Skip to main content

Arif Muntasa

Followers

0

Following

11

Public Views

Johan Edwart L. Hutabarat

Institut Teknologi Nasional Yogyakarta (ITNY)

IPG KRM

Universitas Jambi

Armando Marques-Guedes

UNL - New University of Lisbon

Gwen Robbins Schug

University of North Carolina at Greensboro

Gabriel Gutierrez-Alonso

University of Salamanca

Macquarie University

Swansea University

Interests

Uploads

Papers by Arif Muntasa

Cross-Language Tourism News Retrieval System Using Google Translate API on SEBI Search Engine

Elinvo (Electronics, Informatics, and Vocational Education)

Cross-Language Information Retrieval (CLIR) is responsible for retrieving information stored in a... more Cross-Language Information Retrieval (CLIR) is responsible for retrieving information stored in a language different from the language of the query provided by the user. Some translation methods commonly used in CLIR are Dictionary, Parallel corpora, Comparable corpora, Machine translator, Ontology, and Transitive-based. The query must be translated to the target language, followed by preprocessing and calculating the similarity between the query and all documents in the corpus. The problem is the time and accuracy of query translation. Moreover, the queries are not written as complete sentences according to certain language rules. Stemming, for example, every language has its own method. Indonesian has basic words and affixes in the form of prefixes, suffixes, infixes, and confixes, while English only has suffixes. Stemming takes a long time in text processing. In the Indonesian search engine (SEBI), the provision of cross-language tourism news retrieval is realized using the Googl...

Adaptive Boosting Classifier for Pedestrian Attributes Identification with Color and Texture Features

2018 International Conference on Computer Engineering, Network and Intelligent Multimedia (CENIM)

Combination of color and texture features along with adaptive boosting classifier is proposed in ... more Combination of color and texture features along with adaptive boosting classifier is proposed in this paper for attributes identification of the pedestrian images. Color histogram and Multi Block Local Binary Pattern (MB-LBP) is used as a feature extraction method, since both methods do not require the same size of pedestrian images. This is suitable with the dataset that is trained and tested in this research, PETA dataset since this dataset has various size, a point of view, resolution, etc. Color histogram and MB-LBP capture two kinds of information in the pedestrian images, i.e., color and texture information. The Adaboost is used as a classifier for the attribute identification, and for the weak classifier, the decision stump is applied. The performance accuracy of texture, color, and the combination of both features, are compared in the attribute identification process. In the conducted experiments, the number of bins for the color histogram, and MBLBP are 8, 16, and 32. The results showed that the combination of both features, color histogram and MB-LBP, achieved the highest accuracy.

Studi Litelatur: Peningkatan Kinerja Digital Forensik Dan Pencegahan Cyber Crime

Jurnal Aplikasi Teknologi Informasi dan Manajemen (JATIM)

The more complex a problem that arises due to an ease in information technology makes cyberattack... more The more complex a problem that arises due to an ease in information technology makes cyberattacks increasingly dangerous and dangerous for systems within a country or entity. Crimes thatarise must be detected and a preventive action taken immediately so that no party is harmed for atechnological crime. This study aims to find various methods to review and investigate cyberattacks for security in digital forensics and can provide input and additional insight to all partiesincluding investigators or practitioners and users in the field of digital forensics about variousmethods that can be used to detect and prevent cyber attacks. This research was conducted byconducting a literature study of various studies on cyber attacks. The research method used isdescriptive qualitative method by taking data from literacy studies on research on cyber attacks.The results of the literature study show that the factors that can improve the performance of digitalforensics are the Digital Forensics Fr...

Feature selection with Genetic Algorithm for alcoholic detection using electroencephalogram

2017 International Conference on Sustainable Information Engineering and Technology (SIET)

Electroencephalography is a technique of recording human brain waves through electrodes mounted o... more Electroencephalography is a technique of recording human brain waves through electrodes mounted on the human scalp. The obtained signal data or Electroencephalogram (EEG) can be used for the alcoholic detection as a substitution of the urine test or breathalyzer test, since in the urine or breathalyzer test, is time limited. Therefore, the usage of EEG for alcoholic detection is proposed in this research. There are four stages are proposed in this research. First, noise removal of the recorded signal by separating the signal using Independent Component Analysis. Second, feature extraction with Discrete Wavelet Transform. Third, extracted features from the previous stage are selected based on the importance of each feature using Genetic Algorithm, and final stage is the alcoholic detection with Backpropagation neural network. Experiments were conducted on 64 channels of EEG data. The feature selection stage using Genetic Algorithm made the accuracy of the detection process is increased, i.e. the maximum average accuracy of the detection is 79.38%. This feature selection stage also reduced the used feature for the detection process, up to 48% of all obtained features, in the extraction feature stage.

Actor-Network Theory (ANT) for Indonesia e-Government implementation

2020 6th Information Technology International Seminar (ITIS), 2020

An Actor-Network Theory (ANT) is a well-known theory in the information system area. This approac... more An Actor-Network Theory (ANT) is a well-known theory in the information system area. This approach is employed to capture actors, heterogeneous networks, and power relations. In Indonesia, there is a presidential regulation number 95 of 2018 about electronic government system. This regulation set all of the e-government implementation in Indonesia, such as governance, management, information technology audit, the organizer, acceleration, monitoring, and evaluation. Many e-government research using ANT for a theoretical lens, however, there is limited research in Indonesia districts. Therefore, this research aims to apply ANT in eGovernment case studies in two districts in Indonesia. The first case study is conducted at a district of East Java in 2018 and the second one in a district of Lampung in 2019. Both cases have different social, economic, cultural, and political contexts, even though in one country, Indonesia. The first case study is an established district and another one is the new district. This research captures the global and local network actors as well as classifies their power over the e-government implementation in both cases. The results show that different contexts produce different results and achievements in e-Government implementation. The dominant factor in the e-Government success factor is the commitment of the district head. It has an impact on the strategy to implement e-Government.

Invited Speech 1: A New Model To Classify The Acute Lymphoblastic Leukemia Image

2019 2nd International Conference of Computer and Informatics Engineering (IC2IE), 2019

Acute lymphoblastic leukemia is a type of cancer which attacks blood cells and spinal cord. A pat... more Acute lymphoblastic leukemia is a type of cancer which attacks blood cells and spinal cord. A patient needs to conduct early detection to avoid more severe conditions. In this research, we proposed a new model to detect acute lymphoblastic leukemia by microscopic images. We divide fives stages to classify acute lymphoblastic leukemia, i.e. An image enhancement, an image segmentation, noise removal, feature extraction, and classification. Firstly, we enhanced the green channel before the segmentation process. Secondly, we employed the entropy method to obtain the best threshold value. However, the segmentation results have delivered small regions besides the main object. Thirdly, we have to remove the small objects before feature extraction. In addition, We extract the main object using energy, entropy, shanon entropy, and the object circularity. Lastly, we modified a learning vector quantization. The proposed algorithm synchronized the weight value, which has the maximum probability. We have evaluated our proposed algorithm using the Acute Lymphoblastic Leukemia Image Database (ALL-IDB2). Our proposed method has delivered 96.15% maximum accuracy. It proved that the proposed method has outperformed to the other methods, i.e. leukemia detection using Fuzzy, perceptron, and support vector machine

A New Modeling of the Landmark Movement Based on the Previous Movement Results to Detect the Facial Sketch Features

Deformable model is one of interest topic in biometrics. The deformable model applications have b... more Deformable model is one of interest topic in biometrics. The deformable model applications have been used to detect objects included facial sketch feature images. Many methods have been discovered and developed to detect objects of deformable. However, any method can solve a problem, but raises other issues. Popular method to solve the deformable object is Active Shape Model. It is the most used to detect facial feature, but it has weakness. This method is very depended on the shape initialization. If the movement direction is not covered by training sets, then the landmark movement will be move to the undesirable direction. In this research, we proposed method to create a new modeling of the landmark movement based on the previous movement results. Four algorithms are used to improve the landmark movement when the detection process. The experimental results show that our proposed method produces the average of detection accuracies more than 90% for all scenarios.

Line Detection Model and Adaptive Threshold Based Image Segmentation For Handwriting Word Recognition

Handwriting word recognition has been researched many researchers. The most method used is Line b... more Handwriting word recognition has been researched many researchers. The most method used is Line based representation. However, it has a weakness, which is high cost to recognize object. In this research, line detection model is proposed to determine the right, left, top and bottom object boundary. In order to separate all of objects, the line detection has been conducted for segmentation iteratively. All of objects are labeled to get number of object in image. One of parameters that have effect to the segmentation results is threshold value. Errors in the determination of the threshold will affect to the segmentation results. However, it is necessary to determine the adaptive threshold value. In this research, Otsu’s method is proposed to achieve the best threshold value. The threshold value depends on the testing image. 84 images have been used as training set. Proposed method has been tested by using 30 images. In this case, 20 images have single line handwriting word and the 10 i...

Identification of Pedestrians Attributes Based on Multi-Class Multi-Label Classification using Convolutional Neural Network (CNN)

The usage of computer vision in identifying pedestrians attributes has received a great attention... more The usage of computer vision in identifying pedestrians attributes has received a great attention, especially in the visual surveillance systems. For instance, searching for system based on the attributes. Attributes Identification using Convolutional Neural Network architecture is presented in this article, since the architecture can perform feature learning. CNN consist of convolution layer, ReLU, Pooling, and Fully-connected. There are three experiment scenarios are conducted based on the number of convolution layers, to determine the effect of layers on CNN performance. Three different CNN architectures were trained and tested using a PETA dataset with 35 attributes. The highest accuracy achieved is 75.66% based on number of convolutional layers. The conducted experiments showed that more numbers of convolution layers used would produce the better CNN's performance.

E-government learning media through augmented reality technology

Bulletin of Social Informatics Theory and Application, 2020

Augmented reality (AR) is one of the newest learning technologies besides Artificial Intelligent,... more Augmented reality (AR) is one of the newest learning technologies besides Artificial Intelligent, Virtual Reality, Virtual World, Alternates Reality Games, and Wearable Computing. E-Government is becoming an evolving academic field and transforming to be E-Governance, E-Participation, E-Service, E-Democracy, E-Voting issues. Therefore, some universities have E-Government course in their curriculum. AR can be one of the learning media to learn and teach E-Government effectively and interestingly. This research aims and contributes to develop a novel model of E-Government learning media using AR. Furthermore, this work has three novelties as follows:First, a method to develop e-Government learning media using AR. Second, a novel model of e-government learning media based on augmented reality technology. Third, the material courses are command centre and public service mall materials based in Surabaya, Indonesia. The new model of learning media using AR consists of main components, i....

A Conceptual Model of Indonesian Question Answering System based on Semantic Web

Journal of Physics: Conference Series, 2020

This study aims to develop a conceptual model of question answering system in Bahasa using Semant... more This study aims to develop a conceptual model of question answering system in Bahasa using Semantic web technology. A trigger was built to periodically collect data scattered on the internet in a data center to shorten the query process on data without reducing the data completeness. The test case domain was automotive. This study used four types of ontologies: main ontology, ontologies from manufacturers, dealers, and rentals. This research has some novelties, such as solving the weakness of distributed and centralized data and combining the advantages of both by storing distributed data into a pool (centralized data) periodically, then querying the pool. Furthermore, it makes contribution by providing a conceptual model of Indonesian question answering system, particularly in automotive domain. This has some implications, such as extend the body of knowledge of computer science, especially semantic web and ontology fields.

Classification of EEG signals uses the coefficient of wavelet transform and K-nearest neighbor

IOP Conference Series: Materials Science and Engineering, 2020

The research that was built was used to explain the application of Electroencephalography (EEG) s... more The research that was built was used to explain the application of Electroencephalography (EEG) signal waves. EEG data is used to move the cursor up and the cursor down. Characterization of each EEG signal uses the Wavelet method taken at each the subband of the wavelet process. Wavelet process by taking the average value and the standard deviation value of the wavelet coefficient. The average value and the standard deviation value is used as an EEG feature. K-Nearest Neighbor is used as an identification whether the cursor will move up or vice versa. This study uses of 100 EEG signal data consisting of 50 test data and 50 testing data. The accuracy of identification uses the 80% K-NN method.

A novel framework of e-participation for smart cities

Bulletin of Social Informatics Theory and Application, 2019

The smart city is an exciting concept for improving the quality of the city. However, a smart cit... more The smart city is an exciting concept for improving the quality of the city. However, a smart city needs participation from citizens and all related stakeholders to use the technologies effectively in order to achieve the goal and solve the problems. Even though the city already has high-tech infrastructures, participation still required to provide ideas, inputs, and roles within the development of smart city concepts. This research aims to describe and analyze the current state of the e-Participation frameworks and propose a novel framework for smart cities. The analysis will go through a systematic literature review. Hopefully, this study makes contributions by providing a novel framework of e-Participation for smart cities. This research has implications for theory and practice. For theory, the novel framework can be added to the body of knowledge of e-participation, e-government, and smart cities fields. For practice, the framework will be useful for practitioners, policymakers,...

Extraction of ElectroEncephaloGraph (EEG) Signal Using the Subband Coefficient of Wavelet Transform on Cursor Moves

IOP Conference Series: Materials Science and Engineering, 2019

This research explains the application of ElectroEncephaloGraph (EEG) signal waves used to move t... more This research explains the application of ElectroEncephaloGraph (EEG) signal waves used to move the up cursor and down cursor. In each sub band of the waveform, Electro Encephalo Graph (EEG) will produce the average and standard deviation to be used as a feature of the EEG. Artificial Neural Network Backpropagation as the basis for determining whether the cursor moves up or the cursor moves down. The data used in this study is EEG data derived from BCI Competition 2003 (BCI Competition 2003). Decision-making is done in two stages. In the first stage, the mean and standard deviation values on each wavelet subband as a feature extraction of EEG data. This feature is an input to the Backpropagation Neural Network. In the second stage of the process of inserting into two classes (class 0 and class 1) EEG data files, there are 260 EEG and 293 file training data files from EEG file test data files, totaling to 553 EEG data files. The results obtained for the classification of EEG results ...

The Dimensionality Reduction Based on the Average of the Left and the Right Structure Preserving Projection

International Review on Computers and Software (IRECOS), 2016

The large dimensionalities are the crucial problem in computer vision, especially in biometrics f... more The large dimensionalities are the crucial problem in computer vision, especially in biometrics field. In this research, a new method to reduce the dimensionality based on the left and right preserving projection is proposed. The proposed approach transforms the image input to integrate the row and column information. The image data transformation is employed through the shifting to the left and the right. The shifting results, for both the left and right are computed showing the average values needed to realize the dominant features. In this research, the proposed approach uses the Eigenfaces process calculation to avoid the singularity problem. The projection results of the Eigenfaces to the average value of the calculation results are employed to further process on the covariance of the left and the right Structure Preserving Projection. The Eigenvalue problem solution is applied to determine the Eigenvectors associated to the largest Eigenvalues. The results of the experiments demonstrated that the maximum acceptance rates are 100% for the ORL, 98.78% for the YALE, and 96.52% for the UoB face database. The best average acceptance rates achieved are 95.3%, 97.94%, and 95.21% for the ORL, YALE, and UoB face database respectively. The results have also exhibited that the proposal has outperformed the other appearance approaches.

Windowing System Facial Detection based on Gabor Kernel Filter, Fast Fourier Transform, and Probabilistic Learning Vector Quantization

International Journal of Technology, 2017

Facial detection is a crucial stage in the facial recognition process. Misclassification during t... more Facial detection is a crucial stage in the facial recognition process. Misclassification during the facial detection process will impact recognition results. In this research, windowing system facial detection using the Gabor kernel filter and the fast Fourier transform was proposed. The training set images, for both facial and non-facial images, were processed to obtain the local features by using the Gabor kernel filter and the fast Fourier transform. The local features were measured using probabilistic learning vector quantization. In this process, facial and non-facial features were classified using label 1 and -1. The proposed method was evaluated using facial and non-facial image testing sets, which were taken from the MIT+CMU image database. The testing images were enhanced first before the detection process using four different enhancement methods: histogram equalization, adaptive histogram equalization, contrast limited adaptive histogram equalization, and the single-scale retinex method. The detection results demonstrated that the highest average accuracy was 83.44%.

Three-directional of the ID-diagonal fisherface modeling based feature extraction

2016 International Seminar on Intelligent Technology and Its Applications (ISITIA), 2016

Three directional of the 1D diagonal fisherface for feature extraction is proposed in this resear... more Three directional of the 1D diagonal fisherface for feature extraction is proposed in this research for the face recognition system. Three directional of the diagonal matrix is built in the proposed method, i.e. the input face image is transformed into a diagonal matrix. Therefore, another form of the input image is created for the feature extraction using the diagonal matrix of the input image. The diagonal matrix is then transformed into 1D matrix for each training set. The feature extraction used this matrix to create the projection and weight matrix using Fisher Face. Two until ten features are extracted using the proposed method. There are four distance methods are used in the experiments, i.e. Euclidian, Manhattan, Canberra, and Chebychev. The experiments of the proposed system are conducted on three databases, i.e. ORL, University of Bern, and YALE-A databases. The recognition accuracy of the proposed model for face recognition achieved up to 92% accuracy.

A new approach: the local feature extraction based on the new regulation of the locally preserving projection

Applied Mathematical Sciences, 2015

A novelty of the local feature extraction was proposed for face recognition. To optimize the Eige... more A novelty of the local feature extraction was proposed for face recognition. To optimize the Eigenvalue and Eigenvector, a new regulation has been embedded to the Locally Preserving Projection. The proposed method has reduced computation time to obtain the new subspace of the original of the Locally Preserving Projection. The proposed method has also produced orthogonal basis function matrix. However, orthogonal basis function matrix can reconstruct easier than non-orthogonal function. The proposed method has been evaluated by using three face image databases, they are the University of Bern, the YALE-A, and the ORL face image databases. The experimental results of the proposed method have produced the recognition rate 96% for the University of Bern, 96.19% for the YALE-A and 98% for the ORL facial image databases. The experimental results of the proposed method have produced higher recognition rate than the Principal Component Analysis, the Linear Discriminant Analysis and the Locally Preserving Projection.

Appearance Global and Local Structure Fusion for Face Image Recognition

TELKOMNIKA (Telecommunication Computing Electronics and Control), 2011

Analisis komponen utama (PCA) dan analisis deskriminan linear (LDA) merupakan metode ekstraksi be... more Analisis komponen utama (PCA) dan analisis deskriminan linear (LDA) merupakan metode ekstraksi berbasis penampakan yang menghasilkan fitur-fitur dengan struktur global. Fitur-fitur dengan struktur global mempunyai kelemahan, yaitu fitur-fitur dengan struktur lokal tidak dapat dicirikan. Proyeksi pelestarian lokalitas (LPP) dan wajah-Laplacian orthogonal (OLF) merupakan metode ekstraksi model penampakan yang menghasilkan fitur-fitur dengan struktur lokal, namun fitur struktur global diabaikan.

Enhancement of the Adaptive Shape Variants Average Values by Using Eight Movement Directions for Multi-Features Detection of Facial Sketch with Tilt Angle 15°

ITB Journal of Information and Communication Technology, 2012

This paper aims to detect multi features of a facial sketch by using a novel approach. The detect... more This paper aims to detect multi features of a facial sketch by using a novel approach. The detection of multi features of facial sketch has been conducted by several researchers, but they mainly considered frontal face sketches as object samples. In fact, the detection of multi features of facial sketch with certain angle is very important to assist police for describing the criminal's face, when criminal's face only appears on certain angle. Integration of the maximum line gradient value enhancement and the level set methods was implemented to detect facial features sketches with tilt angle to 15 degrees. However, these methods tend to move towards non features when there are a lot of graffiti around the shape. To overcome this weakness, the author proposes a novel approach to move the shape by adding a parameter to control the movement based on enhancement of the adaptive shape variants average values with 8 movement directions. The experimental results show that the proposed method can improve the detection accuracy up to 92.74%.

Cross-Language Tourism News Retrieval System Using Google Translate API on SEBI Search Engine

Elinvo (Electronics, Informatics, and Vocational Education)

Cross-Language Information Retrieval (CLIR) is responsible for retrieving information stored in a... more Cross-Language Information Retrieval (CLIR) is responsible for retrieving information stored in a language different from the language of the query provided by the user. Some translation methods commonly used in CLIR are Dictionary, Parallel corpora, Comparable corpora, Machine translator, Ontology, and Transitive-based. The query must be translated to the target language, followed by preprocessing and calculating the similarity between the query and all documents in the corpus. The problem is the time and accuracy of query translation. Moreover, the queries are not written as complete sentences according to certain language rules. Stemming, for example, every language has its own method. Indonesian has basic words and affixes in the form of prefixes, suffixes, infixes, and confixes, while English only has suffixes. Stemming takes a long time in text processing. In the Indonesian search engine (SEBI), the provision of cross-language tourism news retrieval is realized using the Googl...

Adaptive Boosting Classifier for Pedestrian Attributes Identification with Color and Texture Features

2018 International Conference on Computer Engineering, Network and Intelligent Multimedia (CENIM)

Combination of color and texture features along with adaptive boosting classifier is proposed in ... more Combination of color and texture features along with adaptive boosting classifier is proposed in this paper for attributes identification of the pedestrian images. Color histogram and Multi Block Local Binary Pattern (MB-LBP) is used as a feature extraction method, since both methods do not require the same size of pedestrian images. This is suitable with the dataset that is trained and tested in this research, PETA dataset since this dataset has various size, a point of view, resolution, etc. Color histogram and MB-LBP capture two kinds of information in the pedestrian images, i.e., color and texture information. The Adaboost is used as a classifier for the attribute identification, and for the weak classifier, the decision stump is applied. The performance accuracy of texture, color, and the combination of both features, are compared in the attribute identification process. In the conducted experiments, the number of bins for the color histogram, and MBLBP are 8, 16, and 32. The results showed that the combination of both features, color histogram and MB-LBP, achieved the highest accuracy.

Studi Litelatur: Peningkatan Kinerja Digital Forensik Dan Pencegahan Cyber Crime

Jurnal Aplikasi Teknologi Informasi dan Manajemen (JATIM)

The more complex a problem that arises due to an ease in information technology makes cyberattack... more The more complex a problem that arises due to an ease in information technology makes cyberattacks increasingly dangerous and dangerous for systems within a country or entity. Crimes thatarise must be detected and a preventive action taken immediately so that no party is harmed for atechnological crime. This study aims to find various methods to review and investigate cyberattacks for security in digital forensics and can provide input and additional insight to all partiesincluding investigators or practitioners and users in the field of digital forensics about variousmethods that can be used to detect and prevent cyber attacks. This research was conducted byconducting a literature study of various studies on cyber attacks. The research method used isdescriptive qualitative method by taking data from literacy studies on research on cyber attacks.The results of the literature study show that the factors that can improve the performance of digitalforensics are the Digital Forensics Fr...

Feature selection with Genetic Algorithm for alcoholic detection using electroencephalogram

2017 International Conference on Sustainable Information Engineering and Technology (SIET)

Electroencephalography is a technique of recording human brain waves through electrodes mounted o... more Electroencephalography is a technique of recording human brain waves through electrodes mounted on the human scalp. The obtained signal data or Electroencephalogram (EEG) can be used for the alcoholic detection as a substitution of the urine test or breathalyzer test, since in the urine or breathalyzer test, is time limited. Therefore, the usage of EEG for alcoholic detection is proposed in this research. There are four stages are proposed in this research. First, noise removal of the recorded signal by separating the signal using Independent Component Analysis. Second, feature extraction with Discrete Wavelet Transform. Third, extracted features from the previous stage are selected based on the importance of each feature using Genetic Algorithm, and final stage is the alcoholic detection with Backpropagation neural network. Experiments were conducted on 64 channels of EEG data. The feature selection stage using Genetic Algorithm made the accuracy of the detection process is increased, i.e. the maximum average accuracy of the detection is 79.38%. This feature selection stage also reduced the used feature for the detection process, up to 48% of all obtained features, in the extraction feature stage.

Actor-Network Theory (ANT) for Indonesia e-Government implementation

2020 6th Information Technology International Seminar (ITIS), 2020

An Actor-Network Theory (ANT) is a well-known theory in the information system area. This approac... more An Actor-Network Theory (ANT) is a well-known theory in the information system area. This approach is employed to capture actors, heterogeneous networks, and power relations. In Indonesia, there is a presidential regulation number 95 of 2018 about electronic government system. This regulation set all of the e-government implementation in Indonesia, such as governance, management, information technology audit, the organizer, acceleration, monitoring, and evaluation. Many e-government research using ANT for a theoretical lens, however, there is limited research in Indonesia districts. Therefore, this research aims to apply ANT in eGovernment case studies in two districts in Indonesia. The first case study is conducted at a district of East Java in 2018 and the second one in a district of Lampung in 2019. Both cases have different social, economic, cultural, and political contexts, even though in one country, Indonesia. The first case study is an established district and another one is the new district. This research captures the global and local network actors as well as classifies their power over the e-government implementation in both cases. The results show that different contexts produce different results and achievements in e-Government implementation. The dominant factor in the e-Government success factor is the commitment of the district head. It has an impact on the strategy to implement e-Government.

Invited Speech 1: A New Model To Classify The Acute Lymphoblastic Leukemia Image

2019 2nd International Conference of Computer and Informatics Engineering (IC2IE), 2019

Acute lymphoblastic leukemia is a type of cancer which attacks blood cells and spinal cord. A pat... more Acute lymphoblastic leukemia is a type of cancer which attacks blood cells and spinal cord. A patient needs to conduct early detection to avoid more severe conditions. In this research, we proposed a new model to detect acute lymphoblastic leukemia by microscopic images. We divide fives stages to classify acute lymphoblastic leukemia, i.e. An image enhancement, an image segmentation, noise removal, feature extraction, and classification. Firstly, we enhanced the green channel before the segmentation process. Secondly, we employed the entropy method to obtain the best threshold value. However, the segmentation results have delivered small regions besides the main object. Thirdly, we have to remove the small objects before feature extraction. In addition, We extract the main object using energy, entropy, shanon entropy, and the object circularity. Lastly, we modified a learning vector quantization. The proposed algorithm synchronized the weight value, which has the maximum probability. We have evaluated our proposed algorithm using the Acute Lymphoblastic Leukemia Image Database (ALL-IDB2). Our proposed method has delivered 96.15% maximum accuracy. It proved that the proposed method has outperformed to the other methods, i.e. leukemia detection using Fuzzy, perceptron, and support vector machine

A New Modeling of the Landmark Movement Based on the Previous Movement Results to Detect the Facial Sketch Features

Deformable model is one of interest topic in biometrics. The deformable model applications have b... more Deformable model is one of interest topic in biometrics. The deformable model applications have been used to detect objects included facial sketch feature images. Many methods have been discovered and developed to detect objects of deformable. However, any method can solve a problem, but raises other issues. Popular method to solve the deformable object is Active Shape Model. It is the most used to detect facial feature, but it has weakness. This method is very depended on the shape initialization. If the movement direction is not covered by training sets, then the landmark movement will be move to the undesirable direction. In this research, we proposed method to create a new modeling of the landmark movement based on the previous movement results. Four algorithms are used to improve the landmark movement when the detection process. The experimental results show that our proposed method produces the average of detection accuracies more than 90% for all scenarios.

Line Detection Model and Adaptive Threshold Based Image Segmentation For Handwriting Word Recognition

Handwriting word recognition has been researched many researchers. The most method used is Line b... more Handwriting word recognition has been researched many researchers. The most method used is Line based representation. However, it has a weakness, which is high cost to recognize object. In this research, line detection model is proposed to determine the right, left, top and bottom object boundary. In order to separate all of objects, the line detection has been conducted for segmentation iteratively. All of objects are labeled to get number of object in image. One of parameters that have effect to the segmentation results is threshold value. Errors in the determination of the threshold will affect to the segmentation results. However, it is necessary to determine the adaptive threshold value. In this research, Otsu’s method is proposed to achieve the best threshold value. The threshold value depends on the testing image. 84 images have been used as training set. Proposed method has been tested by using 30 images. In this case, 20 images have single line handwriting word and the 10 i...

Identification of Pedestrians Attributes Based on Multi-Class Multi-Label Classification using Convolutional Neural Network (CNN)

The usage of computer vision in identifying pedestrians attributes has received a great attention... more The usage of computer vision in identifying pedestrians attributes has received a great attention, especially in the visual surveillance systems. For instance, searching for system based on the attributes. Attributes Identification using Convolutional Neural Network architecture is presented in this article, since the architecture can perform feature learning. CNN consist of convolution layer, ReLU, Pooling, and Fully-connected. There are three experiment scenarios are conducted based on the number of convolution layers, to determine the effect of layers on CNN performance. Three different CNN architectures were trained and tested using a PETA dataset with 35 attributes. The highest accuracy achieved is 75.66% based on number of convolutional layers. The conducted experiments showed that more numbers of convolution layers used would produce the better CNN's performance.

E-government learning media through augmented reality technology

Bulletin of Social Informatics Theory and Application, 2020

Augmented reality (AR) is one of the newest learning technologies besides Artificial Intelligent,... more Augmented reality (AR) is one of the newest learning technologies besides Artificial Intelligent, Virtual Reality, Virtual World, Alternates Reality Games, and Wearable Computing. E-Government is becoming an evolving academic field and transforming to be E-Governance, E-Participation, E-Service, E-Democracy, E-Voting issues. Therefore, some universities have E-Government course in their curriculum. AR can be one of the learning media to learn and teach E-Government effectively and interestingly. This research aims and contributes to develop a novel model of E-Government learning media using AR. Furthermore, this work has three novelties as follows:First, a method to develop e-Government learning media using AR. Second, a novel model of e-government learning media based on augmented reality technology. Third, the material courses are command centre and public service mall materials based in Surabaya, Indonesia. The new model of learning media using AR consists of main components, i....

A Conceptual Model of Indonesian Question Answering System based on Semantic Web

Journal of Physics: Conference Series, 2020

This study aims to develop a conceptual model of question answering system in Bahasa using Semant... more This study aims to develop a conceptual model of question answering system in Bahasa using Semantic web technology. A trigger was built to periodically collect data scattered on the internet in a data center to shorten the query process on data without reducing the data completeness. The test case domain was automotive. This study used four types of ontologies: main ontology, ontologies from manufacturers, dealers, and rentals. This research has some novelties, such as solving the weakness of distributed and centralized data and combining the advantages of both by storing distributed data into a pool (centralized data) periodically, then querying the pool. Furthermore, it makes contribution by providing a conceptual model of Indonesian question answering system, particularly in automotive domain. This has some implications, such as extend the body of knowledge of computer science, especially semantic web and ontology fields.

Classification of EEG signals uses the coefficient of wavelet transform and K-nearest neighbor

IOP Conference Series: Materials Science and Engineering, 2020

The research that was built was used to explain the application of Electroencephalography (EEG) s... more The research that was built was used to explain the application of Electroencephalography (EEG) signal waves. EEG data is used to move the cursor up and the cursor down. Characterization of each EEG signal uses the Wavelet method taken at each the subband of the wavelet process. Wavelet process by taking the average value and the standard deviation value of the wavelet coefficient. The average value and the standard deviation value is used as an EEG feature. K-Nearest Neighbor is used as an identification whether the cursor will move up or vice versa. This study uses of 100 EEG signal data consisting of 50 test data and 50 testing data. The accuracy of identification uses the 80% K-NN method.

A novel framework of e-participation for smart cities

Bulletin of Social Informatics Theory and Application, 2019

The smart city is an exciting concept for improving the quality of the city. However, a smart cit... more The smart city is an exciting concept for improving the quality of the city. However, a smart city needs participation from citizens and all related stakeholders to use the technologies effectively in order to achieve the goal and solve the problems. Even though the city already has high-tech infrastructures, participation still required to provide ideas, inputs, and roles within the development of smart city concepts. This research aims to describe and analyze the current state of the e-Participation frameworks and propose a novel framework for smart cities. The analysis will go through a systematic literature review. Hopefully, this study makes contributions by providing a novel framework of e-Participation for smart cities. This research has implications for theory and practice. For theory, the novel framework can be added to the body of knowledge of e-participation, e-government, and smart cities fields. For practice, the framework will be useful for practitioners, policymakers,...

Extraction of ElectroEncephaloGraph (EEG) Signal Using the Subband Coefficient of Wavelet Transform on Cursor Moves

IOP Conference Series: Materials Science and Engineering, 2019

This research explains the application of ElectroEncephaloGraph (EEG) signal waves used to move t... more This research explains the application of ElectroEncephaloGraph (EEG) signal waves used to move the up cursor and down cursor. In each sub band of the waveform, Electro Encephalo Graph (EEG) will produce the average and standard deviation to be used as a feature of the EEG. Artificial Neural Network Backpropagation as the basis for determining whether the cursor moves up or the cursor moves down. The data used in this study is EEG data derived from BCI Competition 2003 (BCI Competition 2003). Decision-making is done in two stages. In the first stage, the mean and standard deviation values on each wavelet subband as a feature extraction of EEG data. This feature is an input to the Backpropagation Neural Network. In the second stage of the process of inserting into two classes (class 0 and class 1) EEG data files, there are 260 EEG and 293 file training data files from EEG file test data files, totaling to 553 EEG data files. The results obtained for the classification of EEG results ...

The Dimensionality Reduction Based on the Average of the Left and the Right Structure Preserving Projection

International Review on Computers and Software (IRECOS), 2016

The large dimensionalities are the crucial problem in computer vision, especially in biometrics f... more The large dimensionalities are the crucial problem in computer vision, especially in biometrics field. In this research, a new method to reduce the dimensionality based on the left and right preserving projection is proposed. The proposed approach transforms the image input to integrate the row and column information. The image data transformation is employed through the shifting to the left and the right. The shifting results, for both the left and right are computed showing the average values needed to realize the dominant features. In this research, the proposed approach uses the Eigenfaces process calculation to avoid the singularity problem. The projection results of the Eigenfaces to the average value of the calculation results are employed to further process on the covariance of the left and the right Structure Preserving Projection. The Eigenvalue problem solution is applied to determine the Eigenvectors associated to the largest Eigenvalues. The results of the experiments demonstrated that the maximum acceptance rates are 100% for the ORL, 98.78% for the YALE, and 96.52% for the UoB face database. The best average acceptance rates achieved are 95.3%, 97.94%, and 95.21% for the ORL, YALE, and UoB face database respectively. The results have also exhibited that the proposal has outperformed the other appearance approaches.

Windowing System Facial Detection based on Gabor Kernel Filter, Fast Fourier Transform, and Probabilistic Learning Vector Quantization

International Journal of Technology, 2017

Facial detection is a crucial stage in the facial recognition process. Misclassification during t... more Facial detection is a crucial stage in the facial recognition process. Misclassification during the facial detection process will impact recognition results. In this research, windowing system facial detection using the Gabor kernel filter and the fast Fourier transform was proposed. The training set images, for both facial and non-facial images, were processed to obtain the local features by using the Gabor kernel filter and the fast Fourier transform. The local features were measured using probabilistic learning vector quantization. In this process, facial and non-facial features were classified using label 1 and -1. The proposed method was evaluated using facial and non-facial image testing sets, which were taken from the MIT+CMU image database. The testing images were enhanced first before the detection process using four different enhancement methods: histogram equalization, adaptive histogram equalization, contrast limited adaptive histogram equalization, and the single-scale retinex method. The detection results demonstrated that the highest average accuracy was 83.44%.

Three-directional of the ID-diagonal fisherface modeling based feature extraction

2016 International Seminar on Intelligent Technology and Its Applications (ISITIA), 2016

Three directional of the 1D diagonal fisherface for feature extraction is proposed in this resear... more Three directional of the 1D diagonal fisherface for feature extraction is proposed in this research for the face recognition system. Three directional of the diagonal matrix is built in the proposed method, i.e. the input face image is transformed into a diagonal matrix. Therefore, another form of the input image is created for the feature extraction using the diagonal matrix of the input image. The diagonal matrix is then transformed into 1D matrix for each training set. The feature extraction used this matrix to create the projection and weight matrix using Fisher Face. Two until ten features are extracted using the proposed method. There are four distance methods are used in the experiments, i.e. Euclidian, Manhattan, Canberra, and Chebychev. The experiments of the proposed system are conducted on three databases, i.e. ORL, University of Bern, and YALE-A databases. The recognition accuracy of the proposed model for face recognition achieved up to 92% accuracy.

A new approach: the local feature extraction based on the new regulation of the locally preserving projection

Applied Mathematical Sciences, 2015

A novelty of the local feature extraction was proposed for face recognition. To optimize the Eige... more A novelty of the local feature extraction was proposed for face recognition. To optimize the Eigenvalue and Eigenvector, a new regulation has been embedded to the Locally Preserving Projection. The proposed method has reduced computation time to obtain the new subspace of the original of the Locally Preserving Projection. The proposed method has also produced orthogonal basis function matrix. However, orthogonal basis function matrix can reconstruct easier than non-orthogonal function. The proposed method has been evaluated by using three face image databases, they are the University of Bern, the YALE-A, and the ORL face image databases. The experimental results of the proposed method have produced the recognition rate 96% for the University of Bern, 96.19% for the YALE-A and 98% for the ORL facial image databases. The experimental results of the proposed method have produced higher recognition rate than the Principal Component Analysis, the Linear Discriminant Analysis and the Locally Preserving Projection.

Appearance Global and Local Structure Fusion for Face Image Recognition

TELKOMNIKA (Telecommunication Computing Electronics and Control), 2011

Analisis komponen utama (PCA) dan analisis deskriminan linear (LDA) merupakan metode ekstraksi be... more Analisis komponen utama (PCA) dan analisis deskriminan linear (LDA) merupakan metode ekstraksi berbasis penampakan yang menghasilkan fitur-fitur dengan struktur global. Fitur-fitur dengan struktur global mempunyai kelemahan, yaitu fitur-fitur dengan struktur lokal tidak dapat dicirikan. Proyeksi pelestarian lokalitas (LPP) dan wajah-Laplacian orthogonal (OLF) merupakan metode ekstraksi model penampakan yang menghasilkan fitur-fitur dengan struktur lokal, namun fitur struktur global diabaikan.

Enhancement of the Adaptive Shape Variants Average Values by Using Eight Movement Directions for Multi-Features Detection of Facial Sketch with Tilt Angle 15°

ITB Journal of Information and Communication Technology, 2012

This paper aims to detect multi features of a facial sketch by using a novel approach. The detect... more This paper aims to detect multi features of a facial sketch by using a novel approach. The detection of multi features of facial sketch has been conducted by several researchers, but they mainly considered frontal face sketches as object samples. In fact, the detection of multi features of facial sketch with certain angle is very important to assist police for describing the criminal's face, when criminal's face only appears on certain angle. Integration of the maximum line gradient value enhancement and the level set methods was implemented to detect facial features sketches with tilt angle to 15 degrees. However, these methods tend to move towards non features when there are a lot of graffiti around the shape. To overcome this weakness, the author proposes a novel approach to move the shape by adding a parameter to control the movement based on enhancement of the adaptive shape variants average values with 8 movement directions. The experimental results show that the proposed method can improve the detection accuracy up to 92.74%.