cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Yuhefizar
Contact Email
jurnal.resti@gmail.com
Phone
+628126777956
Journal Mail Official
ephi.lintau@gmail.com
Editorial Address
Politeknik Negeri Padang, Kampus Limau Manis, Padang, Indonesia.
Location
,
INDONESIA
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
Published by Ikatan Ahli Informatika Indonesia
ISSN : 25800760     EISSN : 25800760     DOI : https://doi.org/10.29207/resti.v2i3.606
Core Subject : Science, Engineering,
Computer Science & IT Engineering
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) dimaksudkan sebagai media kajian ilmiah hasil penelitian, pemikiran dan kajian analisis-kritis mengenai penelitian Rekayasa Sistem, Teknik Informatika/Teknologi Informasi, Manajemen Informatika dan Sistem Informasi. Sebagai bagian dari semangat menyebarluaskan ilmu pengetahuan hasil dari penelitian dan pemikiran untuk pengabdian pada Masyarakat luas dan sebagai sumber referensi akademisi di bidang Teknologi dan Informasi. Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) menerima artikel ilmiah dengan lingkup penelitian pada: Rekayasa Perangkat Lunak Rekayasa Perangkat Keras Keamanan Informasi Rekayasa Sistem Sistem Pakar Sistem Penunjang Keputusan Data Mining Sistem Kecerdasan Buatan/Artificial Intelligent System Jaringan Komputer Teknik Komputer Pengolahan Citra Algoritma Genetik Sistem Informasi Business Intelligence and Knowledge Management Database System Big Data Internet of Things Enterprise Computing Machine Learning Topik kajian lainnya yang relevan
Arjuna Subject : Ilmu Komputer (semua kategori) - Ilmu Komputer (Lain-Lain)
Articles 1,185 Documents
 115   116   117   118   119 
Comparative Analysis of CNN Architectures and Optimizers for ALL Classification Sugiyarto Surono; Fanita Damayanti; Frenky Riski Gilang Pratama; Attarik Mohammad; Maretta Mia Audina; Marfungah Wati
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 10 No 4 (2026): August 2026
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v10i7506

Abstract

Deep learning has become an important approach for medical image analysis, particularly for automated image classification. Among its various architectures, Convolutional Neural Networks (CNNs) are frequently employed to extract and distinguish visual patterns from medical images. However, the performance of a CNN may vary depending on the optimization algorithm used during the training process. This study investigates the effect of four optimization methods on different CNN architectures for the classification of Peripheral Blood Smear (PBS) images. A publicly available Kaggle dataset comprising 3,242 images and four diagnostic categories was used in the experiments. Six architectures, namely ResNet50V2, Xception, EfficientNetB2, EfficientNetB4, EfficientNetV2-S, and EfficientNetV2-M, were evaluated in combination with Adam, AdamW, NAdam, and RMSprop. Before training, the images underwent resizing, normalization, and augmentation procedures. The dataset was subsequently partitioned into training, validation, and test subsets using proportions of 80%, 10%, and 10%, respectively. Classification performance was analyzed using accuracy, loss, and confusion matrix-based evaluation. Among all tested combinations, Xception with RMSprop produced the best result, obtaining 99.69% accuracy and a loss value of 0.0194. Conversely, the lowest result was observed for ResNet50V2 with RMSprop, which achieved 93.85% accuracy. The experimental findings indicate that the optimizer can substantially influence CNN classification performance, and the effectiveness of an optimization method is dependent on the architecture with which it is combined.
Enhancing Software Defect Prediction Performance using NR-Clustering SMOTE to Address Class Imbalance Hairani Hairani; Muhamad Masjun Efendi; Gede Yogi Pratama; Rahayun Amrullah Husaini; M.Khaerul Ihsan
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 10 No 4 (2026): August 2026
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v10i7558

Abstract

Software defect detection is important to prevent system failures and increased maintenance costs. However, the complexity of modern software makes manual testing inefficient, so machine learning approaches are used. The main challenge of this approach is data imbalance, where defective cases are far fewer, causing the model to overlook the minority class and reducing detection capability, even though accuracy appears high. This study aims to address class imbalance in software defect detection by applying the NR-Clustering SMOTE method to improve machine learning performance—the classification methods using Random Forest. NR-Clustering SMOTE not only oversamples the minority class but also incorporates a noise-reduction mechanism to remove minority data that may degrade classification performance. The results show that NR-Clustering SMOTE improves the performance of Random Forest compared with the original data, SMOTE, and NR-Modified SMOTE across all evaluation metrics, namely accuracy, recall, and F1-score. These findings indicate that integrating noise reduction and SMOTE-based data balancing using Manhattan distance within each cluster produces a more representative data distribution, thereby improving the model’s ability to classify software defect cases more accurately. Therefore, this study confirms that NR-Clustering SMOTE effectively improves Random Forest performance for software defect detection compared with existing approaches.
An Improved Spectral Residual Algorithm for Sparrow Pest Detection in Rice Field CCTV Video Raja Ayu Mahessya; Yuhandri Yuhandri; Rini Sovia
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 10 No 4 (2026): August 2026
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v10i4.7775

Abstract

Agricultural pest surveillance in open field environments presents a persistent challenge for food security, particularly when early detection is limited by the inadequacy of traditional monitoring methods. Existing Spectral Residual (SR) algorithms, while capable of saliency-based object detection, exhibit notable limitations in adapting to dynamic agricultural conditions resulting in degraded segmentation quality and elevated false-positive rates. This study proposes an improved Spectral Residual algorithm that independently processes Red, Green, and Blue (RGB) channels via Fast Fourier Transform (FFT), preserving richer spectral information that is otherwise lost in conventional grayscale-based approaches. The proposed pipeline integrates spatial saliency maps with temporal motion cues extracted through frame-difference analysis, suppressing static background noise while reinforcing responses in regions exhibiting genuine movement. Adaptive binarization and Support Vector Machine (SVM) classification are further incorporated to strengthen object segmentation robustness across diverse lighting and background conditions. The model was evaluated on 136 frames extracted from 2MP Hikvision CCTV footage installed at rice field sites, benchmarked against the classical Spectral Residual baseline using accuracy, precision, recall, and F1-score. The proposed method achieved an accuracy of 98.15%, precision of 91.67%, recall of 98.98%, and F1-score of 95.18% — representing substantial improvements over the baseline (74.00%, 70.00%, 65.00%, and 69.00%, respectively). These results demonstrate that multi-channel spectral processing combined with motion-aware saliency yields a robust and computationally efficient saliency framework whose low algorithmic complexity makes it a strong candidate for real-time sparrow pest detection in CCTV-based agricultural monitoring systems.
Composite Loss-Enhanced DeepLabV3+ for Crack Segmentation in Photogrammetric Mining Imagery Yesri Elva; Iskandar Fitri; Syafri Arlis
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 10 No 4 (2026): August 2026
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v10i4.7807

Abstract

Crack detection in photogrammetric mining imagery remains challenging due to severe class imbalance, thin and irregular crack morphology, and complex soil textures that hinder accurate boundary delineation. This study proposes a DeepLabV3+ semantic segmentation model with a ResNet-50 backbone trained using a Composite Loss that integrates Focal Tversky Loss, Weighted Cross-Entropy Loss, and Boundary Loss into a unified multi-objective framework. The dataset comprised 425 photogrammetric crack images acquired from open-pit mining sites, divided into 380 training and 45 test images. Data augmentation, including horizontal flip, vertical flip, channel shuffle, and random brightness and contrast adjustment, was applied to improve model generalization under varying image conditions. The proposed configuration was evaluated against Default Loss, Binary Cross-Entropy Loss, and Weighted Cross-Entropy Loss. The model achieved a Global Accuracy of 0.9808, Mean Intersection over Union of 0.7890, and Mean Boundary F1-Score of 0.8642, outperforming all baseline configurations. Boundary fidelity showed the largest improvement, with a Mean BFScore gain of 27.3% over Default Loss, demonstrating the effectiveness of Boundary Loss for precise crack contour delineation. The trained model was also deployed in a GUI-based system supporting automated crack analysis and PDF report generation for practical field application.
Enhancing Transparency in Income Prediction: A SHAP-based Explainable AI Approach with Random Forest Marsani Asfi; Budi Warsito; Adi Wibowo
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 10 No 5 (2026): October 2026 (In progress)
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v10i5.7461

Abstract

Explainable Artificial Intelligence (XAI) has gained increasing importance as modern machine learning models operate as opaque black-box systems, limiting transparency and trust in automated decision-making. While most XAI frameworks merely focus on static post-hoc visualizations, this study goes beyond passive diagnostics by introducing an algorithmic recourse framework for income prediction. By integrating SHAP (SHapley Additive exPlanations) with an optimized Random Forest classifier trained on the UCI Adult Income dataset, we partition socio-economic attributes into actionable features (e.g., weekly working hours, education level) and immutable demographic variables (e.g., age, sex). The dataset was preprocessed using missing-value handling, categorical encoding, class balancing via SMOTE, and feature scaling. Hyperparameter tuning optimization determined that the configuration of 150 trees, a maximum depth of 15, and a minimum split size of 15 yielded peak training accuracy rates of 85.49%, 86.13%, and 86.05% respectively. The Random Forest model achieved strong predictive performance, yielding an accuracy of 84.68%, precision of 85.15%, recall of 84.68%, F1-score of 84.88%, and a high ROC-AUC of 90.95%. Global SHAP analysis identified relationship, marital status, and education as the primary drivers of income classification. Crucially, local instance-level counterfactual optimization demonstrates that strategically adjusting actionable features can successfully neutralize negative SHAP momentum and flip unfavorable low-income predictions (≤ 50K) into positive outcomes (>50K) while holding immutable traits perfectly constant. This integration demonstrates that combining Random Forest with an algorithmic recourse pipeline transforms AI from a descriptive transparency tool into an active, prescriptive mechanism that empowers data-driven decision-making and human solutions in socio-economic applications.
Leveraging Assistive Image Captioning on VizWiz through Multi-Caption Supervision Fauzan Firdaus; Ema Rachmawati; Angel Metanosa Afinda; Edward Ferdian; Muhammad Maulana Yusuf
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 10 No 5 (2026): October 2026 (In progress)
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v10i5.7472

Abstract

Most image captioning datasets provide several human captions for the same image, yet model training is frequently carried out in a way that effectively treats each image as having only one caption. Our study revisits this mismatch by examining the impact of training with multiple captions per image on caption generation performance. For this purpose, we utilized the VizWiz-Captions dataset, which reflects a real assistive scenario with images taken by blind or visually impaired users with five reference captions per image. We evaluated two conventional CNN-based encoder–decoder models, namely CNN–LSTM and CNN–GRU. In both models, the visual encoder was fixed using InceptionV3 image features, while the recurrent decoder was varied between LSTM and GRU. We performed experiments by using training data by varying the number of captions used per image from one to five (k = 1–5), while keeping the overall pipeline unchanged. The models were evaluated using standard multi-reference metrics (BLEU-1 to BLEU-4, CIDEr, and SPICE). Across both baselines, performance generally improves when more captions are included in training, although the gains are not perfectly monotonic for every metric. The strongest results are observed at k = 5. For CNN–LSTM, the best setting achieved BLEU-4 = 0.175, CIDEr = 0.206, and SPICE = 0.120. On the other hand, CNN–GRU achieved BLEU-4 = 0.192, CIDEr = 0.226, and SPICE = 0.119. These findings suggest that, even without architectural changes, simply leveraging the available caption diversity during training can provide a practical boost for VizWiz-style captioning tasks.
Modified Quantization Strategy for Enhanced Lossless RAW Audio Compression Tonny Hidayat; Hendra Kurniawan; Hastari Utama; Dewi Anisa Istiqomah; Joko Dwi Santoso
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 10 No 5 (2026): October 2026 (In progress)
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v10i5.7512

Abstract

The amount of audio data required in today's world is rapidly expanding, posing a significant problem and challenge in terms of media storage and transfer. RAW audio data compression has grown in popularity due to its ability to lower file size without sacrificing quality. To anticipate the complexity of increasing the lossless audio compression ratio, this study proposes a lossless audio data compression algorithm that theoretically increases its efficiency ratio, thus providing multiple compression that combines linear prediction methods, integer wavelet transforms, and as a back-end process using Huffman code for binary output coefficients. The combined transformation produced employs a new prediction form and coefficient process, which is a novelty and the primary contribution of the coding of this algorithm structure. Based on the test results in many audio categories, the obtained results demonstrate that the proposed algorithm surpasses the performance of previous studies. The compression ratio increased from 2% to 5%. The algorithm discovered in this work is ideal for compressing audio with complex patterns and high irregularity using a quick compression procedure.
Comparative Analysis of CNN and Vision Transformer for Alzheimer’s Prediction via Cross-Dataset Validation Jasman Pardede; Hilman Syawaluddin; Muhammad Rama Reyswara; Asep Nana Hermana; Marisa Premitasari
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 10 No 5 (2026): October 2026 (In progress)
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v10i5.7532

Abstract

The automated detection of Alzheimer's Disease (AD) is increasingly critical for global healthcare systems, as conventional diagnostic pathways remain subjective, resource-intensive, and poorly scalable to resource-constrained clinical settings such as those prevalent in Indonesia. This study presents a rigorous comparative evaluation of three deep learning architectures - a custom Convolutional Neural Network (CNN), ResNet-50, and Vision Transformer (ViT/B-16)- for binary AD classification from Axial T1-weighted MRI. To assess genuine generalizability rather than in-distribution performance, a cross-dataset validation strategy was adopted: models were trained on the public dataset, i.e., the Falah/HuggingFace dataset, and then evaluated on the private dataset collected from Rumah Sakit Hasan Sadikin (RSHS), Bandung, Indonesia. To eliminate hyperparameter bias as a confounding variable, a systematic Grid Search across 27 configurations was conducted, combining three optimizers (Adam, SGD, and AdamW) with three learning rates (0.001, 0.0001, and 0.00001) per architecture. ResNet-50 with SGD at lr = 0.0001 emerged as the optimal configuration, achieving the highest internal accuracy of 0.989 (F1 = 0.980) and retaining a robust external accuracy of 0.823 (F1 = 0.825), surpassing all six prior benchmarked studies, none of which conducted external clinical validation. ViT exhibited severe domain sensitivity, collapsing from 0.985 internal to 0.617 external accuracy, a 37.36% degradation attributable to attention miscalibration under distribution shift. The Custom CNN achieved a moderate external accuracy of 0.653, outperforming ViT by 5.83% and confirming the structural advantage of convolutional inductive biases in limited-data regimes. These findings indicate that CNN-based architectures offer a safer, computationally more efficient, and clinically applicable solution for AD diagnosis compared to Transformer models, which rely heavily on large amounts of data.
Optimized CNN-Based Transfer Learning Approach with ReduceLROnPlateau for Early Potato Leaf Disease Classification Andi Sunyoto; Yoga Pristyanto; Anggit Ferdita Nugraha
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 10 No 5 (2026): October 2026 (In progress)
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v10i5.7533

Abstract

Potato leaf disease classification is important for supporting early disease detection and improving crop management. This study proposes an optimized CNN-based transfer learning approach for classifying potato leaf images into three classes: Early Blight, Healthy, and Late Blight. Six pre-trained architectures—InceptionV3, Xception, DenseNet121, EfficientNetB0, ResNet50, and MobileNet were evaluated using ImageNet weights. Each model was modified with a custom classification head consisting of Global Average Pooling, a dense layer with 1,024 neurons, dropout, and a Softmax output layer. The models were evaluated using 5-fold stratified cross-validation. A frozen baseline configuration was first established, followed by fine-tuning configurations that varied learning rate, dropout rate, number of unfrozen layers, and the use of ReduceLROnPlateau. Data augmentation was applied only to the training subset in each fold. Performance was assessed using accuracy, AUC, validation loss, confusion matrix, ROC curves, training time, inference time, and GFLOPs. The results showed that optimized fine-tuning improved the average accuracy from 0.9876 to 0.9975 and reduced the average validation loss from 0.0316 to 0.0025. The Xception with the HP3 configuration achieved the best performance, with a mean accuracy of 0.999444 and an AUC of 0.999997. However, EfficientNetB0 achieved a high accuracy of 0.998333 with only 0.80 GFLOPs per image, making it more suitable for efficiency-oriented deployment. These findings demonstrate that fine-tuned transfer learning combined with adaptive learning rate scheduling is effective for potato leaf disease classification. Because the optimized configurations differ from the frozen baseline in the number of unfrozen layers, the dropout rate, the learning rate, and the use of ReduceLROnPlateau simultaneously, the reported gain reflects the combined effect of these components rather than the isolated contribution of ReduceLROnPlateau. The results were obtained on PlantVillage images acquired under controlled conditions and should therefore be regarded as an upper reference rather than an estimate of performance under field condition.
Real-Time Multi-Face Detection for Automated Attendance Management Using YOLOv8 Danang Arbian Sulistyo; Lia Farokhah; Emilsa T Bantug; Rachel T Alegado
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 10 No 5 (2026): October 2026 (In progress)
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v10i5.7544

Abstract

Accurate and efficient attendance management remains challenging in classroom environments because face-based systems are sensitive to illumination, pose variation, occlusion, and simultaneous student entry. This study proposes a YOLOv8n-based attendance management framework that integrates entropy-guided frame selection, automated face annotation, classroom-oriented augmentation, closed-set student identification, and temporal attendance validation. The dataset comprised 1,800 frames from six enrolled students and was expanded to 12,600 training images through augmentation applied exclusively to the training partition. The model achieved a precision of 0.620, recall of 0.692, mAP@0.5 of 0.566, and mAP@0.5:0.95 of 0.497 on the held-out test set. Model-only inference required 1.344 ms per image under the specified computational configuration, whereas the complete attendance pipeline, including frame acquisition, preprocessing, detection, rendering, and attendance logging, required an average of 0.122 s per frame, corresponding to approximately 8.2 FPS. Temporal confirmation and check-in/check-out rules prevented short-lived or repeated detections from generating duplicate attendance records. These findings demonstrate the feasibility of the framework as a proof of concept for controlled, closed-set classroom attendance involving a limited cohort. Further evaluation with more participants, independent recording sessions, and cross-environment data is required before broader deployment can be claimed.

Page 117 of 119 | Total Record : 1185

 115   116   117   118   119 

Filter by Year

2017 2026


Reset
Filter By Issues
All Issue Vol 10 No 5 (2026): October 2026 (In progress) Vol 10 No 4 (2026): August 2026 Vol 10 No 3 (2026): June 2026 Vol 10 No 2 (2026): April 2026 Vol 10 No 1 (2026): February 2026 Vol 9 No 6 (2025): December 2025 Vol 9 No 5 (2025): October 2025 Vol 9 No 4 (2025): August 2025 Vol 9 No 3 (2025): June 2025 Vol 9 No 2 (2025): April 2025 Vol 9 No 1 (2025): February 2025 Vol 8 No 6 (2024): December 2024 Vol 8 No 5 (2024): October 2024 Vol 8 No 4 (2024): August 2024 Vol 8 No 3 (2024): June 2024 Vol 8 No 2 (2024): April 2024 Vol 8 No 1 (2024): February 2024 Vol 7 No 6 (2023): December 2023 Vol 7 No 5 (2023): October 2023 Vol 7 No 4 (2023): August 2023 Vol 7 No 3 (2023): Juni 2023 Vol 7 No 2 (2023): April 2023 Vol 7 No 1 (2023): February 2023 Vol 6 No 6 (2022): Desember 2022 Vol 6 No 5 (2022): Oktober 2022 Vol 6 No 4 (2022): Agustus 2022 Vol 6 No 3 (2022): Juni 2022 Vol 6 No 2 (2022): April 2022 Vol 6 No 1 (2022): Februari 2022 Vol 5 No 6 (2021): Desember 2021 Vol 5 No 5 (2021): Oktober2021 Vol 5 No 4 (2021): Agustus 2021 Vol 5 No 3 (2021): Juni 2021 Vol 5 No 2 (2021): April 2021 Vol 5 No 1 (2021): Februari 2021 Vol 4 No 6 (2020): Desember 2020 Vol 4 No 5 (2020): Oktober 2020 Vol 4 No 4 (2020): Agustus 2020 Vol 4 No 3 (2020): Juni 2020 Vol 4 No 2 (2020): April 2020 Vol 4 No 1 (2020): Februari 2020 Vol 3 No 3 (2019): Desember 2019 Vol 3 No 2 (2019): Agustus 2019 Vol 3 No 1 (2019): April 2019 Vol 2 No 3 (2018): Desember 2018 Vol 2 No 2 (2018): Agustus 2018 Vol 2 No 1 (2018): April 2018 Vol 1 No 3 (2017): Desember 2017 Vol 1 No 2 (2017): Agustus 2017 Vol 1 No 1 (2017): April 2017 More Issue