![International Journal of Trend in Scientific Research and Development (IJTSRD)
Volume 3 Issue 5, August 2019 Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470
@ IJTSRD | Unique Paper ID – IJTSRD26775 | Volume – 3 | Issue – 5 | July - August 2019 Page 1610
National Flags Recognition Based on
Principal Component Analysis
Soe Moe Myint, Moe Moe Myint, Aye Aye Cho
University of Computer Studies, Pathein, Myanmar
How to cite this paper: Soe Moe Myint |
Moe Moe Myint | Aye Aye Cho "National
Flags Recognition Based on Principal
Component Analysis" Published in
International
Journal of Trend in
Scientific Research
and Development
(ijtsrd), ISSN: 2456-
6470, Volume-3 |
Issue-5, August
2019,pp.1610-1614,
https://doi.org/10.31142/ijtsrd26775
Copyright © 2019 by author(s) and
International Journalof Trendin Scientific
Research and Development Journal. This
is an Open Access article distributed
under the terms of
the Creative
CommonsAttribution
License (CC BY 4.0)
(http://creativecommons.org/licenses/by
/4.0)
ABSTRACT
Recognizing an unknown flag in a scene is challenging due to the diversity of
the data and to the complexity of the identification process. And flags are
associated with geographical regions, countries and nations. But flag
identification of different countries is a challenging and difficult task.
Recognition of an unknown flag image in a scene is challenging due to the
diversity of the data and to the complexity of the identification process. The
aim of the study is to propose a feature extraction based recognition system
for Myanmar’s national flag. Image features are acquired from the region and
state of flags which are identifiedbyusingprincipalcomponentanalysis(PCA).
PCA is a statistical approach used for reducing the number of features in
National flags recognition system.
KEYWORDS: PCA, texture feature
1. INTRODUCTION
Flag recognition systems are part of image processing applications and their
significance as a research areas are increasing recently. J. Nagi, S. K. Ahmed [2]
proposed a new technique for human facerecognition thatusesan image-based
approach towards artificial intelligence by removing redundant data from face
images through image compression using the two-dimensional discrete cosine
transform (2D-DCT). Feature vectors are constructed by computing DCT
coefficients.
A self-organizing map (SOM) usinganunsupervised learning
techniques used to classify DCT-based feature vectors into
groups to identify if the subject in theinputimageis“present
or “not present” in the image database.
Md. Zamilur Rahman and Mohammad Shameemmhossain
Kawsar Ahmed [5] proposed an approach based on support
vector machine (SVM). This machine is trained with the
percentage of different colors in the flag. Totrain uptheSVM
have used a package LIBSVM which classify multiple classes
and find out an accurate country id (name) as output. The
system is designed to detect a flag more accurately and
implement a system that is more efficient in this problem
domain by reducing the time for flag detection andthevisual
recognition for man.
K. Sulovská, S. Bělašková, M. Adámek [3] proposed the
analytical-statistical method for the face recognition.
Although the bases of the face recognition are known by
researchers worldwide, the statistical tests of data obtained
by measuring chosen anthropometrical points can be found
in several articles. Their aim was to show how the data act
during the various emotions of one face, which will be
helpful for deeper knowledge of how the face behaves.
Acquired results reflect the difficulty of describing the face
and the applicability of combination of different recognition
methods (e.g. methods based on neural networks,
recognition of facial contours, distribution of the gray scale
in the image, deformation models) to get the best results in
the verification/identification of a human.
The aim of this study is to propose flag recognition based on
image featuresextractionwith principal componentanalysis.
The proposed system overcomes certain limitation of the
existing recognition system. It is based on extracting the
texture features of a set of Myanmar’snationalflagsstored in
the database and performing mathematical operations on
the values corresponding to them. The proposed system is
better mainly due to the use of image features rather than
the entire flag. Their advantages are in terms of: recognition
accuracy and better discriminatory power computational
cost because smaller images require less processing to train
the PCA. Because of the use of dominant features and hence
can be used as an effective means of authentication.
This paper is composed as follows: Methodology of the
system is described in section two. In section three, Data
Acquisition and in section four includes Result and
Discussion. Finally, the paper has been concluded.
2. Methodology
In this section, related methodologies of the proposed
system which are discussed with two parts. The first part
will explain the texture features in digital image processing.
The second part will describe the procedure of principal
component analysis forrecognizinganddetectingofnational
flags. The block diagram of the proposed flag recognition
methods is given in Figure1.
IJTSRD26775](https://image.slidesharecdn.com/304nationalflagsrecognitionbasedonprincipalcomponentanalysis-190913085810/85/National-Flags-Recognition-Based-on-Principal-Component-Analysis-1-320.jpg)
![International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
@ IJTSRD | Unique Paper ID – IJTSRD26775 | Volume – 3 | Issue – 5 | July - August 2019 Page 1611
Figure1. Flag Recognition system
The proposed flag recognition system is shown in Figure1,
the first step need to extract texture features from the
training and testing flag images. In thesecond step, extracted
features from the dataset and testing flag image are
recognized by using principal component analysis for
dimension reduction.
2.1. Feature Extraction
The feature based methods do not directly work with image
intensity values, but use salient features extracted from two
images, which has been shown to be more suitable for such
situation that intensity changes and complicated geometric
deformations are encountered. Texture is a key component
of human visual perception. Everyone can recognize texture
but, it is more difficult to define. Texture has qualities such
as periodicity and scale; it can be described in terms of
direction, coarseness, contrast and so on. In this system
chose three very different approaches to computing texture
features: the first takes a statistical approach in the form of
co-occurrence matrices, next the psychological view of
Tamura’s features and finally signal processing with Gabor
wavelets.
2.1.1. Texture feature
Texture is a main component of human visual perception.
Everyone can determine texture but it is more difficult to
define. Table 1 shows the normalized co-occurrence matrix.
Table1: Features Calculated from the Normalized Co-
occurrence Matrix P(i, j).
Texture Features Formula
Energy ∑i∑j P2(i,j)
Entropy ∑i∑j P(i,j)log P(i,j)
Contrast ∑i∑j (i – j)2P(i,j)
Homogeneity ∑i∑j (P(i,j)/1 + |i – j|
where P = co-occurrence matrix, µ = mean of the co-
occurrence matrix P, σ = standard variation of co-occurrence
matrix P.
Co-occurrence matrices
Haralick [1] suggested the use of grey level co-occurrence
matrices (GLCM) to extract second order statistics from an
image.
Tamura
Tamura et al took the approach of devising texture features
that correspond to human visual perception [5]. The first
three attained very successful results and are used in our
evaluation.
Gabor
One of the most popular signalprocessing-based approaches
for texture feature extraction has been the use of Gabor
filters. This system implementation is based on that of
Manjunath et al. [3, 6]. Gabor wavelet transform is then
defined to be:
( ) ( ) ( )*
1 1 1 1, , d dmn mnW x y I x y g xx y x y−∫ (1)
The mean and standard deviationofthemagnitude|Wmn| are
used to for the feature vector.
Table2: First three Tamura features.
Tamura Formula
Coarseness
Contrast
Directionality
magnitude of the vector,
angle of the vector,
2.2. Principal Component Analysis
Principal Components Analysis (PCA) is a dimensionality
reduction algorithm that can be used to significantly speed
up your unsupervised feature learning algorithm. More
importantly, understanding PCA will enable us to later
implement whitening,which isan importantpre-processing
step for many algorithms [10]. PCA is mostlyused as atool in
exploratory data analysis and for making predictive models.
It is often used to visualize genetic distance and relatedness
between populations [9].
Standard Principle ComponentAnalysis(PCA) isoftenuseful
preprocessing strategy in ICA is to first whiten the observed
variables. This transform observed vector linearly so that
obtain a new vector which is white. Its component are
uncorrelated and their variances equal unity.
(2)
One popular method for whitening is to use the eigenvalue
decomposition (EVD) of the covariance
matrix , where E is orthogonal matrix of
eigenvectors of and D is the diagonal matrix of its
eigenvalues ).Notethat canbe
estimated in a standard way from the variable sample
whitening can now be done by
(3)
For example, rank of D is equal to two for image features,
meaning that observed and training features are
uncorrelated. On the other hand, if the flag is not recognized,
this mixtures are actually the combination of one features
only, hence, the rank of D will be reduced to one.
3. Data Acquisition
Flag recognition system is composed of two stages: Data
Acquisition and Result and Discussion that consists of two
parts: Feature Extraction and Flag Recognition with PCA.](https://image.slidesharecdn.com/304nationalflagsrecognitionbasedonprincipalcomponentanalysis-190913085810/85/National-Flags-Recognition-Based-on-Principal-Component-Analysis-2-320.jpg)
![International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
@ IJTSRD | Unique Paper ID – IJTSRD26775 | Volume – 3 | Issue – 5 | July - August 2019 Page 1614
Principal Component Analysis (PCA) is a well-established mathematical technique for reducing the dimensionality of data,
while keeping as much variation as possible. PCA achieves dimension reduction by creating new, artificial variables called
principal components. Each principal component is a linear combination of the observed variables. According to the results,
features extraction based flag recognition system can recognize the correct flag image for testing data. In this experiment,
receiver operating characteristic (ROC) curve has been used to verify the effectiveness of the proposed method. Figure3gives
the ROC curves as the flag recognition results. It can be seen that the variation of the performance function for training and
testing.
Figure3. ROC curve under different type of flag image
5. Conclusion
National flag recognition system is implemented by using
texture features and principal component analysis. The
computer system can automatically recognized all national
flags of Myanmar via loaded from camera, web cameras or
natural scenes. Applying texture based features extraction
techniques are combined to use with PCA, which reported
better accuracy for flags recognition results. The eigenvector
for PCA approach thus provided a practical solution that is
well fitted to the problem of flag recognition system. PCA
achieves dimension reduction by creating new, artificial
variables called principal components. Each principal
component is a linear combination of theobserved variables.
Recognition accuracy of the proposed system from 90 % to
95 % depending on weather the flag is among 14 state and
region.
REFERENCE
[1] G. N. Srinivasan, and Shobha G: “Statistical Texture
Analysis,” PROCEEDINGS OF WORLD ACADEMY OF
SCIENCE, ENGINEERING AND TECHNOLOGY VOLUME
36 DECEMBER 2008 ISSN 2070-3740.
[2] J. Nagi, S. K. Ahmed: “A MATLAB based Face
Recognition SystemusingImageProcessingand Neural
Networks,” 4th International Colloquium on Signal
Processing and its Applications, March 7-9,2008, Kuala
Lumpur, Malaysia.© Faculty of Electrical Engineering,
UiTM Shah Alam, Malaysia. ISBN: 978-983-42747-9-5.
[3] K. Sulovská, S. Bělašková, M. Adámek: “Study of Face
Recognition Using Statistical Analysis,” International
Journal of Video&Image Processing and Network
Security IJVIPNS-IJENS Vol:13 No:01.
[4] Komal Vij, Yaduvir singh: “Enhancement of Images
Using Histogram Processing Techniques,” Int. J. Comp.
Tech. Appl., Vol 2 (2), 309-313, ISSN: 2229-6093. [4] J.
Nagi, S. K. Ahmed: “A MATLAB based Face Recognition
System using Image Processing and Neural Networks,”
4th International Colloquium on Signal Processingand
its Applications, March 7-9, 2008, Kuala Lumpur,
Malaysia.© Faculty of Electrical Engineering, UiTM
Shah Alam, Malaysia. ISBN: 978-983-42747-9-5.
[5] Md. Zamilur Rahman, Mohammad Shameemmhossain
Kawsar Ahmed: “Flag Identification Using Support
Vector Machine,” JU Journal of Information Technology
(JIT), Vol. 2, June 2013 11.
[6] Mohit P. Gawande, Prof. Dhiraj G. Agrawal: “Face
recognition using PCA and different distance
classifiers”, IOSR Journal of Electronics and
Communication Engineering (IOSR-JECE) e-ISSN:
2278-2834, p- ISSN: 2278-8735.Volume 9, Issue1, Ver.
VI (Feb. 2014), PP 01-05 www.iosrjournals.org. [3]
Komal Vij, Yaduvir singh: “Enhancement of Images
Using Histogram Processing Techniques,” Int. J. Comp.
Tech. Appl., Vol 2 (2), 309-313, ISSN: 2229-6093.
[7] Qiu Chen, Koji Kotani, Feifei Lee, and Tadahiro Ohmi:
“Face Recognition Using Histogram-based Features in
Spatial and Frequency Domains,” copyright (c) IARIA,
2014. ISBN: 978-1-61208-320-9.
[8] S. P. P. Thwe, N. A. A. Htwe: “Implementation of Face
Recognition System Using Principal Component
Analysis and Euclidean Distance,”
[9] Taranpreet Singh Ruprah: “Face Recognition Based on
PCA Algorithm,” Special Issue of International Journal
of Computer Science and Informatics (IJCSI), ISSN
(PRINT): 2231-5292, Vol.-II, Issue- 1, 2.
[10] Y. Tint, “ Steganalysis for MP3Stego using Independent
Component Analysis” Proceedings of 2013
International Conference on Information and
Communication Technology for Education(ICTE 2013
V)](https://image.slidesharecdn.com/304nationalflagsrecognitionbasedonprincipalcomponentanalysis-190913085810/85/National-Flags-Recognition-Based-on-Principal-Component-Analysis-5-320.jpg)
National Flags Recognition Based on Principal Component Analysis
- 1. International Journal of Trend in Scientific Research and Development (IJTSRD) Volume 3 Issue 5, August 2019 Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470 @ IJTSRD | Unique Paper ID – IJTSRD26775 | Volume – 3 | Issue – 5 | July - August 2019 Page 1610 National Flags Recognition Based on Principal Component Analysis Soe Moe Myint, Moe Moe Myint, Aye Aye Cho University of Computer Studies, Pathein, Myanmar How to cite this paper: Soe Moe Myint | Moe Moe Myint | Aye Aye Cho "National Flags Recognition Based on Principal Component Analysis" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456- 6470, Volume-3 | Issue-5, August 2019,pp.1610-1614, https://doi.org/10.31142/ijtsrd26775 Copyright © 2019 by author(s) and International Journalof Trendin Scientific Research and Development Journal. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (CC BY 4.0) (http://creativecommons.org/licenses/by /4.0) ABSTRACT Recognizing an unknown flag in a scene is challenging due to the diversity of the data and to the complexity of the identification process. And flags are associated with geographical regions, countries and nations. But flag identification of different countries is a challenging and difficult task. Recognition of an unknown flag image in a scene is challenging due to the diversity of the data and to the complexity of the identification process. The aim of the study is to propose a feature extraction based recognition system for Myanmar’s national flag. Image features are acquired from the region and state of flags which are identifiedbyusingprincipalcomponentanalysis(PCA). PCA is a statistical approach used for reducing the number of features in National flags recognition system. KEYWORDS: PCA, texture feature 1. INTRODUCTION Flag recognition systems are part of image processing applications and their significance as a research areas are increasing recently. J. Nagi, S. K. Ahmed [2] proposed a new technique for human facerecognition thatusesan image-based approach towards artificial intelligence by removing redundant data from face images through image compression using the two-dimensional discrete cosine transform (2D-DCT). Feature vectors are constructed by computing DCT coefficients. A self-organizing map (SOM) usinganunsupervised learning techniques used to classify DCT-based feature vectors into groups to identify if the subject in theinputimageis“present or “not present” in the image database. Md. Zamilur Rahman and Mohammad Shameemmhossain Kawsar Ahmed [5] proposed an approach based on support vector machine (SVM). This machine is trained with the percentage of different colors in the flag. Totrain uptheSVM have used a package LIBSVM which classify multiple classes and find out an accurate country id (name) as output. The system is designed to detect a flag more accurately and implement a system that is more efficient in this problem domain by reducing the time for flag detection andthevisual recognition for man. K. Sulovská, S. Bělašková, M. Adámek [3] proposed the analytical-statistical method for the face recognition. Although the bases of the face recognition are known by researchers worldwide, the statistical tests of data obtained by measuring chosen anthropometrical points can be found in several articles. Their aim was to show how the data act during the various emotions of one face, which will be helpful for deeper knowledge of how the face behaves. Acquired results reflect the difficulty of describing the face and the applicability of combination of different recognition methods (e.g. methods based on neural networks, recognition of facial contours, distribution of the gray scale in the image, deformation models) to get the best results in the verification/identification of a human. The aim of this study is to propose flag recognition based on image featuresextractionwith principal componentanalysis. The proposed system overcomes certain limitation of the existing recognition system. It is based on extracting the texture features of a set of Myanmar’snationalflagsstored in the database and performing mathematical operations on the values corresponding to them. The proposed system is better mainly due to the use of image features rather than the entire flag. Their advantages are in terms of: recognition accuracy and better discriminatory power computational cost because smaller images require less processing to train the PCA. Because of the use of dominant features and hence can be used as an effective means of authentication. This paper is composed as follows: Methodology of the system is described in section two. In section three, Data Acquisition and in section four includes Result and Discussion. Finally, the paper has been concluded. 2. Methodology In this section, related methodologies of the proposed system which are discussed with two parts. The first part will explain the texture features in digital image processing. The second part will describe the procedure of principal component analysis forrecognizinganddetectingofnational flags. The block diagram of the proposed flag recognition methods is given in Figure1. IJTSRD26775
- 2. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470 @ IJTSRD | Unique Paper ID – IJTSRD26775 | Volume – 3 | Issue – 5 | July - August 2019 Page 1611 Figure1. Flag Recognition system The proposed flag recognition system is shown in Figure1, the first step need to extract texture features from the training and testing flag images. In thesecond step, extracted features from the dataset and testing flag image are recognized by using principal component analysis for dimension reduction. 2.1. Feature Extraction The feature based methods do not directly work with image intensity values, but use salient features extracted from two images, which has been shown to be more suitable for such situation that intensity changes and complicated geometric deformations are encountered. Texture is a key component of human visual perception. Everyone can recognize texture but, it is more difficult to define. Texture has qualities such as periodicity and scale; it can be described in terms of direction, coarseness, contrast and so on. In this system chose three very different approaches to computing texture features: the first takes a statistical approach in the form of co-occurrence matrices, next the psychological view of Tamura’s features and finally signal processing with Gabor wavelets. 2.1.1. Texture feature Texture is a main component of human visual perception. Everyone can determine texture but it is more difficult to define. Table 1 shows the normalized co-occurrence matrix. Table1: Features Calculated from the Normalized Co- occurrence Matrix P(i, j). Texture Features Formula Energy ∑i∑j P2(i,j) Entropy ∑i∑j P(i,j)log P(i,j) Contrast ∑i∑j (i – j)2P(i,j) Homogeneity ∑i∑j (P(i,j)/1 + |i – j| where P = co-occurrence matrix, µ = mean of the co- occurrence matrix P, σ = standard variation of co-occurrence matrix P. Co-occurrence matrices Haralick [1] suggested the use of grey level co-occurrence matrices (GLCM) to extract second order statistics from an image. Tamura Tamura et al took the approach of devising texture features that correspond to human visual perception [5]. The first three attained very successful results and are used in our evaluation. Gabor One of the most popular signalprocessing-based approaches for texture feature extraction has been the use of Gabor filters. This system implementation is based on that of Manjunath et al. [3, 6]. Gabor wavelet transform is then defined to be: ( ) ( ) ( )* 1 1 1 1, , d dmn mnW x y I x y g xx y x y−∫ (1) The mean and standard deviationofthemagnitude|Wmn| are used to for the feature vector. Table2: First three Tamura features. Tamura Formula Coarseness Contrast Directionality magnitude of the vector, angle of the vector, 2.2. Principal Component Analysis Principal Components Analysis (PCA) is a dimensionality reduction algorithm that can be used to significantly speed up your unsupervised feature learning algorithm. More importantly, understanding PCA will enable us to later implement whitening,which isan importantpre-processing step for many algorithms [10]. PCA is mostlyused as atool in exploratory data analysis and for making predictive models. It is often used to visualize genetic distance and relatedness between populations [9]. Standard Principle ComponentAnalysis(PCA) isoftenuseful preprocessing strategy in ICA is to first whiten the observed variables. This transform observed vector linearly so that obtain a new vector which is white. Its component are uncorrelated and their variances equal unity. (2) One popular method for whitening is to use the eigenvalue decomposition (EVD) of the covariance matrix , where E is orthogonal matrix of eigenvectors of and D is the diagonal matrix of its eigenvalues ).Notethat canbe estimated in a standard way from the variable sample whitening can now be done by (3) For example, rank of D is equal to two for image features, meaning that observed and training features are uncorrelated. On the other hand, if the flag is not recognized, this mixtures are actually the combination of one features only, hence, the rank of D will be reduced to one. 3. Data Acquisition Flag recognition system is composed of two stages: Data Acquisition and Result and Discussion that consists of two parts: Feature Extraction and Flag Recognition with PCA.
- 3. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470 @ IJTSRD | Unique Paper ID – IJTSRD26775 | Volume – 3 | Issue – 5 | July - August 2019 Page 1612 Firstly, an input flag image is required to perform image processing techniques such as resizing, changing resolution and cropping. Table 3 shows the list of different type for flag images. Data are acquired from the 14 National flag of Myanmar. These are represented of Myanmar’s 7 states and 7 regions. Table3. Myanmar’s national flags dataset Division and States Total images (Resizing, converting resolution, different type of cropping Kachin State 20 Kayah State 20 Kayin State 20 Chin State 20 Mon State 20 Rakhine State 20 Shan State 20 Yangon Region 20 Mandalay Region 20 Sagaing Region 20 Magway Region 20 Bago Region 20 Ayeyarwaddy Region 20 Tanintharyi Region 20 Table4. Sample Dataset for Ayeyarwaddy Region Flag Original image Resizing image Changing resolution image Cropping image The input original flag images are cropped, resized about 20 different portions and resizinginthe varietyrangeof pixels(1024 x 768 px), (800 x 600 px), (640 x 480 px), (448 x 336 px) and (314 x 235 px) for flag images. In this experiment, MATLAB is used to crop, resize and change resolution of the original image. Table4 and table5 are shown the sample dataset of Ayeyarwaddy Region and Yangon Region flags. Table5. Sample Dataset for Yangon Region Flag Original image Resizing image Changing resolution image Cropping image 4. Result And Discussion 4.1. Results of Feature Extraction Extraction of features is made from flag images by using histogram and probability density. If all features in the feature vector were statistically independent, one could simply eliminate the least discriminative features from this vector. Many features depend on each other or on an underlying unknown variable. A single feature could therefore represent a combination of multiple types of information by a single value. Table6 shows sample features fordifferenttypesofAyeyarwaddyDivision flag.
- 4. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470 @ IJTSRD | Unique Paper ID – IJTSRD26775 | Volume – 3 | Issue – 5 | July - August 2019 Page 1613 Table6. Sample features for different types of flag Flag types Features aye-pdf1 aye-pdf2 aye-pdf3 aye-pdf4 341593 303510 310097 301693 342012 303977 310531 302058 342462 304453 311012 302510 342647 304635 311207 302694 343229 305193 311372 303274 aye-reso-pdf1 aye-reso-pdf2 aye-reso-pdf3 aye-reso-pdf4 83442 63077 65407 62749 86337 65541 68043 65192 87587 66685 69237 66342 88554 67408 70033 67094 89443 68123 70804 67806 aye-resize-pdf1 aye-resize-pdf2 aye-resize-pdf3 aye-resize-pdf4 16119 9511 9476 9501 16306 9655 9651 9643 16419 9726 9737 9721 16512 9783 9798 9785 16618 9849 9883 9862 aye-crop-pdf1 aye-crop-pdf2 aye-crop-pdf3 aye-crop-pdf4 115301 102976 109399 100996 115631 103362 109744 101276 115928 103684 110074 101581 116065 103821 110223 101719 116596 104330 110339 102249 4.2. PCA based flag recognition system The paper has presented a flag recognition system using PCA in the context of flag verification and flag recognition using photometric normalization for comparison. This system trained above 100 images in many kinds of flag. PCA is very high dimensional nature of many data sets makes direct visualization impossible as we humans can only comprehend three dimensions. The solution is to work with data dimension reduction techniques. When reducing the dimensions of data, it’s important not to lose more information than is necessary. The variation in a data set can be seen as representing the information that we would like to keep. Figure 2 shows the testing results of flag recognition system. Figure2. Testing results of flag recognition system
- 5. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470 @ IJTSRD | Unique Paper ID – IJTSRD26775 | Volume – 3 | Issue – 5 | July - August 2019 Page 1614 Principal Component Analysis (PCA) is a well-established mathematical technique for reducing the dimensionality of data, while keeping as much variation as possible. PCA achieves dimension reduction by creating new, artificial variables called principal components. Each principal component is a linear combination of the observed variables. According to the results, features extraction based flag recognition system can recognize the correct flag image for testing data. In this experiment, receiver operating characteristic (ROC) curve has been used to verify the effectiveness of the proposed method. Figure3gives the ROC curves as the flag recognition results. It can be seen that the variation of the performance function for training and testing. Figure3. ROC curve under different type of flag image 5. Conclusion National flag recognition system is implemented by using texture features and principal component analysis. The computer system can automatically recognized all national flags of Myanmar via loaded from camera, web cameras or natural scenes. Applying texture based features extraction techniques are combined to use with PCA, which reported better accuracy for flags recognition results. The eigenvector for PCA approach thus provided a practical solution that is well fitted to the problem of flag recognition system. PCA achieves dimension reduction by creating new, artificial variables called principal components. Each principal component is a linear combination of theobserved variables. Recognition accuracy of the proposed system from 90 % to 95 % depending on weather the flag is among 14 state and region. REFERENCE [1] G. N. Srinivasan, and Shobha G: “Statistical Texture Analysis,” PROCEEDINGS OF WORLD ACADEMY OF SCIENCE, ENGINEERING AND TECHNOLOGY VOLUME 36 DECEMBER 2008 ISSN 2070-3740. [2] J. Nagi, S. K. Ahmed: “A MATLAB based Face Recognition SystemusingImageProcessingand Neural Networks,” 4th International Colloquium on Signal Processing and its Applications, March 7-9,2008, Kuala Lumpur, Malaysia.© Faculty of Electrical Engineering, UiTM Shah Alam, Malaysia. ISBN: 978-983-42747-9-5. [3] K. Sulovská, S. Bělašková, M. Adámek: “Study of Face Recognition Using Statistical Analysis,” International Journal of Video&Image Processing and Network Security IJVIPNS-IJENS Vol:13 No:01. [4] Komal Vij, Yaduvir singh: “Enhancement of Images Using Histogram Processing Techniques,” Int. J. Comp. Tech. Appl., Vol 2 (2), 309-313, ISSN: 2229-6093. [4] J. Nagi, S. K. Ahmed: “A MATLAB based Face Recognition System using Image Processing and Neural Networks,” 4th International Colloquium on Signal Processingand its Applications, March 7-9, 2008, Kuala Lumpur, Malaysia.© Faculty of Electrical Engineering, UiTM Shah Alam, Malaysia. ISBN: 978-983-42747-9-5. [5] Md. Zamilur Rahman, Mohammad Shameemmhossain Kawsar Ahmed: “Flag Identification Using Support Vector Machine,” JU Journal of Information Technology (JIT), Vol. 2, June 2013 11. [6] Mohit P. Gawande, Prof. Dhiraj G. Agrawal: “Face recognition using PCA and different distance classifiers”, IOSR Journal of Electronics and Communication Engineering (IOSR-JECE) e-ISSN: 2278-2834, p- ISSN: 2278-8735.Volume 9, Issue1, Ver. VI (Feb. 2014), PP 01-05 www.iosrjournals.org. [3] Komal Vij, Yaduvir singh: “Enhancement of Images Using Histogram Processing Techniques,” Int. J. Comp. Tech. Appl., Vol 2 (2), 309-313, ISSN: 2229-6093. [7] Qiu Chen, Koji Kotani, Feifei Lee, and Tadahiro Ohmi: “Face Recognition Using Histogram-based Features in Spatial and Frequency Domains,” copyright (c) IARIA, 2014. ISBN: 978-1-61208-320-9. [8] S. P. P. Thwe, N. A. A. Htwe: “Implementation of Face Recognition System Using Principal Component Analysis and Euclidean Distance,” [9] Taranpreet Singh Ruprah: “Face Recognition Based on PCA Algorithm,” Special Issue of International Journal of Computer Science and Informatics (IJCSI), ISSN (PRINT): 2231-5292, Vol.-II, Issue- 1, 2. [10] Y. Tint, “ Steganalysis for MP3Stego using Independent Component Analysis” Proceedings of 2013 International Conference on Information and Communication Technology for Education(ICTE 2013 V)