Garuda - Garba Rujukan Digital

Face recognition for occluded face with mask region convolutional neural network and fully convolutional network: a literature review Rahmat Budiarsa; Retantyo Wardoyo; Aina Musdholifah
International Journal of Electrical and Computer Engineering (IJECE) Vol 13, No 5: October 2023
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijece.v13i5.pp5662-5673

Face recognition technology has been used in many ways, such as in the authentication and identification process. The object raised is a piece of face image that does not have complete facial information (occluded face), it can be due to acquisition from a different point of view or shooting a face from a different angle. This object was raised because the object can affect the detection and identification performance of the face image as a whole. Deep leaning method can be used to solve face recognition problems. In previous research, more focused on face detection and recognition based on resolution, and detection of face. Mask region convolutional neural network (mask R-CNN) method still has deficiency in the segmentation section which results in a decrease in the accuracy of face identification with incomplete face information objects. The segmentation used in mask R-CNN is fully convolutional network (FCN). In this research, exploration and modification of many FCN parameters will be carried out using the CNN backbone pooling layer, and modification of mask R-CNN for face identification, besides that, modifications will be made to the bounding box regressor. it is expected that the modification results can provide the best recommendations based on accuracy.

Face recognition with occluded face using improve intersection over union of region proposal network on Mask region convolutional neural network Budiarsa, Rahmat; Wardoyo, Retantyo; Musdholifah, Aina
International Journal of Electrical and Computer Engineering (IJECE) Vol 14, No 3: June 2024
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijece.v14i3.pp3256-3265

Face recognition entails detecting and identifying facial attributes. Mask region convolutional neural network (R-CNN) method is a prominent approach, while prior research predominantly delved into refining loss functions and perfecting object and face detection, recognizing, and identifying faces using imperfect data remained relatively unexplored. This study focuses on an occluded dataset comprising Indonesian faces, wherein 'occluded' denotes facial data that lacks complete visibility-encompassing instances where objects obscure faces or are partially cropped. This investigation involves a deliberate experiment that tailors the intersection over union (IoU) of the region proposal network (RPN) to suit the nuances of occluded Indonesian faces, thereby augmenting accuracy in recognition and segmentation tasks. The innovation IoU in the strategic utilization of Anchors, which involves the exclusion of anchors falling beyond the image borders to optimize computational efficiency. The outcomes of this research are striking; it showcases a remarkable 14.75%, 10.9%, and 12.97% surge based on mean average precision (mAP), mean average recall (mAR), and F1-Scores compared to the conventional Mask R-CNN approach. Notably, our proposed model elevates the average accuracy by 10% to 15% and decreases running time by 21%, a noteworthy enhancement compared to the preceding model. This progress is substantiated by validation utilizing 300 instances dataset, reinforcing the robustness of our approach.

Perbandingan Algoritma Long Short-Term Memory (LSTM) Dan XGBoost Dalam Memprediksi Kualitas Udara Di Jakarta Luthfi radyansyah; Nizirwan; Riya Widayanti; Rahmat Budiarsa
ROUTERS: Jurnal Sistem dan Teknologi Informasi Vol. 4 No. 2, Juli 2026 (In Progress)
Publisher : Program Studi Teknologi Rekayasa Internet, Politeknik Negeri Lampung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25181/rt.v4i2.5013

This study compares the Extreme Gradient Boosting (XGBoost) and Long Short-Term Memory (LSTM) algorithms in predicting air quality in Jakarta using the Air Pollution Standard Index (ISPU) dataset. The dataset consists of 2,874 observations collected from January 1, 2024, to July 31, 2025. The results show that both models are capable of producing accurate predictions; however, XGBoost demonstrates superior performance. XGBoost achieves an RMSE of 1.9704, MAE of 1.0509, MAPE of 1.54%, and an R² of 0.9913. In contrast, LSTM produces an RMSE of 3.0667, MAE of 2.0968, MAPE of 3.71%, and an R² of 0.9790. These findings indicate that XGBoost has lower prediction error and a stronger ability to explain data variability. Additionally, particulate pollutants such as PM2.5 and PM10 are identified as the dominant factors influencing air quality. Overall, XGBoost proves to be more effective, stable, and efficient in modeling air quality data.

IMPLEMENTASI MODEL EFFICIENTNET-B5 DENGAN TRANSFER LEARNING UNTUK KLASIFIKASI TINGKAT ROASTING BIJI KOPI BERBASIS CITRA Jody Kalim; Nenden Siti Fatonah; Riya Widayanti; Rahmat Budiarsa
JATI (Jurnal Mahasiswa Teknik Informatika) Vol. 10 No. 2 (2026): JATI Vol. 10 No. 2
Publisher : Institut Teknologi Nasional Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36040/jati.v10i2.17848

Kopi merupakan komoditas strategis yang berkontribusi besar terhadap perekonomian nasional Indonesia, dengan produksi mencapai 810.000 ton pada tahun 2024. Kualitas produk kopi sangat ditentukan oleh proses roasting yang menghasilkan empat kategori tingkat kematangan dengan karakteristik visual berbeda. Namun, penilaian tingkat roasting secara manual rentan terhadap inkonsistensi akibat ketergantungan pada persepsi visual subjektif operator, sehingga berpotensi menurunkan konsistensi mutu dan daya saing produk kopi di pasar. Penelitian ini bertujuan mengembangkan sistem klasifikasi tingkat roasting biji kopi berbasis citra secara otomatis dan objektif menggunakan arsitektur EfficientNet-B5 dengan pendekatan Transfer Learning. Dataset yang digunakan berjumlah 1.600 citra biji kopi dari Kaggle yang terdistribusi seimbang dalam empat kategori, yaitu Green, Light, Medium, dan Dark, dengan pembagian data 60% latih, 15% validasi, dan 25% uji. Model dikembangkan menggunakan bobot awal ImageNet, pembekuan 150 layer awal, optimizer Adam, serta lapisan klasifikasi tambahan yang terdiri atas GlobalAveragePooling2D, BatchNormalization, Dense, dan Dropout. Hasil evaluasi menunjukkan akurasi pengujian sebesar 98,68% dengan rata-rata Precision, Recall, dan F1-Score sebesar 0,98, mencerminkan kemampuan generalisasi yang tinggi. Eksperimen komparatif membuktikan bahwa Transfer Learning bersifat esensial, karena model tanpa bobot pre-trained mengalami kegagalan konvergensi dengan akurasi hanya 26%. Model juga diintegrasikan ke dalam prototipe aplikasi web berbasis Streamlit untuk inferensi citra secara real-time.

Sistem Pakar Dengan Case Base Reasoning Dan Certainty Factor Untuk Diagnosa THT Budiarsa, Rahmat; Perdana Putra, Nova
Jurnal Pustaka AI (Pusat Akses Kajian Teknologi Artificial Intelligence) Vol 6 No 2 (2026): Pustaka AI (Pusat Akses Kajian Teknologi Artificial Intelligence)
Publisher : Pustaka Galeri Mandiri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55382/jurnalpustakaai.v6i2.1718

Perbandingan jumlah dokter spesialis telinga hidung dan tenggorokan (THT) dengan jumlah provinsi di Indonesia adalah 1: 48 dokter spesialis THT. Penyakit yang dibahas dalam penelitian ini adalah penyakit otitis media akut, presbikusis, vestibulitis, sinusitis, Faringitis, Tonsilitis, Perforasi Septum, Deviasi Septum, dan Abses Parafaringeal dengan 40 gejala. Menurut dr.Nova Perdana Putra, Sp.THT-KL kesembilan penyakit yang disebutkan sebelumnya adalah penyakit yang biasa diderita oleh masyarakat Kerinci, Jambi. Tujuan dan Manfaat penelitian ini adalah mempermudah dan mempercepat pasien atau masyarakat dalam mendiagnosa awal 9 penyakit THT dan memberikan pengetahuan tentang penyakit yang kemungkinan diderita sebelum ke dokter THT serta membantu diagnosa awal yang dilakukan dokter THT. Metode pengumpulan data yang dilakukan adalah wawancara, studi literatur dan observasi berjumlah 150 data. Analisis yang dilakukan adalah analisis data, user dan kebutuhan sistem. Pada tahap perancangan sistem pakar dilakukan akuisisi pengetahuan, membuat tabel basis pengetahuan, membuat basis aturan, Decision Tree, membuat alur keputusan sistem dan perancangan basis data. Implementasi dilakukan dalam bentuk web sesuai dengan alur keputusan sistem dimulai dari menginputkan data (pasien dan gejala), decision tree, menghitung similarity, menghitung certainty factor, menyimpan data kasus baru, konfirmasi data, dan menampilkan hasil diagnosa. Tingkat akurasi sistem diuji menggunakan metode accuracy testing. Dalam Penelitian ini penentuan penyakit dengan kasus baru menggunakan kasus lama dapat diterapkan jika similarity gejala pada kedua kasus ? 80%. Penelitian ini telah diuji menggunakan 105 data uji kasus dan menghasilkan tingkat akurasi 93,33% yang berhasil didiagnosa oleh sistem dan disetujui oleh pakar THT dr. Nova Perdana Putra, Sp.THT-KL.

Pembuatan Model Machine Learning untuk Klasifikasi Risiko Pinjaman Perbankan Menggunakan Random Forest dan XGboost Zacki Ferdinansyah; Rahmat Budiarsa
JURNAL PENELITIAN SISTEM INFORMASI (JPSI) Vol. 4 No. 3 (2026): Agustus : JURNAL PENELITIAN SISTEM INFORMASI
Publisher : Institut Teknologi dan Bisnis (ITB) Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54066/jpsi.v4i3.4644

Credit risk is one of the major challenges in managing loan portfolios in the banking sector, highlighting the need for approaches capable of identifying loan risk patterns and factors associated with problematic loans based on historical data. This study aims to develop and compare the performance of Random Forest and Extreme Gradient Boosting (XGBoost) algorithms in classifying loan risk patterns and to identify the features that contribute to the classification results. The study uses a historical loan dataset processed through exploratory data analysis, categorical data transformation, an 80:20 training and testing data split, and class imbalance handling using Random Under Sampling and SMOTE. Model optimization was performed through hyperparameter tuning using GridSearchCV, while model performance was evaluated based on accuracy, precision, recall, and F1-score. Model interpretation was conducted using Feature Importance, SHAP (SHapley Additive Explanations), and decision tree visualization. The results show that XGBoost with preprocessing and the second hyperparameter tuning achieved the best performance, with an accuracy of 98%, precision of 97%, recall of 85%, and F1-score of 90%, outperforming Random Forest on the dataset used. Interpretability analysis indicates that recoveries, total_rec_prncp, out_prncp, last_pymnt_amnt, and funded_amnt are among the features that contribute substantially to the classification results. These findings indicate that preprocessing, class imbalance handling, and parameter optimization can improve classification performance while providing insights into the factors contributing to loan risk patterns based on historical data.

Pembuatan Model Machine Learning untuk Klasifikasi Risiko Pinjaman Perbankan Menggunakan Random Forest dan XGboost Zacki Ferdinansyah; Rahmat Budiarsa
JURNAL PENELITIAN SISTEM INFORMASI (JPSI) Vol. 4 No. 3 (2026): Agustus : JURNAL PENELITIAN SISTEM INFORMASI
Publisher : Institut Teknologi dan Bisnis (ITB) Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54066/jpsi.v4i3.4644

Credit risk is one of the major challenges in managing loan portfolios in the banking sector, highlighting the need for approaches capable of identifying loan risk patterns and factors associated with problematic loans based on historical data. This study aims to develop and compare the performance of Random Forest and Extreme Gradient Boosting (XGBoost) algorithms in classifying loan risk patterns and to identify the features that contribute to the classification results. The study uses a historical loan dataset processed through exploratory data analysis, categorical data transformation, an 80:20 training and testing data split, and class imbalance handling using Random Under Sampling and SMOTE. Model optimization was performed through hyperparameter tuning using GridSearchCV, while model performance was evaluated based on accuracy, precision, recall, and F1-score. Model interpretation was conducted using Feature Importance, SHAP (SHapley Additive Explanations), and decision tree visualization. The results show that XGBoost with preprocessing and the second hyperparameter tuning achieved the best performance, with an accuracy of 98%, precision of 97%, recall of 85%, and F1-score of 90%, outperforming Random Forest on the dataset used. Interpretability analysis indicates that recoveries, total_rec_prncp, out_prncp, last_pymnt_amnt, and funded_amnt are among the features that contribute substantially to the classification results. These findings indicate that preprocessing, class imbalance handling, and parameter optimization can improve classification performance while providing insights into the factors contributing to loan risk patterns based on historical data.

PREDIKSI TREN PUBLIKASI BLOCKCHAIN DAN KRIPTOGRAFI DI INDONESIA MENGGUNAKAN BIBLIOMETRIK DAN LSTM Fathimah Azzahro Hasni Rifayanti; Rahmat Budiarsa
SKANIKA: Sistem Komputer dan Teknik Informatika Vol 9 No 2 (2026): Jurnal SKANIKA Juli 2026
Publisher : Universitas Budi Luhur

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36080/skanika.v9i2.3843

The rapid adoption of digital technologies in Indonesia has driven increasing research on Blockchain and Cryptography, yet comprehensive mapping and projections of their development remain limited. This study aims to analyze the intellectual landscape and predict publication trends in these two technologies in Indonesia using a quantitative approach integrating bibliometric analysis and Machine Learning. Data were obtained from Dimensions.ai for the 2020–2024 period, yielding 2,193 relevant documents after data cleansing. Network analysis using VOSviewer reveals a shift in Blockchain research from financial assets toward practical applications, particularly Supply Chain management and system efficiency. Meanwhile, Cryptography research shows polarization between the development of classical algorithms and applications for IoT security. A Long Short-Term Memory (LSTM) model was then developed to predict publication trends for 2025–2029 based on annual data from 2020–2024. The model achieved a Mean Absolute Percentage Error (MAPE) below 5% on the test data; however, the limited number of data points should be considered when interpreting this accuracy. LSTM projects a flattening publication trend through 2029, with Cryptography expected to surpass Blockchain by approximately 55 documents. These findings indicate that security and trust may become increasingly critical priorities compared with Blockchain infrastructure development.

Title

Found 8 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 8 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 8 Documents
Search