Deny Wiria Nugraha
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Deny Wiria Nugraha has a total of 36 co-authors. Here are the top co-authors by number of shared publications:
- Amil Ahmad Ilham
Department of Informatics, Universitas Hasanuddin, Jl. Poros Malino Km. 6, Bontomarannu, Gowa, Sulawesi Selatan 92171, Indonesia
2 shared publications - Amriana Amriana
Universitas Tadulako
2 shared publications - Andani Achmad
Universitas Hasanuddin
2 shared publications - Ardiaty Arief
Hasanuddin Univ, Dept Elect Engn, Gowa 92171, Indonesia
2 shared publications - Septiano Anggun Pratama
Computer Science, Tadulako University
2 shared publications - Syahrullah Syahrullah
Universitas Tadulako
2 shared publications
Latest Publications
- Open Access
- 28 August 2026
- Published by Universitas Informatika dan Bisnis Indonesia in SisInfo : Jurnal Sistem Informasi dan Informatika
- Vol. 8 (2) , 154-163
Analysis And Design of the Electronic-Based Tax Information System (SIPABETA) Architecture Using the TOGAF Architecture Development Method (ADM) at the Palu City Regional Revenue Agency
This study aims to analyze and design the architecture of the Electronic-Based Tax Information System (SIPABETA) at the Palu City Regional Revenue Agency (Bapenda) using the TOGAF Architecture Development Method (ADM). The study addresses data fragmentation, manual business processes, and limited system integration, which reduce the effectiveness of local tax administration and constrain the optimization of Locally Generated Revenue (PAD). A descriptive qualitative method was applied through a case-study approach involving interviews, direct observation, and document review. The analysis followed the TOGAF ADM phases, from the Preliminary Phase through Architecture Change Management. The study produced an enterprise architecture blueprint comprising an integrated business architecture, centralized data architecture, modular application architecture, and web-based technology architecture supported by application programming interfaces (APIs) and stronger data security. The proposed architecture is expected to improve operational efficiency, information accuracy, public service quality, and the transparency and sustainability of local tax governance in Palu City.
- 28 August 2026
- Published by Universitas Informatika dan Bisnis Indonesia in SisInfo : Jurnal Sistem Informasi dan Informatika
- Vol. 8 (2) , 154-163
Open AccessAnalysis And Design of the Electronic-Based Tax Information System (SIPABETA) Architecture Using the TOGAF Architecture Development Method (ADM) at the Palu City Regional Revenue Agency
This study aims to analyze and design the architecture of the Electronic-Based Tax Information System (SIPABETA) at the Palu City Regional Revenue Agency (Bapenda) using the TOGAF Architecture Development Method (ADM). The study addresses data fragmentation, manual business processes, and limited system integration, which reduce the effectiveness of local tax administration and constrain the optimization of Locally Generated Revenue (PAD). A descriptive qualitative method was applied through a case-study approach involving interviews, direct observation, and document review. The analysis followed the TOGAF ADM phases, from the Preliminary Phase through Architecture Change Management. The study produced an enterprise architecture blueprint comprising an integrated business architecture, centralized data architecture, modular application architecture, and web-based technology architecture supported by application programming interfaces (APIs) and stronger data security. The proposed architecture is expected to improve operational efficiency, information accuracy, public service quality, and the transparency and sustainability of local tax governance in Palu City.
- Open Access
- 1 August 2026
- Published by Institute of Advanced Engineering and Science in Bulletin of Electrical Engineering and Informatics
- Vol. 15 (4) , 3457-3468
Smart measurement revolution using long range technology and HLW8012 sensor integration
Long range (LoRa) has advantages compared to other communication systems such as power line control (PLC), narrowband internet of things (NB-IoT), and Sigfox. The advantage of low power consumption is the implementation and reliability of LoRa communication systems across all segments of smart grid (SG) and home automation applications. Parameters of the reliability of the communication system can also be seen from the maximum distance, signal strength, and network quality in indoor and outdoor conditions as well as different heights and levels of signal collisions when transmitted. This study aims to utilize unlicensed LoRa Shield SX1276 on the 915 MHz frequency spectrum for SG communication networks to monitor messages in the form of current, voltage, and power factor using the HLW8012 sensor. This research produces a complete LoRa communication system monitoring product according to the information needed by the user.
- 1 August 2026
- Published by Institute of Advanced Engineering and Science in Bulletin of Electrical Engineering and Informatics
- Vol. 15 (4) , 3457-3468
Open AccessSmart measurement revolution using long range technology and HLW8012 sensor integration
Long range (LoRa) has advantages compared to other communication systems such as power line control (PLC), narrowband internet of things (NB-IoT), and Sigfox. The advantage of low power consumption is the implementation and reliability of LoRa communication systems across all segments of smart grid (SG) and home automation applications. Parameters of the reliability of the communication system can also be seen from the maximum distance, signal strength, and network quality in indoor and outdoor conditions as well as different heights and levels of signal collisions when transmitted. This study aims to utilize unlicensed LoRa Shield SX1276 on the 915 MHz frequency spectrum for SG communication networks to monitor messages in the form of current, voltage, and power factor using the HLW8012 sensor. This research produces a complete LoRa communication system monitoring product according to the information needed by the user.
- Open Access
- 25 June 2026
- Published by STMIK AMIKOM Yogyakarta in Journal of Information System Management (JOISM)
- Vol. 8 (1) , 188-194
SISTEM PERINGATAN DINI GEMPA DAN TSUNAMI MENGGUNAKAN ARTIFICIAL NEURAL NETWORK
Penelitian ini bertujuan untuk meningkatkan keselamatan publik melalui penyediaan prediksi tsunami yang cepat dan andal dengan memanfaatkan algoritma Artificial Neural Network (ANN) dalam sistem berbasis web. Masalah utama yang diangkat adalah tingginya risiko kerugian material dan korban jiwa akibat keterlambatan informasi serta potensi false alarm pada pengolahan data konvensional. Metode penyelesaian dilakukan dengan menerapkan arsitektur jaringan saraf ANN (susuna lapisan layer 9-10-8-1) yang mengolah parameter Magnitudo, Kedalaman, Jarak ke Pantai, serta Jenis Patahan sebagai input utama. Hasil perancangan menunjukkan performa sistem yang sangat presisi, dibuktikan dengan perolehan nilai Error Akhir sebesar 0,004950 pada iterasi ke-58 yang merepresentasikan tingkat akurasi pelatihan mencapai 99,50%. Selain itu, pengujian pada skenario gempa Magnitudo 7.5 menghasilkan Probabilitas Tsunami sebesar 93,38% dengan tingkat kepercayaan tinggi. Integrasi ANN dalam platform web ini terbukti mampu menyediakan informasi yang terstruktur, akurat, dan ramah pengguna sebagai instrumen krusial dalam pengambilan tindakan tanggap bencana yang efektif.
- 25 June 2026
- Published by STMIK AMIKOM Yogyakarta in Journal of Information System Management (JOISM)
- Vol. 8 (1) , 188-194
Open AccessSISTEM PERINGATAN DINI GEMPA DAN TSUNAMI MENGGUNAKAN ARTIFICIAL NEURAL NETWORK
Penelitian ini bertujuan untuk meningkatkan keselamatan publik melalui penyediaan prediksi tsunami yang cepat dan andal dengan memanfaatkan algoritma Artificial Neural Network (ANN) dalam sistem berbasis web. Masalah utama yang diangkat adalah tingginya risiko kerugian material dan korban jiwa akibat keterlambatan informasi serta potensi false alarm pada pengolahan data konvensional. Metode penyelesaian dilakukan dengan menerapkan arsitektur jaringan saraf ANN (susuna lapisan layer 9-10-8-1) yang mengolah parameter Magnitudo, Kedalaman, Jarak ke Pantai, serta Jenis Patahan sebagai input utama. Hasil perancangan menunjukkan performa sistem yang sangat presisi, dibuktikan dengan perolehan nilai Error Akhir sebesar 0,004950 pada iterasi ke-58 yang merepresentasikan tingkat akurasi pelatihan mencapai 99,50%. Selain itu, pengujian pada skenario gempa Magnitudo 7.5 menghasilkan Probabilitas Tsunami sebesar 93,38% dengan tingkat kepercayaan tinggi. Integrasi ANN dalam platform web ini terbukti mampu menyediakan informasi yang terstruktur, akurat, dan ramah pengguna sebagai instrumen krusial dalam pengambilan tindakan tanggap bencana yang efektif.
- Open Access
- 31 May 2026
- Published by Institut Teknologi Garut in Jurnal Algoritma
- Vol. 23 (1) , 906–917
Integrasi YOLOv12 dan Konveyor Belt Untuk Mendeteksi dan Menyortir Kerusakan Kaleng Cat Secara Otomatis
Penelitian ini bertujuan mengembangkan sistem otomasi inspeksi dan penyortiran kerusakan kaleng cat berbasis integrasi deep learning dan konveyor belt untuk meningkatkan efisiensi dan konsistensi quality control di industri. Metode yang digunakan meliputi perancangan sistem mekanikal konveyor, integrasi rangkaian elektronik, serta pengembangan model deteksi objek menggunakan YOLOv12 dengan konfigurasi multi-kamera untuk meminimalkan blind spot. Dataset terdiri dari 1.209 citra yang dibagi menjadi data training, validation, dan test, dengan proses augmentasi untuk meningkatkan robustness model. Evaluasi dilakukan menggunakan metrik precision, recall, F1-score, dan mAP, serta pengujian sistem secara end-to-end berdasarkan akurasi sortir, latency, dan throughput pada berbagai kondisi pencahayaan dan kecepatan konveyor. Hasil penelitian menunjukkan bahwa model mencapai precision 0,98, recall 0,96, F1-score 0,97, dan mAP50–95 sebesar 0,96. Namun, implementasi sistem menghasilkan akurasi sortir sebesar 55,6% dengan performa optimal pada kecepatan sedang, yang menunjukkan adanya pengaruh signifikan faktor operasional terhadap kinerja sistem secara keseluruhan.
- 31 May 2026
- Published by Institut Teknologi Garut in Jurnal Algoritma
- Vol. 23 (1) , 906–917
Open AccessIntegrasi YOLOv12 dan Konveyor Belt Untuk Mendeteksi dan Menyortir Kerusakan Kaleng Cat Secara Otomatis
Penelitian ini bertujuan mengembangkan sistem otomasi inspeksi dan penyortiran kerusakan kaleng cat berbasis integrasi deep learning dan konveyor belt untuk meningkatkan efisiensi dan konsistensi quality control di industri. Metode yang digunakan meliputi perancangan sistem mekanikal konveyor, integrasi rangkaian elektronik, serta pengembangan model deteksi objek menggunakan YOLOv12 dengan konfigurasi multi-kamera untuk meminimalkan blind spot. Dataset terdiri dari 1.209 citra yang dibagi menjadi data training, validation, dan test, dengan proses augmentasi untuk meningkatkan robustness model. Evaluasi dilakukan menggunakan metrik precision, recall, F1-score, dan mAP, serta pengujian sistem secara end-to-end berdasarkan akurasi sortir, latency, dan throughput pada berbagai kondisi pencahayaan dan kecepatan konveyor. Hasil penelitian menunjukkan bahwa model mencapai precision 0,98, recall 0,96, F1-score 0,97, dan mAP50–95 sebesar 0,96. Namun, implementasi sistem menghasilkan akurasi sortir sebesar 55,6% dengan performa optimal pada kecepatan sedang, yang menunjukkan adanya pengaruh signifikan faktor operasional terhadap kinerja sistem secara keseluruhan.
- Open Access
- 5 December 2025
- Published by Politeknik Negeri Batam in JOURNAL OF APPLIED INFORMATICS AND COMPUTING
- Vol. 9 (6) , 3012-3024
Performance Comparison of Multilayer Perceptron (MLP) and Random Forest for Early Detection of Cardiovascular Disease
Cardiovascular disease is a disorder of the heart and blood vessels that can lead to heart attacks, strokes, and heart failure, so early detection is essential. This study compares Multilayer Perceptron (MLP) and Random Forest for risk classification in a Kaggle dataset containing 70,000 samples with balanced targets. Pre-processing included age conversion, outlier cleaning, standardization, and feature selection based on feature importance. Both models were optimized using RandomizedSearchCV and evaluated using accuracy, precision, recall, F1-score, AUC-ROC, confusion matrix, and k-fold cross-validation. The results show that the accuracy of MLP is 73.90% and Random Forest is 74.23% with an AUC of 0.80 for both. Random Forest is more stable across all folds and performs better on the negative class, while MLP is slightly more sensitive to the positive class. Independent t-test and Mann-Whitney U tests show p>0.05, indicating that the difference in performance is not significant. The most influential features were diastolic blood pressure, age, cholesterol, and systolic blood pressure. The non-clinical Streamlit prototype demonstrated the model's potential for education and initial decision support.
- 5 December 2025
- Published by Politeknik Negeri Batam in JOURNAL OF APPLIED INFORMATICS AND COMPUTING
- Vol. 9 (6) , 3012-3024
Open AccessPerformance Comparison of Multilayer Perceptron (MLP) and Random Forest for Early Detection of Cardiovascular Disease
Cardiovascular disease is a disorder of the heart and blood vessels that can lead to heart attacks, strokes, and heart failure, so early detection is essential. This study compares Multilayer Perceptron (MLP) and Random Forest for risk classification in a Kaggle dataset containing 70,000 samples with balanced targets. Pre-processing included age conversion, outlier cleaning, standardization, and feature selection based on feature importance. Both models were optimized using RandomizedSearchCV and evaluated using accuracy, precision, recall, F1-score, AUC-ROC, confusion matrix, and k-fold cross-validation. The results show that the accuracy of MLP is 73.90% and Random Forest is 74.23% with an AUC of 0.80 for both. Random Forest is more stable across all folds and performs better on the negative class, while MLP is slightly more sensitive to the positive class. Independent t-test and Mann-Whitney U tests show p>0.05, indicating that the difference in performance is not significant. The most influential features were diastolic blood pressure, age, cholesterol, and systolic blood pressure. The non-clinical Streamlit prototype demonstrated the model's potential for education and initial decision support.
- Open Access
- 9 August 2025
- Published by Universitas Hamzanwadi in EDUMATIC: Jurnal Pendidikan Informatika
- Vol. 9 (2) , 639-648
Segmentasi Pelanggan Menggunakan Kerangka LRFMV dan Algoritma K-Means untuk Optimalisasi Strategi Pemasaran
In this competitive digital era, customer behavior is key to maintaining loyalty and increasing profitability. This study aims to implement customer segmentation using the Length, Recency, Frequency, Monetary, Volume (LRFMV) approach and the K-Means algorithm to identify customer behavior characteristics and determine high-value segments. The combination of these five dimensions has rarely been used in previous studies, thus providing a new contribution to data-based customer behavior analysis. This study adopts an exploratory descriptive quantitative approach. The data used consists of 2,098 transactions from 452 customers, sourced from a public GitHub dataset. The data analysis process includes preprocessing, determining LRFMV values, and segmentation using K-Means Clustering. The Silhouette Coefficient is used to evaluate cluster quality and determine the optimal number of clusters. The results show that the best configuration is obtained at k=5 with a Silhouette value of 0.842. The findings show five customer segments with different characteristics and Customer Lifetime Value (CLV) values. Clusters 0 and 2 are categorized as Loyal Customers (L↑R↓F↑M↑V↑) with the highest CLV. Clusters 3 and 1 are Inactive New Customers (L↓R↑F↓M↓V↓) with low contribution. Cluster 4 consists of Inactive Customers (L↓R↓F↓M↓V↓), indicating overall inactivity. These segmentation results are used to develop more targeted strategies, such as loyalty programs or reactivation campaigns, to optimize marketing strategies based on customer value.
- 9 August 2025
- Published by Universitas Hamzanwadi in EDUMATIC: Jurnal Pendidikan Informatika
- Vol. 9 (2) , 639-648
Open AccessSegmentasi Pelanggan Menggunakan Kerangka LRFMV dan Algoritma K-Means untuk Optimalisasi Strategi Pemasaran
In this competitive digital era, customer behavior is key to maintaining loyalty and increasing profitability. This study aims to implement customer segmentation using the Length, Recency, Frequency, Monetary, Volume (LRFMV) approach and the K-Means algorithm to identify customer behavior characteristics and determine high-value segments. The combination of these five dimensions has rarely been used in previous studies, thus providing a new contribution to data-based customer behavior analysis. This study adopts an exploratory descriptive quantitative approach. The data used consists of 2,098 transactions from 452 customers, sourced from a public GitHub dataset. The data analysis process includes preprocessing, determining LRFMV values, and segmentation using K-Means Clustering. The Silhouette Coefficient is used to evaluate cluster quality and determine the optimal number of clusters. The results show that the best configuration is obtained at k=5 with a Silhouette value of 0.842. The findings show five customer segments with different characteristics and Customer Lifetime Value (CLV) values. Clusters 0 and 2 are categorized as Loyal Customers (L↑R↓F↑M↑V↑) with the highest CLV. Clusters 3 and 1 are Inactive New Customers (L↓R↑F↓M↓V↓) with low contribution. Cluster 4 consists of Inactive Customers (L↓R↓F↓M↓V↓), indicating overall inactivity. These segmentation results are used to develop more targeted strategies, such as loyalty programs or reactivation campaigns, to optimize marketing strategies based on customer value.
- Open Access
- 3 August 2025
- Published by Politeknik Negeri Batam in JOURNAL OF APPLIED INFORMATICS AND COMPUTING
- Vol. 9 (4) , 1969-1577
Implementation of ResNet-50-Based Convolutional Neural Network For Mobile Skin Cancer Classification
The skin is one of the most important parts of the human body, serving vital functions such as protecting internal organs from injury, shielding against direct bacterial exposure, regulating body temperature, and more. However, the skin is also susceptible to diseases, one of which is skin cancer. Skin cancer can be extremely dangerous if not treated promptly, as it can lead to death. Therefore, early detection is crucial. This study proposes a technology-based solution by classifying skin cancer using a convolutional neural network (CNN) with a ResNet50 architecture implemented into a mobile application via a REST API using Flask. The HAM10000 dataset, consisting of 10,015 skin lesion images across seven classes, was used for model training. Various testing scenarios were conducted to determine the optimal parameter combination. The best results were achieved with an accuracy of 83.84%, precision and recall of 83%, and an F1-score of 83%, using a training data configuration of 70%, dropout of 0.4, and a batch size of 64. The model implemented in this Android application can perform early detection of skin cancer quickly, practically, and easily accessible to the general public, though healthcare professionals must still supervise it. However, although this model can assist users in making early predictions, the prediction results from this model are only a tool for early detection and do not replace clinical diagnosis by professional medical personnel.2) Figure 8 shows the display for taking pictures through the gallery or camera. Users can choose the image they want to upload from the gallery or the camera to be analysed and predicted by the model.
- 3 August 2025
- Published by Politeknik Negeri Batam in JOURNAL OF APPLIED INFORMATICS AND COMPUTING
- Vol. 9 (4) , 1969-1577
Open AccessImplementation of ResNet-50-Based Convolutional Neural Network For Mobile Skin Cancer Classification
The skin is one of the most important parts of the human body, serving vital functions such as protecting internal organs from injury, shielding against direct bacterial exposure, regulating body temperature, and more. However, the skin is also susceptible to diseases, one of which is skin cancer. Skin cancer can be extremely dangerous if not treated promptly, as it can lead to death. Therefore, early detection is crucial. This study proposes a technology-based solution by classifying skin cancer using a convolutional neural network (CNN) with a ResNet50 architecture implemented into a mobile application via a REST API using Flask. The HAM10000 dataset, consisting of 10,015 skin lesion images across seven classes, was used for model training. Various testing scenarios were conducted to determine the optimal parameter combination. The best results were achieved with an accuracy of 83.84%, precision and recall of 83%, and an F1-score of 83%, using a training data configuration of 70%, dropout of 0.4, and a batch size of 64. The model implemented in this Android application can perform early detection of skin cancer quickly, practically, and easily accessible to the general public, though healthcare professionals must still supervise it. However, although this model can assist users in making early predictions, the prediction results from this model are only a tool for early detection and do not replace clinical diagnosis by professional medical personnel.2) Figure 8 shows the display for taking pictures through the gallery or camera. Users can choose the image they want to upload from the gallery or the camera to be analysed and predicted by the model.
- Open Access
- 1 April 2025
- Published by Institute of Advanced Engineering and Science in Bulletin of Electrical Engineering and Informatics
- Vol. 14 (2) , 1391-1406
Transformers for aerial images semantic segmentation of natural disaster-impacted areas in natural disaster assessment
Aerial image segmentation of natural disaster-impacted areas and detailed and automatic natural disaster assessment are the main focus of this study. Detecting and recognizing objects on aerial images of areas impacted by natural disasters and assessing natural disaster-impacted areas are still difficult problems. To solve these problems, this study utilizes four of the latest transformer-based semantic segmentation network models, bidirectional encoder representation from image transformers (BEIT), dense prediction transformer (DPT), OneFormer, and SegFormer, and proposes a detailed and automatic natural disaster assessment of the segmented image. The SegFormer model achieved the first-best result, and the OneFormer model achieved the second-best result. The SegFormer model outperformed OneFormer by 1.58% higher for the mean accuracy value and 4.28% for the mean intersection over union (mIoU) value. All receiver operating characteristics (ROC) curves have mean area under curve (AUC) values above 0.9, which means that the SegFormer model performs well in generating semantic segmentation images. The fuzzy c-means (FCM) clustering algorithm performed well and could automatically cluster the natural disaster assessments into four categories. This study has produced semantic segmentation of aerial images of areas impacted by natural disasters and natural disaster assessments, which can be used in natural disaster management systems.
- 1 April 2025
- Published by Institute of Advanced Engineering and Science in Bulletin of Electrical Engineering and Informatics
- Vol. 14 (2) , 1391-1406
Open AccessTransformers for aerial images semantic segmentation of natural disaster-impacted areas in natural disaster assessment
Aerial image segmentation of natural disaster-impacted areas and detailed and automatic natural disaster assessment are the main focus of this study. Detecting and recognizing objects on aerial images of areas impacted by natural disasters and assessing natural disaster-impacted areas are still difficult problems. To solve these problems, this study utilizes four of the latest transformer-based semantic segmentation network models, bidirectional encoder representation from image transformers (BEIT), dense prediction transformer (DPT), OneFormer, and SegFormer, and proposes a detailed and automatic natural disaster assessment of the segmented image. The SegFormer model achieved the first-best result, and the OneFormer model achieved the second-best result. The SegFormer model outperformed OneFormer by 1.58% higher for the mean accuracy value and 4.28% for the mean intersection over union (mIoU) value. All receiver operating characteristics (ROC) curves have mean area under curve (AUC) values above 0.9, which means that the SegFormer model performs well in generating semantic segmentation images. The fuzzy c-means (FCM) clustering algorithm performed well and could automatically cluster the natural disaster assessments into four categories. This study has produced semantic segmentation of aerial images of areas impacted by natural disasters and natural disaster assessments, which can be used in natural disaster management systems.
- Open Access
- 10 January 2025
- Published by Politeknik Negeri Batam in JOURNAL OF APPLIED INFORMATICS AND COMPUTING
- Vol. 9 (1) , 16-22
Analysis of the Use of MTCNN and Landmark Technology to Improve the Accuracy of Facial Recognition on Official Documents
A face recognition system consists of two stages: face detection and face recognition. Detection of features such as eyes and mouth is important in facial image processing, especially for official documents such as identity cards. To ensure identification accuracy, this research applies facial landmark extraction technology and MTCNN (Multi-Task Cascaded Convolutional Neural Network). The purpose of this research is to evaluate the accuracy of MTCNN in detecting facial features at the Department of Population and Civil Registration (dukcapil) Palu City, using facial landmarks and waterfall methods as an application development methodology. The evaluation results show that MTCNN has high face recognition accuracy and good positioning ability regardless of what GPU in use as long have right CPU and System Operation. In comparison, the Viola-Jones algorithm is effective for high-speed applications, while SSD offers balanced performance with GPU device requirements for optimal performance. While MTCNN proved to be effective, challenges still exist, such as false positives and false negatives, especially in poor lighting conditions and extreme poses. Image and camera quality, including resolution and facial expression, also affects detection accuracy. These findings suggest that the application of MTCNN can improve face recognition accuracy for official documents, although it requires addressing existing challenges. With this technology, it is expected that errors in facial recognition can be minimized, resulting in more reliable data that meets the standards for issuing identity documents. This research contributes to the development of a more accurate and efficient face recognition system for personal identification applications.
- 10 January 2025
- Published by Politeknik Negeri Batam in JOURNAL OF APPLIED INFORMATICS AND COMPUTING
- Vol. 9 (1) , 16-22
Open AccessAnalysis of the Use of MTCNN and Landmark Technology to Improve the Accuracy of Facial Recognition on Official Documents
A face recognition system consists of two stages: face detection and face recognition. Detection of features such as eyes and mouth is important in facial image processing, especially for official documents such as identity cards. To ensure identification accuracy, this research applies facial landmark extraction technology and MTCNN (Multi-Task Cascaded Convolutional Neural Network). The purpose of this research is to evaluate the accuracy of MTCNN in detecting facial features at the Department of Population and Civil Registration (dukcapil) Palu City, using facial landmarks and waterfall methods as an application development methodology. The evaluation results show that MTCNN has high face recognition accuracy and good positioning ability regardless of what GPU in use as long have right CPU and System Operation. In comparison, the Viola-Jones algorithm is effective for high-speed applications, while SSD offers balanced performance with GPU device requirements for optimal performance. While MTCNN proved to be effective, challenges still exist, such as false positives and false negatives, especially in poor lighting conditions and extreme poses. Image and camera quality, including resolution and facial expression, also affects detection accuracy. These findings suggest that the application of MTCNN can improve face recognition accuracy for official documents, although it requires addressing existing challenges. With this technology, it is expected that errors in facial recognition can be minimized, resulting in more reliable data that meets the standards for issuing identity documents. This research contributes to the development of a more accurate and efficient face recognition system for personal identification applications.
- Open Access
- 1 January 2024
- Published by AIP Publishing in AIP Conference Proceedings
- Vol. 3145 (1) , 020001
Aerial imagery segmentation of natural disaster-affected areas using deep convolutional networks for disaster assessment
After a natural disaster, aerial imagery from an unmanned aerial vehicle (UAV) is an invaluable resource for impact assessment and disaster management. Traditio After a natural disaster, aerial imagery from an unmanned aerial vehicle (UAV) is an invaluable resource for impact assessment and disaster management. Traditio
- 1 January 2024
- Published by AIP Publishing in AIP Conference Proceedings
- Vol. 3145 (1) , 020001
Open AccessAerial imagery segmentation of natural disaster-affected areas using deep convolutional networks for disaster assessment
After a natural disaster, aerial imagery from an unmanned aerial vehicle (UAV) is an invaluable resource for impact assessment and disaster management. Traditio After a natural disaster, aerial imagery from an unmanned aerial vehicle (UAV) is an invaluable resource for impact assessment and disaster management. Traditio