between a minimum and a maximum. Presently,
grayscale images are commonly stored with 8 bits per
sampled pixel, which allows 256 different intensities (i.e.,
shades of gray). The binary representations assume that 0
is black and the maximum value 255 is white.
2.3.2. Image Ma tching and Recogni tio n
Pattern recognition aims to classify data or patterns on
the basis of either a priori knowledge or statistical in-
formation extracted from the patterns. The patterns to be
classified are usually groups of measurements or obser-
vations, defining points in an appropriate multidimen-
sional space. This is in contrast to pattern matching,
where the pattern is rigidly specified. Pattern recognition
is used to test whether things have a desired structure, to
find relevant structure, to retrieve the aligning parts, and
to substitute the matching part with something else.
In human vision-based recognition of an image, the
first thing that will catch the attention is something that
is familiar. To be recognized, an object must have some
feature that our consciousness can assign. Behind this
process, the mental model captures the important char-
acteristics of the object. It is unfortunate that, in many
scientific experiments, the task assigned to human vision
is not the recognition of familiar objects, but the detec-
tion and description of unfamiliar ones, which is far
more difficult. According to the McCulloch and Pitts
simplified neuron model, the weighted sum of many in-
puts exceeds a threshold, and then the output is turned on.
(see Figure 3). Learning consists of adjusting the
weights, which can be either positive or negative [7].
The current research applies image processing meth-
odology based on grayscale image features of the sono-
gram. The motivation of applying such image processing
is to find a simple and generalized way for the automa-
tion as a human brain does in the recognition process by
applying pattern matching.
3. Methodology
The proposed automation process is divided into two
steps. First, from the song sonogram image, we detect
the song elements on the basis of the local property of
Figure 3. McCulloch and Pitts simplified model of a neuron
and its implementation as a threshold logic unit [7].
Figure 4. Process flow diagram of the songnote detection
and recognition.
the sonogram image. Then, on the basis of the detected
elements, we apply image matching to assign a label to
the extracted elements, and thus, we obtain the songnote
sequence of the song. Figure 4 shows the process flow
diagram of the proposed methodology.
3.1. Songnote Detection
From the sonogram image, we first detect the elements.
On the basis of the extracted statistical features of the
detected elements, we carry out the recognition process.
For this reason, the detection process is very important.
3.1.1. Detection Met hod
The detection process is carried out by analyzing the
sonogram image for intensity values; we can obtain a
graph for the average pixel intensity value. If the sono-
gram image has many noises at the beginning, which are
ignored in the visual inspection by human, the present
system does not ignore them as noises. For this reason,
we pre-process the sonogram image. Then, if we take the
average intensity value along the vertical line and draw a
graph where the Y-axis represents the average intensity
value or gray value and the X-axis represents the pixel
index x, which is the distance from the (0, 0) pixel along
the X-axis, we have a graph as follows:
The above graph (see Figure 5) is generated from the

A Feasible Approach for Automatic Detection and Recognition of the Bengalese Finch Songnotes and Their Sequences
224
sample sonogram image shown in Figure 6. It is clearly
visible that from the graph we can find some clear gaps
between the elements. By defining parameters (see Fig-
ure 6) such as minimum element width, minimum gap
between elements, and the intensity threshold, we can
execute our algorithm to find the song elements. If some
region does not fit with the three above mentioned pa-
rameters, we consider it to be noise. Note that these pa-
rameters can vary from bird to bird. The detected song
elements and the features of the elements, such as width
information, are used for the recognition process.
3.1.2. Detection Al go rithm
The song element detection algorithm takes the array of
the average intensity values as the input. On the basis of
the defined parameter values, the proposed detection
algorithm produces an unlabeled list of song elements.
Input: array of intensity values.
Output: a list of elements.
Procedure:
(1) Initialize the parameters.
(2) If the intensity value exceeds threshold and
next is not a gap
· set start element flag true;
· set start index to current index;
(3) If start element flag is true and next
minimum gap is detected
· set start element flag false;
· set end index to current index;
· add to element list;
(4) Continue step 2 and 3 until end of the
intensity array
(5) Return element list.
Detect ion Algo rithm
3.2. Songnote Recognition
For extracting the songnote sequence from the sonogram
image, we extract local statistical features and then carry
out the statistical pattern matching for recognition.
3.2.1. Recog nition Method
As discussed in the previous section, similar patterns are
assigned with the same label in the recognition process.
Our recognition method is based on the local property of
the sonogram image. By executing the note detection
algorithm, we obtain element list information. This un-
labeled element list provides the start pixel and the end
pixel information for every element.
As for the Bengalese finch song, note patterns differ
from bird to bird. Therefore, we decided not to use any
prior knowledge; rather, we use the statistical informa-
tion extracted from the patterns. See Figure 7. First, we
divide every note into N regions, and every region is
Figure 5. Average intensity value graph derived from the
sonogram image.
Figure 6. Sample sonogram image and the parameters.
12310171 20
27 23037 3
{ ,...,,,,
...,,,,...,, }
c
cc
RRR RgRgRgRg
Rg RgRgRg Rg

Figure 7. Explains the procedure while N = 3.
divided into nine (3×3) cells. We denote the center cell
as gc and the other cells as gn in a clockwise direction,
where n = 0, 1, …, 7. Thus, we obtain a set of values for
every single element. Then, we apply a statistical test
called the chi-square test to find the similarity between
elements. Note that the value of N should not be greater
than 3 because if the set size exceeds thirty, the Chi-square
distribution tends toward a normal distribution.
3.2.2. Chi-square Goodness Fit Test
The chi-square test (χ2) is a statistical hypothesis test
whose results are evaluated by reference to the

A Feasible Approach for Automatic Detection and Recognition of the Bengalese Finch Songnotes and Their Sequences 225
chi-square distribution. Pearson’s chi-square test is the
best-known of several chi-square tests. Its properties
were first investigated by Karl Pearson [8]. Pearson’s
chi-square is the original and most widely-used
chi-square test.
When an analyst attempts to fit a statistical model to
observed data to find how well the model actually re-
flects the data, i.e., how close are the observed values to
those that would be expected under the fitted model, one
statistical test that addresses this issue is the chi-square
goodness of fit test. This test is commonly used to test
the association of variables in two-way tables, where the
assumed model of independence is evaluated against the
observed data. In general, the chi-square test statistic is
of the following form:


Expected
ExpectedObserved 2
2

In the equation, χ2 is of the form:





n
ii
ii
E
EO
1
2
2

where,
χ2 = the test statistic that asymptotically
approaches a χ2 distribution.
Oi = an observed frequency;
Ei = an expected frequency, asserted by
the null hypothesis;
n = the number of possible outcomes;
The chi-square statistic is calculated by finding the
difference between each observed and theoretical fre-
quency for each possible outcome, squaring these values,
dividing each by the theoretical frequency, and taking
the sum of the results. The chi-square statistic can be
used for calculating a p-value by comparing the value of
the statistic to a chi-square distribution. The number of
degrees of freedom is equal to the number of possible
outcomes minus 1. If the computed test statistic is larger
than the chi-square table [8] with (n – 1) degrees of
freedom, the observed and expected values are not close
and the model is a poor fit to the data.
For pattern recognition one famous statistical machine
learning approach is Support Vector Machines (SVM).
SVM separates the data space into two clusters over a
separation boundary defined by a non-linear function [9].
We can apply SVM when supervised learning is possible
and also the number of clusters is known. It is difficult to
apply this technique in our application where there are
several cluster exist which cannot be predefined. During
our experiment we also try to employ another image pat-
tern recognition technique presented by Ojala et al. [10]
but unfortunately, we could not obtain good result. As
the songnote patterns are dependent to the birds we have
limitations for preparing the training set. For that reason,
we find that chi-square test is suitable for our application
comparing to other state-of-the-art techniques for pattern
recognition.
3.2.3. Recog nition Algorit hm
The songnote recognition algorithm takes the unlabeled
list of song elements. It applies the goodness of fit test to
find the similarity between elements and produces the
songnote sequence.
Input: unlabeled list of elements.
Output: labeled list of elements.
Procedure:
(1) For each element in element list divide
into N × 9 cells where 0 < N < 4
(2) Calculate the average intensity value
for every cell
(3) For each element until there is any
unlabeled element
· set one as expected and others as observed;
· if expected is not labeled set it with a new
label;
· test the Chi-square statistics;
(4) If the observed element pass the test then
set the element with same label
(5) Return updated element list
Recogn ition A lgorithm
4. Results
In this section, we present the results of our methodology
for analyzing the Bengalese finch song. First, we explain
the nature of our real song data, and then discuss the re-
sults of the automatic detection and recognition of the
songnote.
4.1. Description of Data
For testing our proposed method we use five different
song unit or phrase for each of the three matured Ben-
galese finch song; the names of the finches are Hikari 52,
Hikari 49 and Kuro 0362. The song data were recorded
at the Okanoya laboratory of RIKEN. The spectrogram
image of the matured Bengalese finch has similar prop-
erties, where the note patterns are clearly visible and
almost each songnote is separated by considerable blank
space. Figure 8 shows the partial sonogram images for
the three birds.
Table 1 shows the sample sonogram images contains
forty six to fifty four notes for Hikari 52, fifty one to
fifty nine notes for Hikari 49 and fifty two to sixty one
notes for Kuro 0362. From the sample sonogram image
Figure 8, it is clearly visible that the sonogram image of

A Feasible Approach for Automatic Detection and Recognition of the Bengalese Finch Songnotes and Their Sequences
226
Figure 8. Spectrogram for Hikari 49 (top), Hikari 52 (middle) and Kuro 0362 (bottom).
Table 1. Results of automatic detection of song elements.
Bird name Number of ap-
peared elements Average accuracy rate
Hikari 52 46 - 54 98%
Hikari 49 51 - 59 90%
Kuro 0362 52 - 61 95%
Hikari 49 is more complex than that of Hikari 52 and
Kuro 0362, i.e., for Hikari 52 and Kuro 0362, the song
notes are almost clearly separated from one to another,
but for Hikari 49, the song notes are not clearly separated
from one another.
Figure 9. Description of noise and effect of applying cutoff
level for Hikari 52.
at the intensity value graph, we obtain 40 extracted ele-
ments while the original numbers of elements are 46.
Therefore, the accuracy rate decreases, and certain ele-
ments lose some necessary information, which is not
desirable. Figure 9 describes the noise situation.
By applying our methodology, we implemented an
application in JAVA, which takes the sonogram image as
an input and provides extracted song elements and their
sequence as the output. ImageJ API [11] is used for ana-
lyzing the image property. In the case of Hikari 49, when we inspect the extracted patterns, we find that some song notes are not extracted
correctly. Initially, we have an accuracy rate of 75% with
our default parameter value as the gaps between the ele-
ments are too short to separate. Figure 9 describes the
errors in the detection process.
4.2. Songnote Detection
In Section 3.1, we discussed the song note extraction
methodology and explained the algorithm used for ex-
tracting the song notes from a sonogram image. We used
parameters such as minimum note width, intensity thre-
shold, and minimum gap between notes. We set the pa-
rameter values for minimum note width as 10 pixels,
intensity threshold as 250, and minimum gap between
notes as 5 pixels for every bird. After executing the algo-
rithm mentioned in Section 3.1.2, we obtain the result for
the best case as follows:
Figure 10, except Figure 10(d), shows some incorrect
extracted notes for Hikari 49. If we carefully inspect
Figure 10, we can observe that Figure 10(a) and Figure
10(b) should be extracted as two different elements be-
cause the right pattern in Figure 10(a) and Figure 10(b)
appears separately (see Figure 10(c)) in the sonogram
image, and Figure 10(c) should be extracted as three
different elements However, Figure 10(d) is considered
to be extracted as a right pattern although it has the same
nature as the patterns shown in Figure 10(a, b, and c)
because the two patterns are very close and the left and
the right patterns do not appear separately in the song.
Now in the case of Hikari 52, when we inspect the ex-
tracted patterns, we find that there are some noises with
the extracted patterns although we have a good accuracy
rate. To avoid the noise, if we apply a cutoff level of 30

A Feasible Approach for Automatic Detection and Recognition of the Bengalese Finch Songnotes and Their Sequences 227
(a) (b) (c) (d)
Figure 10. Description of the error in the detection for Hi-
kari 49.
Table 2. Results of the song note recognition.
Bird name Accuracy rate
Hikari 52 86%
Hikari 49 85%
Kuro 0362 78%
We adjust the default parameter value of the minimum
gap between the notes to be two pixels and use the cutoff
level of nine. Thus, for the best case result we obtain an
accuracy rate of 90%.
4.3. Songnote Recognition
In Section 3.2, we discussed the songnote recognition
methodology and explained the algorithm. The first step
is to divide every extracted element into N parts, and
then calculate the average intensity value for every re-
gion. Thus, for every element, we have a set of 27 ele-
ment while N = 3. Then, we apply the Chi-square test
considering the note width information. In the proposed
method, we compare the elements if the note width is
greater than three-fourths or smaller than five-fourths of
the observed element. After executing the algorithm
mentioned in Section 3.2.3, we obtain the songnote se-
quences.
We can summarize the result for the recognition as
follows (See Table 2):
Notice that for Hikari 49, the result is based on ex-
tracted patterns in the previous step. If we consider the
wrong extracted pattern, then the accuracy rate become
around 70%.
For further discussion the songnote sequence of one
song unit that is produced by our system and the se-
quence by human annotation for Hikari 52 have been
shown below where the bold letters show the different
outcomes in recognition.
System (Hikari 52):
AABACDDEFGHEFGHIBJKLDEFAABACD-
DEFGHEFGHIBJKLDEF
Correct (Hikari 52):
AABLBDDEFGHEFGHICJKDDE-
FAABLBDDEFGHEFGHICJKDDEF
If we inspect the wrong decisions made by the system
for Hikari 52, we find that note B is labeled as C and
note L is labeled as D. This is because the incorrectly
labeled note contains a considerable noise (white part),
which affects the matching process. In the case of incor-
rectly labeling note L note A for Hikari 52, by carefully
observing each note, we find that the intensity density is
the same for both the notes (see Figure 11).
From Figure 11 it is clearly visible that the distribu-
tion of intensity density is the same for both the notes.
This causes the recognition error and is a limitation of
the proposed image matching algorithm. Notice that Note
1 and Note 4 are recognized as A, but originally by hu-
man annotation by inspecting the image and hearing to
the song, Note 4 was labeled L. We notice a similar rec-
ognition error in the case of Hikari 49 and Kuro 0362.
5. Conclusions and Discussion
The present study proposes a brand-new approach to
automatic recognition of song elements and its sequences
other then sound processing, and by applying image
processing, we obtain good results for the approach.
There are good possibilities to improve the accuracy rate
for both the extraction and the recognition methods to
some extent. From the obtained results, we find that the
element extraction process is very important and has a
significant effect on the recognition process. The major
advantage of the proposed approach is its simplicity and
feasibility. The approach is focused on a generalized (does
not depend on the bird) process same as humans do.
Figure 11. Note 1 (A, top left), note 4 (L, top right) and dis-
tribution of intensity density value (bottom) for Hikari 52.

A Feasible Approach for Automatic Detection and Recognition of the Bengalese Finch Songnotes and Their Sequences
228
Table 3(a). Comparison results for automatic detection.
Bird name accuracy rate
(sound processing)
accuracy rate
(proposed method)
Hikari 52 96% 98%
Hikari 49 94% 90%
Table 3(b). Comparison results for automatic recognition.
Bird name accuracy rate
(sound processing)
accuracy rate
(proposed method)
Hikari 52 83% 86%
Hikari 49 -- 85%
 Satisfactory clustering was not possible for Hikari 49 based on the pa-
rameter used.
The accuracy rate of the proposed approach is better
than that of other methods such as sound processing
which was previously carried out at our laboratory. The
following tables show comparison of the accuracy rate
between sound processing and our proposed method for
sngnote detection (see Table 3(a)) and recognition (see
Table 3(b)). For comparison we use the song data of
Hikari 52 and Hikari 49.
For the detection process sound processing method
uses Amplitude and Wiener Entropy. For the recognition
process it applies K-means clustering algorithm which
uses Duration
， Amplitude, Wiener Entropy, Mean Fre-
quency and Harmonic Pitch as parameters. However, the
sound processing requires considerable human effort for
fixing the parameter values (manual labeling has to be
done once if k-means algorithm has been applied) or for
training the system (each songnote has to be manually
separated to build the database if HMM approach has
been applied) for detecting and recognizing the song-
notes for every bird. Furthermore, it is not possible to
make a corpus for bird phonemes. If we employ sound
processing the only thing we can do is to train the system
for a specific bird family as similar patterns do not ap-
pear in different bird families. This is not practical for an
automated system. In contrast, the proposed methodol-
ogy is almost automated and feasible for songbirds as our
approach represents the human inspection method and
does not depend on birds. The default parameter values
ave been used for detecting the songnotes is almost
good for any bird but can be changed by couple of click
by the user if necessary.
h
For the element detection process, the accuracy rate is
100% for some birds, and for other birds, the accuracy
rate is also satisfactorily high. Thus, our approach saves
time and is practical as an automated system. In the rec-
ognition process, we obtain a high accuracy rate of more
than 80%.
REFERENCES
[1] E. Honda and K. Okanoya, “Acoustical and Syntactical
Comparisons between Songs of the White-Backed Munia
(Lonchura striata) and Its Domesticated Strain, the Ben-
galese Finch (Lonchura Striata Var. Domestica),” Zoo-
logical Science, Vol. 16, 1999, pp. 319-326.
[2] K. Okanoya, “Song Syntax in Bengalese Finches: Proxi-
mate and Ultimate Analyses,” Advances in the Study of
Behavior, Vol. 34, 2004, pp. 297-346.
[3] Y. Kakishita, K. Sasahara, T. Nishino, M. Takahasi and
K. Okanoya, “Ethological Data Mining: an Auto-
mata-Based Approach to Extract Behavioral Units and
Rules,” Data Mining and Knowledge Discovery, Vol. 18,
No. 3, 2009, pp. 446-471.
[4] C. K. Catchpole and P. J. B. Slater, “Bird Song: Biologi-
cal Themes and Variations,” 2nd Edition, Cambridge
University Press, UK, 2003.
[5] J. Nishikawa and K. Okanoya, Ornithological Science,
2006, pp. 95-103.
[6] J. Doupe and P. K. Kuhl, “Birdsong and Human Speech:
Common Themes and Mechanisms,” Annual Reviews
Neuroscience, Vol. 22, 1999, pp. 567-631.
[7] J. C. Russ, “The Image Processing Handbook,” 5th Edi-
tion, CRC Press, 2006.
[8] S. M. Ross, “Introduction to Probability and Statistics for
Engineers and Scientists,” 3rd Edition, Elsevier Aca-
demic Press, USA, 2004.
[9] J. Shawe-Taylor and N. Cristianini, “An Introduction to
Support Vector Machines and Other Kernel-Based
Learning Methods,” 1st Edition, Cambridge University
Press, UK, 2000.
[10] T. Ojala, M. Pietikainen and D. Harwood, “A Compara-
tive Study of Texture Measures with Classification Based
on Feature Distributions,” Pattern Recognition, Vol. 29,
No. 1, 1996, pp. 51-59.
[11] ImageJ 1.41, National Institutes of Health, USA, 2010.
http://rsbweb.nih.gov/ij/