Garuda - Garba Rujukan Digital

PENERAPAN MODEL ESRGAN UNTUK UPSCALING CITRA DAN VIDEO DIGITAL Suhardi, Syahrul; Habi Talib, Emil Agusalim; Rachman, Fahrim Irhamna; Wahyuni, Titin; Faisal, Muhammad; S.Kuba, Muhammad Syafaat
PROGRESS Vol 18 No 1 (2026): April
Publisher : P3M STMIK Profesional Makassar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.56708/progres.v18i1.539

Low-resolution images and videos remain a common problem in various digital applications due to limited visual quality. Conventional interpolation-based upscaling methods often produce blurry results and lead to the loss of important texture details. This study aims to apply the Enhanced Super-Resolution Generative Adversarial Network (ESRGAN) to improve the resolution of digital images and videos. The dataset used consists of low-resolution images and videos that are processed through preprocessing, model training, and testing stages using the Google Colab environment. The ESRGAN model is trained to generate high-resolution images while preserving visual details and structural information. Model performance is evaluated using the Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), and visual comparison between images before and after the upscaling process. The results show that ESRGAN significantly improves the quality of images and videos compared to conventional interpolation methods, both quantitatively and qualitatively. Therefore, the application of ESRGAN is considered effective for enhancing the resolution of digital images and videos and can be utilized in applications that require high visual quality.

PENERAPAN MODEL ESRGAN UNTUK UPSCALING CITRA DAN VIDEO DIGITAL Suhardi, Syahrul; Habi Talib, Emil Agusalim; Rachman, Fahrim Irhamna; Wahyuni, Titin; Faisal, Muhammad; S.Kuba, Muhammad Syafaat
PROGRESS Vol 18 No 1 (2026): April
Publisher : P3M STMIK Profesional Makassar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.56708/progres.v18i1.539

Low-resolution images and videos remain a common problem in various digital applications due to limited visual quality. Conventional interpolation-based upscaling methods often produce blurry results and lead to the loss of important texture details. This study aims to apply the Enhanced Super-Resolution Generative Adversarial Network (ESRGAN) to improve the resolution of digital images and videos. The dataset used consists of low-resolution images and videos that are processed through preprocessing, model training, and testing stages using the Google Colab environment. The ESRGAN model is trained to generate high-resolution images while preserving visual details and structural information. Model performance is evaluated using the Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), and visual comparison between images before and after the upscaling process. The results show that ESRGAN significantly improves the quality of images and videos compared to conventional interpolation methods, both quantitatively and qualitatively. Therefore, the application of ESRGAN is considered effective for enhancing the resolution of digital images and videos and can be utilized in applications that require high visual quality.

MONITORING DAN NOTIFIKASI REAL-TIME PERUBAHAN FILE PADA WEB SERVER MENGGUNAKAN WATCHDOG DAN TELEGRAM BOT SEBAGAI SISTEM PERINGATAN DINI Hasbir, Syahrul; Habi Talib, Emil Agusalim; Rachman, Fahrim Irhamna; Wahyuni, Titin; Faisal, Muhammad; S.Kuba, Muhammad Syafaat
PROGRESS Vol 18 No 1 (2026): April
Publisher : P3M STMIK Profesional Makassar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.56708/progres.v18i1.540

Web servers are critical infrastructures for delivering digital services and are highly vulnerable to unauthorized file changes that may threaten system security and service availability. However, many conventional monitoring systems still rely on periodic checking mechanisms, which often fail to provide timely detection of security incidents. This study aims to design and implement a real-time file change monitoring system on a web server using the Watchdog library and a Telegram Bot as an early warning mechanism. The research adopts an applied research method with an experimental approach. The system is developed using the Python programming language and evaluated in a local XAMPP-based web server environment, with the uploads directory selected as the monitoring target. Experimental results demonstrate that the proposed system is capable of detecting various file change events, including file creation, deletion, content modification, and file renaming, in real time without event loss. Notifications delivered via the Telegram Bot provide clear, timely, and actionable information to administrators. These findings indicate that the proposed event-driven monitoring system is effective and efficient in enhancing web server security and improving incident response capabilities.

MONITORING DAN NOTIFIKASI REAL-TIME PERUBAHAN FILE PADA WEB SERVER MENGGUNAKAN WATCHDOG DAN TELEGRAM BOT SEBAGAI SISTEM PERINGATAN DINI Hasbir, Syahrul; Habi Talib, Emil Agusalim; Rachman, Fahrim Irhamna; Wahyuni, Titin; Faisal, Muhammad; S.Kuba, Muhammad Syafaat
PROGRESS Vol 18 No 1 (2026): April
Publisher : P3M STMIK Profesional Makassar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.56708/progres.v18i1.540

Web servers are critical infrastructures for delivering digital services and are highly vulnerable to unauthorized file changes that may threaten system security and service availability. However, many conventional monitoring systems still rely on periodic checking mechanisms, which often fail to provide timely detection of security incidents. This study aims to design and implement a real-time file change monitoring system on a web server using the Watchdog library and a Telegram Bot as an early warning mechanism. The research adopts an applied research method with an experimental approach. The system is developed using the Python programming language and evaluated in a local XAMPP-based web server environment, with the uploads directory selected as the monitoring target. Experimental results demonstrate that the proposed system is capable of detecting various file change events, including file creation, deletion, content modification, and file renaming, in real time without event loss. Notifications delivered via the Telegram Bot provide clear, timely, and actionable information to administrators. These findings indicate that the proposed event-driven monitoring system is effective and efficient in enhancing web server security and improving incident response capabilities.

Student Emotion Recognition from Low-Quality Videos Using Multimodal Deep Learning ANDI MAWADDA TAIBA MAWADDA TAIBA; Rizki Yusliana Bakti; Muhammad Faisal; Muhammad Syafaat S. Kuba; Lukman Anas; Emil Agusalim H. T; Fahrim I. Rahman
JURNAL INFOTEL Vol 18 No 1 (2026): February
Publisher : LPPM INSTITUT TEKNOLOGI TELKOM PURWOKERTO

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.20895/infotel.v18i1.1523

Emotion recognition plays a critical role in intelligent e-learning systems by enabling adaptive feedback and timely pedagogical interventions based on students’ affective states. However, most existing approaches rely heavily on visual facial cues, which are highly vulnerable to real-world conditions such as low-resolution video, partial facial occlusion, poor lighting, and unstable network connections commonly encountered in online learning environments. These limitations significantly degrade the performance of unimodal deep learning models. To address this challenge, this study proposes a multimodal deep learning framework for student emotion recognition that is robust to low-quality and occluded video input. The proposed model integrates visual and audio modalities through a hybrid architecture, combining a lightweight CNN-based visual feature extractor with a BiLSTM-based speech emotion model. An attention-based fusion mechanism is employed to adaptively weight cross-modal features, allowing the system to compensate for degraded or missing visual information using complementary acoustic cues. Experimental evaluations are conducted using publicly available datasets representative of realistic online learning scenarios, including DAiSEE and RAVDESS, with additional augmentation to simulate varying levels of occlusion and video degradation. The results demonstrate that the multimodal approach consistently outperforms unimodal baselines, particularly under high occlusion conditions, while maintaining computational efficiency suitable for near real-time deployment. These findings confirm that multimodal fusion with attention mechanisms provides a more resilient and practical solution for emotion-aware e-learning systems operating under non-ideal input conditions

Optimasi Kinerja Arsitektur CNN Ringan Menggunakan Pendekatan Bayesian untuk Identifikasi Skrip Bima Dayang Aisyah; Muhammad Faisal; Lukman Anas; Abd Rakhim Nanda; Syadiah Nor Wan Shamsuddin; Muhammad Syafaat S. Kuba
Buletin Sistem Informasi dan Teknologi Islam (BUSITI) Vol 7, No 2 (2026)
Publisher : Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/busiti.v7i2.3462

Identifikasi aksara daerah penting untuk mendukung pelestarian budaya digital, namun masih terkendala keterbatasan dataset, kemiripan karakter, dan kebutuhan model yang efisien. Penelitian ini mengoptimasi arsitektur Lightweight CNN menggunakan Bayesian Optimization untuk identifikasi aksara Bima. Dataset terdiri atas 6.190 citra aksara Bima dalam 44 kelas, mencakup aksara Bima baru dan lama. Model menggunakan MobileNetV3-Large sebagai backbone dengan optimasi learning rate, dropout, batch size, dan konfigurasi fine-tuning melalui Tree-structured Parzen Estimator. Hasil eksperimen menunjukkan accuracy 93,06%, precision 92,26%, recall 92,55%, dan F1-score 91,91%, lebih unggul dibanding machine learning tradisional, CNN konvensional, dan beberapa CNN ringan modern. Target accuracy 90% dicapai pada trial keempat. Dengan 3.253.676 parameter dan waktu inferensi 63,35 ms per citra, model ini terbukti akurat, efisien, dan berpotensi diterapkan pada digitalisasi manuskrip serta OCR aksara daerah.

A Calibrated ROI-Aware Hybrid CNN-Transformer for Kidney Stone Presence Classification on Heterogeneous Axial CT Images Muh Ilham Akbar; Muhammad Faisal; Desi Anggreani; Abd Rakhim Nanda; Try Gustaf Said; Muhammad Syafaat S. Kuba
Buletin Sistem Informasi dan Teknologi Islam (BUSITI) Vol 7, No 2 (2026)
Publisher : Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/busiti.v7i2.3463

Batu ginjal merupakan penyebab umum nyeri pinggang akut, dan CT non-kontras menjadi standar referensi untuk mendeteksi kalkulus. Pada penelitian ini, istilah heterogen merujuk pada variasi protokol akuisisi antarrumah sakit, seperti perbedaan dosis radiasi, ketebalan irisan, rekonstruksi, dan bidang pandang, yang dapat mengubah tampilan citra serta menurunkan konsistensi pembacaan. Penelitian ini mengusulkan model hibrida CNN-Transformer yang sadar ROI (implisit) untuk klasifikasi keberadaan batu ginjal pada citra CT aksial heterogen. Arsitektur menggabungkan EfficientNet-B3, encoder Transformer ringan, dan Convolutional Block Attention Module (CBAM) tanpa anotasi ROI manual. Dataset terdiri dari 3.364 citra (1.577 batu, 1.787 non-batu) dengan pemisahan bertingkat 70/15/15. Evaluasi mencakup akurasi, presisi, sensitivitas, spesifisitas, F1, ROC-AUC, PR-AUC, inspeksi kalibrasi, dan audit Grad-CAM. Hasil menunjukkan bahwa penambahan Transformer meningkatkan kinerja dibanding baseline CNN, sedangkan CBAM menggeser profil kesalahan ke sensitivitas yang lebih tinggi. Varian Hybrid+Attention mencapai akurasi 0,9861, F1 0,9851, dan ROC-AUC 0,9967 pada set uji, dengan jumlah negatif palsu lebih rendah dibanding varian hibrida tanpa perhatian. Temuan ini menunjukkan potensi model sebagai alat bantu dokter untuk triase dan pembacaan awal yang lebih konsisten pada data lintas protokol, meskipun validasi eksternal, pemisahan berbasis pasien, dan metrik kalibrasi kuantitatif masih diperlukan sebelum klaim kesiapan klinis.

Explainable Fake News Detection in Indonesian Language Using IndoBERT and SHAP Muhammad Hasraddin Hasnan; Rizky Yusliana Bakti; Muhammad Faisal; Titik Khawa Abd Rahman; Nurnawaty Nurnawaty; Muhammad Syafaat S. Kuba; Titin Wahyuni
Buletin Sistem Informasi dan Teknologi Islam (BUSITI) Vol 7, No 2 (2026)
Publisher : Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/busiti.v7i2.3466

Perkembangan teknologi informasi dan media sosial telah mempercepat penyebaran berita palsu, sehingga diperlukan sistem deteksi yang akurat, andal, dan mudah diinterpretasikan. Penelitian ini bertujuan mengembangkan sistem deteksi fake news berbahasa Indonesia dengan mengintegrasikan IndoBERT sebagai model klasifikasi teks dan SHAP sebagai pendekatan Explainable Artificial Intelligence (XAI) untuk menjelaskan kontribusi kata terhadap hasil prediksi. Dataset diperoleh dari TurnBackHoax dan Kaggle, kemudian melalui tahapan preprocessing berupa cleaning text, filtering bahasa Indonesia, tokenisasi, serta penyeimbangan data menggunakan random oversampling pada data latih. Dari 5.347 data awal, diperoleh 4.980 data setelah filtering bahasa Indonesia, terdiri atas 3.613 data valid dan 1.367 data hoaks. Data dibagi secara stratifikasi dengan rasio 80% untuk pelatihan dan 20% untuk pengujian. Setelah oversampling, data latih menjadi seimbang dengan masing-masing 2.890 sampel per kelas. Hasil eksperimen menunjukkan bahwa model baseline TF-IDF dan Logistic Regression memperoleh akurasi 77%, sedangkan IndoBERT mencapai akurasi 87%, dengan precision 0,87, recall 0,95, dan F1-score 0,91 pada kelas hoaks. Visualisasi SHAP menunjukkan token penting yang memengaruhi klasifikasi. Hasil ini membuktikan bahwa integrasi IndoBERT dan SHAP efektif meningkatkan deteksi berita palsu sekaligus memberikan transparansi model.

A Hybrid K-Means and Neural Network for Enhancing Students’ Academic Performance Suriani Suriani; Muhammad Faisal; Darniati Darniati; Emil Agusalim H. T; Muhammad Syafaat S. Kuba; Swa Lee Lee; Nurdiansyah Nurdiansyah
Buletin Sistem Informasi dan Teknologi Islam (BUSITI) Vol 7, No 2 (2026)
Publisher : Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/busiti.v7i2.3467

Ketersediaan data pada Learning Management System (LMS) mendorong penerapan pembelajaran adaptif di pendidikan tinggi. Penelitian ini mengusulkan kerangka kerja hybrid berbasis kecerdasan buatan yang mengintegrasikan K-Means clustering dan Neural Network untuk profil mahasiswa berbasis perilaku dan prediksi kinerja akademik. Model divalidasi menggunakan Open University Learning Analytics Dataset yang mencakup data demografi, interaksi, dan performa akademik. Hasil menunjukkan akurasi sebesar 0,68 dan F1-score sebesar 0,66, melampaui metode dasar dengan stabilitas yang lebih baik. Clustering menghasilkan silhouette score 0,62 yang menunjukkan pemisahan kelompok yang jelas. Selain itu, sistem meningkatkan relevansi konten sebesar 27% dan menurunkan risiko putus studi sebesar 18%, dengan waktu inferensi rata-rata 0,85 detik. Temuan ini menunjukkan efektivitas kerangka dalam mendukung pembelajaran adaptif yang dipersonalisasi dan skalabel. Model hybrid yang diusulkan dapat mendukung pembelajaran adaptif melalui jalur belajar yang dipersonalisasi serta membantu perguruan tinggi melakukan intervensi dini terhadap mahasiswa berisiko berdasarkan pemantauan berbasis data.

Integrating multi-criteria decision making and public sentiment analysis for sustainable urban green space planning Muhammad Syafaat S. Kuba; Muhammad Faisal; Nurnawaty Nurnawaty; Titik Khawa Abdul Rahman; Andi Makbul Syamsuri; Muhyiddin AM Hayat; Rizki Yusliana Bakti
Bulletin of Electrical Engineering and Informatics Vol 15, No 2: April 2026
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/eei.v15i2.11168

Sustainable planning of green open spaces (GOS) requires decision-making models that combine expert evaluation with public input. This study proposes a novel hybrid framework that integrates multi-criteria group decision making (MCGDM) with public sentiment analysis to support community-based and data-driven urban planning. The workflow consists of evaluating 25 community-proposed GOS locations using stepwise weight assessment ratio analysis (SWARA) for criteria weighting and MABAC-BORDA for multi-criteria ranking, resulting in 11 feasible alternatives. To incorporate community perspectives, a term frequency-inverse document frequency-support vector machine (TF-IDF–SVM) classifier was applied to 1500 public comments, where SVM achieved the highest accuracy (0.80–0.96). The integrated approach improves ranking stability, reduces decision ambiguity, and strengthens alignment between expert judgment and community sentiment. This study contributes a transparent, participatory decision-support model that unifies MCGDM and sentiment analysis to enhance the effectiveness of sustainable GOS planning.

Title

Found 54 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 54 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 54 Documents
Search