Asian Journal of Research in Computer Science
Not a member yet
792 research outputs found
Sort by
Estimating Nutritional Composition from Food Volume Via Deep Learning-Based Depth and Segmentation Models
No full textNutrition plays a critical role in human health, with a balanced diet being essential for preventing non-communicable diseases, enhancing immune function, and improving quality of life. However, dietary imbalances contribute to significant global health issues, including obesity and malnutrition, which have far-reaching economic and health consequences. This research aims to address these challenges by developing a method for estimating the nutritional content of food items from a single 2D image. Our approach integrates a U-Net architecture with a ResNet18 encoder for depth prediction and employs FoodSAM for precise food segmentation. These components enable the calculation of food volume and mass, which are then used to estimate nutritional content based on the USDA database. Experimental results show that our model achieves a mean relative error (MRE) ranging from 11.18% to 50.35% for individual food items. Furthermore, our method maintains consistent mass predictions across various scenarios, including complex food combinations. This method demonstrates robustness in handling foods with diverse shapes and colors, providing a solid foundation for practical dietary tracking applications. By enabling nutritional monitoring, our approach has the potential to support public health initiatives and promote healthier lifestyles
Enhancing Emotional and Cultural Retention in Ancient Chinese Poetry Translation Using BERT
No full textThis research investigates the use of the BERT (Bidirectional Encoder Representations from Transformers) model to enhance the translation of ancient Chinese poetry into English, with particular focus on overcoming the unique challenges posed by this literary genre. Ancient Chinese poetry is renowned for its intricate rhythm, tonal variations, and dense symbolism, all deeply embedded within cultural and historical contexts. These features create significant difficulties in translation, as maintaining the original’s lyrical quality, emotional depth, and rich cultural references often proves elusive using conventional methods. Through a meticulous process of curating and annotating a dataset comprising ten masterworks of the Tang and Song dynasties, this study explores whether BERT’s advanced contextual understanding and bidirectional encoding can more effectively convey the nuanced emotional and cultural content embedded in the source texts. The results demonstrate that BERT substantially improves both affective resonance and the preservation of cultural imagery in English translations, achieving a higher level of fluency and authenticity. This work not only advances the capabilities of machine translation for complex literary forms but also underscores the potential of cutting-edge AI to foster deeper cross-cultural understanding and greater global appreciation of China’s rich poetic heritage
From Framework to Practice: Barriers and Enablers to RMF Adoption in Mid-sized Enterprises
No full textThis study investigates the barriers and enablers influencing the adoption of the Risk Management Framework (RMF) in mid-sized software development enterprises through a quantitative research design. Four datasets were analyzed: the Stack Overflow Developer Survey, ENISA Threat Landscape Reports, OpenSSF Secure Practices Survey, and Verizon’s Data Breach Investigations Report (DBIR). The study employed descriptive statistics to explore awareness patterns, exploratory factor analysis to identify latent barriers, binary logistic regression to model enablers, and independent t-tests to evaluate security performance outcomes. Findings reveal that RMF awareness is highest among DevOps Engineers and Software Architects (73.3%), yet practical adoption remains limited, with only 34.4% of Asian firms implementing RMF-aligned practices. Lack of leadership support emerged as the strongest barrier (loading = -0.58), while leadership endorsement was the most significant enabler (β = 1.1671, p < .001). Organizations that adopted RMF demonstrated 40% faster threat detection and 66% faster incident response compared to non-adopters. The study highlights the strategic role of leadership commitment, workforce training, and CI/CD pipeline integration in promoting RMF adoption. It recommends contextual RMF toolkits tailored for Agile environments, executive-level cybersecurity briefings, and role-specific certifications to accelerate adoption. By strengthening operational resilience and regulatory compliance, RMF offers a scalable pathway for mid-sized firms navigating the complexities of modern software security risks
Solving Quadratic Equations with Complex Coefficients & Determining the Square Root for any Complex Number
No full textThis paper presents a definitive solution for expressing the second root (z1/2) of any complex number z as a function of z and Δ1/2 as a function of Δ in the context of quadratic equations with complex coefficients. This research has resulted in two direct algorithms and associated software. These algorithms are not only accurate but also exceptionally fast, providing a powerful tool for efficiently determining the second root of any complex number and solving quadratic equations with complex coefficients. This innovative approach addresses a long-standing challenge in the field, significantly contributing to the accuracy and speed of calculations in complex coefficient scenarios either manually or using a programmable computing machine such as smart phones or computers
Web Technology and Cloud Computing in Enterprise System: The Role of AI for Digital Marketing
No full textThis study explores the transformative impact of Artificial Intelligence (AI) and cloud computing on digital marketing and organizational operations. As businesses increasingly leverage AI technologies, they gain the ability to create personalized marketing strategies that enhance customer engagement and foster deeper connections with audiences. Cloud computing provides the necessary infrastructure to support these advancements, allowing for scalable solutions that improve operational efficiency and facilitate real-time data analysis. The research addresses the emerging challenges organizations face, such as data privacy issues and the need for skilled talent, while highlighting the critical role of these technologies in driving innovation and growth. By integrating AI with cloud-based solutions, businesses can optimize marketing efforts, streamline operations, and make informed decisions that enhance overall performance. The findings suggest that the future of digital marketing will heavily rely on the effective combination of AI and cloud computing, enabling organizations to remain competitive in an increasingly digital landscape. This convergence not only represents a significant shift in marketing practices but also offers new opportunities for success across various sectors. Enterprise systems (ERP) are vital for streamlining business processes by integrating functions such as finance, human resources, and supply chain management into a unified platform. Technological advancements like cloud computing, digital marketing, and machine learning are transforming these systems, enhancing operational efficiency, customer engagement, and data-driven decision-making. The rise of web technologies further improves user experience, while machine learning helps enterprises analyze data and optimize processes. These innovations create new opportunities and challenges for businesses. Machine learning, in particular, is revolutionizing healthcare by predicting illnesses through data analysis, while data visualization techniques help manage the ever-growing volumes of big data. Similarly, disruptive technologies, such as AI and cloud computing, are reshaping industries like banking and project management by driving digital transformation, improving customer experiences, and enhancing efficiency. However, challenges such as data privacy, security, and integration remain central in adopting these technologies across sectors
Efficient Fine-grained Bird Classification Via Dynamic Ensembles and Active Learning
No full textFine-grained bird species classification is a challenging task, mainly because the differences between species are often very subtle, and there is typically limited labeled data available for each species. This paper introduces a novel approach that combines multiple deep-learning models with a smart labeling strategy to address these issues. The method uses a dynamic ensemble of three advanced models: ResNet50, which excels in recognizing visual patterns; EfficientNet-B3, known for its efficiency and performance balance; and Swin Transformer, a newer model that captures both local and global features in images. Instead of relying on each model’s prediction separately, an adaptive mechanism (a multi-layer perceptron or MLP) weighs their predictions per sample, allowing the system to make smarter and more accurate decisions. Overall, the approach stands out due to its combination of a dynamic ensemble of models and an active learning strategy. This not only boosts performance but also makes the best use of limited labeled data, offering a more efficient way to tackle complex classification tasks like bird species identification. The results demonstrate that with fewer labels, the method can still achieve high accuracy and generalize well, providing a powerful tool for fine-grained classification in scenarios where data is scarce
Predictive Data Analysis in Forecasting Patient Health Outcomes Using Machine Learning Algorithms
No full textThis study investigates the effectiveness of machine learning algorithms, particularly an optimized eXtreme Gradient Boosting (XGBoost) model, in predicting 30-day ICU readmissions using the MIMIC-III dataset comprising over 40,000 critical care records. Through a quantitative approach, supervised learning models were developed, evaluated, and compared, incorporating hyperparameter tuning via grid search with five-fold cross-validation. SHapley Additive exPlanations (SHAP) were employed for interpretability and to identify key predictors. Results showed that the optimized XGBoost achieved an AUC-ROC of 0.892 and F1-score of 0.826, outperforming Logistic Regression and Random Forest. Integration simulations revealed a 3.2-second latency, 89% success rate, and 0.18 workflow disruption index, validating real-time deployment potential. Recommendations include enforcing interpretability standards, enhancing EHR interoperability, promoting clinician algorithm literacy, and ensuring dataset representativeness for predictive equity. These findings highlight the pivotal role of interpretable AI in supporting proactive, equitable, and data-driven clinical decision-making
3ConFA: A Robust Feature Aggregation Framework for High-dimensional Data Optimization
No full textHigh-dimensional datasets pose significant challenges in machine learning, including overfitting, increased computational complexity, and reduced model interpretability. In response to these challenges, this paper introduces the Three Conditions for Feature Aggregation (3ConFA) framework, a feature selection approach based on ensembles that seeks to reduce dimensionality without reducing predictive performance. The 3ConFA framework integrates three key techniques; Chi-square (χ²) test, Information Gain (IG), and Decision Tree-based Recursive Feature Elimination (DT-RFE), and applies stringent conditions to select only the most relevant features. A feature is retained if and only if it satisfies all three conditions: (1) IG score ≥ mean IG threshold, (2) χ² score ≥ mean χ² threshold, and (3) DT-RFE importance score = 1.
Experimental evaluation on ten benchmark datasets demonstrates the effectiveness of the framework, achieving feature reductions of up to 98.75% (e.g., Dexter dataset) while improving classification accuracy (e.g., Madelon: 78% → 85%). Performance metrics (precision, recall, F1-score) were consistently enhanced after feature selection, confirming that 3ConFA enhances model generalization without tradeoff in critical information. The framework’s adaptability across diverse datasets highlights its potential for applications in fraud detection, healthcare, and image classification.
This study contributes a structured, condition-driven feature aggregation approach that outperforms traditional filter and wrapper methods. Future work may explore adaptive thresholding and integration with deep learning models
Organization of Cross-functional (DevOps·PM) Teams in e-commerce: Processes for Teamwork
No full textThe article investigates how to organize cross-functional e-commerce teams by integrating DevOps and traditional project management (PM) processes with systems engineering and quality assurance practices. Drawing on empirical data, classic Waterfall, Agile, and DevOps models, and a review of advanced techniques, it proposes a unified team structure comprising a Product Owner, Business Analyst, UX/UI Designer, Tech Lead, Developers, QA Specialists, and Systems Analysts. It demonstrates how Agile workflows and CI/CD pipelines align with PMBOK-based PM processes, how incremental system architecture and requirements traceability ensure precise tracking from conception to deployment, and how automated testing and DevSecOps practices balance release speed with platform reliability. Tables detail role distribution and a DevOps–PM alignment matrix, clearly illustrating which DevOps activities correspond to each PMBOK process group. These findings offer a methodological foundation for building high-performance e-commerce teams that accelerate time-to-market without sacrificing quality or scalability, and will be of interest to researchers of organizational behavior and sociotechnical systems, as well as to DevOps and project managers focused on optimizing continuous integration, delivery, and product management
The Strategic Role of AI in Enhancing National Cybersecurity Frameworks
No full textThis study examines the key directions of artificial intelligence (AI) development and its influence on cyber defense strategies in national security. An analysis is conducted on current trends in the implementation of machine learning (ML) in cyberattack detection and prevention systems, demonstrating how AI technologies are transforming traditional approaches to forming National Cybersecurity Strategies (NCSS). The study identifies challenges and opportunities associated with the application of AI in protecting critical infrastructure, optimizing cyber incident management processes, and improving interagency coordination. Ethical considerations, legal regulation, and workforce development are discussed as critical aspects in the large-scale adoption of AI systems in government administration. Recommendations are proposed for enhancing NCSS through centralized management, adaptation of regulatory frameworks, strengthening public-private partnerships, and workforce capacity building. The study concludes by justifying the need for international cooperation, the development of global standards, and coordinated response mechanisms to cyber threats based on advanced AI solutions. This study demonstrates the critical role of artificial intelligence in transforming national cybersecurity strategies. By integrating advanced technologies with interdisciplinary approaches, AI not only enhances the detection and mitigation of threats but also facilitates the reform of traditional governance models amid ongoing digital transformation. The practical value of this work lies in its formulation of concrete recommendations aimed at harmonizing legal, ethical, and technical dimensions—thereby fostering a cohesive protection system for infrastructure through active collaboration among the state, private sector, and academic institutions. The findings will be useful for researchers and cybersecurity specialists, as well as government agencies responsible for shaping and implementing national cyber strategies