Register: Jurnal Ilmiah Teknologi Sistem Informasi

Data Augmentation of Sperm Images Using Generative Adversarial Networks (WGAN-GP)

2025-09-09T20:47:43+00:00

This study analyzes the use of WGAN-GP for data augmentation in the analysis of sperm morphology. WGAN-GP has been the focus in this study for generating sperm microscopy images, which in turn aims to mitigate the problem of data scarcity in medical imaging. A heterogeneous dataset with mixed object categories was initially employed, leading to an FID score of 134, which in turn reflected a high incidence of mode collapse. For this reason, the dataset was divided into subcategories of Normal, Abnormal, and Non-Sperm identifications, with the scores of the subcategories being 59.19, 74.92, and 83.56, respectively, and showing better balanced model stability. This study's primary contribution is the use of WGAN-GP for the first time for sperm image data augmentation and the generation of more realistic synthetic images. Furthermore, this study illustrates the first understanding of the intricacies of data distribution's complexity and its effect on the model's performance, indicating the possibility of improvement using class-based techniques and sophisticated architectures for the generator. The innovation of this study is the application of WGAN-GP to sperm morphology datasets, improving image quality and the stability of the results, coupled with extensive model performance analysis and providing a further understanding of the field of medical image data augmentation.

Unsupervised Optimization of Boundary Information Based on the Coefficient of Variation to Improve Image Segmentation

2025-11-29T14:59:11+00:00

The automatic retrieval of boundary information from image objects suffers from the problem of under and over-segmentation, where the former leads to missed object detection, while the latter delivers an improper object shape. A method to optimize the automatic retrieval of complete and proper boundary information is proposed in this research based on an unsupervised approach. The strategy is to utilize the trade-off between the coefficient of variation from shape distribution against the mean of entropy contribution from segmented regions. This mechanism relies on the assumption that the segmentation result of a natural image contains a prominent main object representation with its details which are presented as a normal distribution of segmented regions. The research also enhances the entropy-based segmentation evaluation by redefining the computation of image entropy and segmentation entropy. The experiment shows that the proposed approach is capable of reducing over-segmentation by 57.20% compared to the existing algorithm, while at the same time reducing the consumption time by 85.26%. The empirical evaluation shows that the proposed approach delivers the highest accuracy among other evaluated methods. Qualitative validation based on groups of human observers shows that the proposed approach is the most desired algorithm for producing boundary information and measuring segmentation quality. These findings suggest that the trade-off between the mean of entropy contribution from the segmented regions and the coefficient of variation from shape distribution becomes an effective feature for unsupervised retrieval of boundary information.

AI Ethics in Indonesian Higher Education: A Systematic Review of Algorithmic Bias, Privacy, and Accountability

2026-03-01T01:24:16+00:00

Artificial Intelligence (AI) is increasingly integrated into higher education to enhance personalised learning, automate assessment, and improve institutional efficiency. However, its rapid adoption also raises ethical concerns related to algorithmic bias, data privacy, and accountability, particularly in Indonesia, where regulatory frameworks and digital infrastructure remain underdeveloped. Despite growing global discussions on AI ethics, limited studies have systematically examined these challenges within Indonesian higher education. This study analyses the ethical implications of AI integration by focusing on algorithmic fairness, data privacy, and governance accountability. Employing a narrative review approach supported by the PRISMA 2020 framework, this study systematically reviews 56 studies retrieved from the Scopus database between 2021 and 2025. Article screening was conducted using Covidence, while VOSviewer was utilised to identify research trends and thematic gaps. The findings reveal three major ethical concerns: (1) algorithmic bias in AI-driven assessment and admissions systems; (2) risks to data privacy and student surveillance associated with learning analytics; and (3) limited transparency and accountability in AI-based decision-making. The study further identifies significant gaps in Indonesia’s policy readiness and institutional governance. As its contribution, this study proposes a culturally grounded approach to AI governance and recommends the development of a National AI in Education Ethics Charter to support responsible and equitable AI integration in higher education.

Agroecological Zoning of Bangkalan Regency Using K-Means and HDBSCAN Based on Integrated Soil Fertility and Climate Features

2025-09-28T03:36:12+00:00

Agroecological heterogeneity poses challenges for agricultural planning in Bangkalan Regency, Indonesia. This study aimed to delineate agroecological zones by integrating soil fertility, climate, and topographic variables using K-Means clustering and Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN). A total of 11,000 geospatial observations obtained from Google Earth Engine were aggregated into 277 village-level units. The dataset included soil nutrients (nitrogen, phosphorus, and potassium), the Soil Quality Index, temperature, rainfall, humidity, elevation, and slope. Data preparation, modeling, and evaluation were performed as the primary methodological steps. Min-Max Scaling was applied to normalize the data. The optimal K-Means configuration (K = 3) achieved a Silhouette Score of 0.2668, an Inertia value of 294.5529, and a Calinski-Harabasz Index (CHI) of 75.8821. The resulting clusters were classified as High-Potential (52 villages), Moderate-Potential (142 villages), and Environmental-Constraint (83 villages) zones. HDBSCAN was used to validate clustering patterns and detect environmental anomalies. The optimal HDBSCAN configuration identified two density-based clusters and five noise villages. These villages showed exceptionally high nitrogen, phosphorus, and Soil Quality Index values, indicating localized agroecological hotspots. The integration of K-Means and HDBSCAN offers a comprehensive framework for agricultural planning, resource allocation, and sustainable land management.

Enhanced Rice Yield Prediction in Indonesia with Integrated Climate and Agricultural Data Using Decision Tree Regression

2025-01-16T06:55:30+00:00

Rice is central to Indonesia’s food security, yet provincial yields are highly sensitive to climatic variability, making reliable forecasting essential for national planning and improving farmer welfare. Most prior Indonesian yield models rely on rainfall and temperature data and omit sunlight exposure duration, which is a limiting factor for photosynthesis in the humid tropics where solar radiation, not temperature, often constrains productivity. This study develops a province-level rice yeild prediction model based on Decision Tree Regression (DTR) that integrates climate data from the Meteorology, Climatology, and Geophysics Agency (BMKG) with agricultural statistics from the Central Statistics Agency (BPS). The dataset comprises data from 34 provinces covering the period from 2018 to 2023 (204 province-year observations), with year, harvested land area, rainfall, and sunlight exposure duration as predictors and rice production as the target variable. The dataset was partitioned into training and testing subsets using an 80:20 ratio. Hyperparameter tuning was performed using k-fold cross-validation, and model performance was performed using Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination (R²). The model attained an average RMSE of 93,046 tons (MAE ≈ 51,824 tons; MAPE ≈ 6.8% on the 2023 hold-out year). A key finding is that absolute RMSE is strongly scale-dependent. When evaluated in relative terms, the highest-producing Java provinces were among the most accurately predicted (relative RMSE ≈ 2.3–2.5%), whereas several small or structurally volatile provinces showed relative errors above 40%. The study contributes to the literature by providing (i) the explicit integration of sunlight exposure into Indonesian rice yield modeling, (ii) a province-disaggregated error analysis that reframes accuracy in scale-independent terms, and (iii) an interpretable decision-support tool for food-policy stakeholders such as Bulog and the Ministry of Agriculture.