Archives of Computational Methods in Engineering最新文献

{"title":"Diagnostic Driven Topology Adaptive Generative Adversarial Networks for Improved Breast Cancer Diagnosis","authors":"Weijia Wang,&nbsp;Francisco Hernando-Gallego,&nbsp;Diego Martín,&nbsp;Mohammad Khishe","doi":"10.1007/s11831-025-10430-5","DOIUrl":"10.1007/s11831-025-10430-5","url":null,"abstract":"<div><p>Breast cancer is one of the major causes of deaths in women. In the meantime, proper and early diagnosis with the help of mammograms can greatly enhance the outcomes of treatment. Nevertheless, the volume of the available data and the inconsistency of lesions are significant obstacles to the development of reliable diagnostic models. Generative Adversarial Networks (GANs) can provide a solution to data augmentation but due to the static nature of their models, they are incapable of capturing diagnostically important features, like irregular mass margins or microcalcifications. The proposed research study proposes a new Diagnostic-Driven Topology-Adaptive GAN (DTA-GAN) framework that improves the performance by adapting the generator and discriminator structures in real-time during the training process based on the diagnosis. DTA-GAN is compared with two baseline models, four state-of-the-art GAN-based models, and five ablation-based DTA-GAN models in 13 metrics on three popular datasets: CBIS-DDSM, INbreast, and Mini-MIAS to perform an extensive assessment. DTA-GANs results remarkably exceed the benchmarks with an AUC of 0.90–0.92 (an increase of 10% over DCGANs 0.82), an FID of 8.40–8.45 (compared to StyleGAN2’s 7.89), and feature preservation LMS metrics of 0.87–0.89 and CDR 91.9–92.7% using both qualitative and quantitative assessment across the three datasets. DTA-GAN synthesizes mammograms with topology controller reward functions focusing on imaging and diagnostics to improve the classification accuracy of the following tasks, offering a breast cancer detection solution that can be scaled to be used widely. This is a breakthrough in medical imaging because it synthesizes data that meets the stringent diagnostic standards, making the systems that are used to make diagnoses more reliable and more generalizable.</p></div>","PeriodicalId":55473,"journal":{"name":"Archives of Computational Methods in Engineering","volume":"33 4","pages":"4793 - 4825"},"PeriodicalIF":12.1,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147959248","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mountainous Flood Resilience: A Comprehensive Systematic Review of Flood Analysis Methods","authors":"Madhab Rijal,&nbsp;Pingping Luo,&nbsp;Binaya Kumar Mishra,&nbsp;Sudeep Thakuri,&nbsp;Yubin Zhang,&nbsp;Yang Zhao","doi":"10.1007/s11831-025-10439-w","DOIUrl":"10.1007/s11831-025-10439-w","url":null,"abstract":"<div><p>Mountainous regions are experiencing increasingly frequent and devastating floods under climate change, demanding urgent advancements in numerical modeling for risk mitigation. Through a comprehensive bibliometric analysis, this study maps the evolution of mountain flood models, revealing that 81% of the research aligns with climate action goals. While traditional tools like HEC-RAS and SWAT dominate current applications, machine learning (ML) adoption has surged exponentially since 2020. However, high-fidelity hydrodynamic models are hampered by prohibitive computational costs, while ML techniques struggle with data scarcity and region-specific biases. Challenges in model structure, parameterization, and historical data reliability further undermine predictive accuracy. The study proposes a framework integrating hybrid AI-physics models, emerging computational technologies, and self-adaptive systems. Physics-Informed Neural Networks (PINNs) and transfer learning to overcome sparse data in regions like the Himalayas. ML-based and quantum computing to accelerate ensemble climate-flood simulations and digital twins to minimize uncertainty through real-time data assimilation. General AI assistants that dynamically adjust models based on live sensor networks. The study highlights the need to revise universal hydrological theories for mountainous contexts, where nonlinear processes such as debris flow and rain-on-snow phenomena defy conventional assumptions. By synthesizing bibliometric trends with cutting-edge technical solutions, this study provides a practical roadmap for hydrologists to select, refine, or design models tailored to the unique challenges of alpine flood risk management. The findings advocate a paradigm shift from static models to adaptive, transparent, and computationally scalable systems to safeguard vulnerable communities in a warming world.</p></div>","PeriodicalId":55473,"journal":{"name":"Archives of Computational Methods in Engineering","volume":"33 4","pages":"4969 - 4997"},"PeriodicalIF":12.1,"publicationDate":"2025-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147958999","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring the Frontiers: A Comprehensive Survey on the Advancements and Prospects of the Spintronics Era","authors":"Alisha P. B,&nbsp;Tripti S. Warrier","doi":"10.1007/s11831-025-10446-x","DOIUrl":"10.1007/s11831-025-10446-x","url":null,"abstract":"<div>\u0000 \u0000 <p>The review aimed to advance both the understanding and practical application of spintronic devices. Through meticulous classification into system-level and device-level approaches, we spotlight design constraints in real-world scenarios. At the system-level domain, exploration extends to broader applications in cache and main memory, addressing issues related to write performance, reliability, and power constraints from a spintronic perspective. Furthermore, the survey meticulously explores device characterization aligned with specific application requirements, aiming to craft spintronic devices uniquely tailored to diverse application needs. It examines trade-offs inherent in various characterization and modeling strategies, providing insights and advancements to the field. Motivated by the impact of spintronic devices in cutting-edge domains like stochastic computing, security, and Artificial Neural Networks, the survey explores spintronics applications’ diverse and evolving landscape. It emphasizes application-centric characterization, dissecting trade-offs to tailor spintronic devices for specific needs and contributing insights that can shape the future of computing. This work offers guidance for technology adoption in specific domains. The work provides nuanced insights to inspire further advancements in spintronics and contribute to the evolution of these transformative technologies.</p>\u0000 </div>","PeriodicalId":55473,"journal":{"name":"Archives of Computational Methods in Engineering","volume":"33 4","pages":"5037 - 5080"},"PeriodicalIF":12.1,"publicationDate":"2025-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147959149","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Comprehensive Review of Digital Video Watermarking Techniques in Spatial, Frequency, and Compressed Domain","authors":"Riddhi,&nbsp;Preeti Garg,&nbsp;Vineet Sharma","doi":"10.1007/s11831-025-10443-0","DOIUrl":"10.1007/s11831-025-10443-0","url":null,"abstract":"<div><p>As there has been an exponential growth of video watching on digital media, protecting multimedia content against piracy, forgery, and unauthorized sharing has become an essential challenge. Conventional cryptographic techniques secure data only while in transit, but do not confirm ownership and authenticity when the video is viewed. In response to this lacuna, researchers have largely studied digital video watermarking, where information is invisibly inserted in video streams to offer copyright protection, authentication, and traceability. A large array of methods has been suggested over the years, which involve spatial-domain strategies like LSB and correlation-based embedding, frequency-domain methods involving DCT, DWT, and SVD, and compressed-domain watermarking with integration of standards like MPEG, H.264, and H.265. Recently, hybrid schemes and deep learning-based methods have been proposed to improve resistance against geometric attacks, desynchronization, and compression while providing imperceptibility. This paper provides a comprehensive overview of these methods, comparing their merits and demerits in various applications. It further explores key performance metrics like PSNR, SSIM, NCC, BER, and BCR, along with the security of watermarking systems against different classes of attacks. Through integration of existing work and an identification of emerging trends such as AI-based adaptive watermarking, blockchain-enabled ownership verification, and codec-aware embedding, this research offers a complete guide for further developing safe, efficient, and viable video watermarking schemes.</p></div>","PeriodicalId":55473,"journal":{"name":"Archives of Computational Methods in Engineering","volume":"33 4","pages":"4999 - 5035"},"PeriodicalIF":12.1,"publicationDate":"2025-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147959053","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Survey of Machine Learning Techniques for Diabetes Prediction: Current Trends and Future Directions","authors":"Mudra Narasimharao,&nbsp;Biswaranjan Swain,&nbsp;Rahul Priyadarshi,&nbsp;Praveen Priyaranjan Nayak,&nbsp;Satyanarayan Bhuyan","doi":"10.1007/s11831-025-10450-1","DOIUrl":"10.1007/s11831-025-10450-1","url":null,"abstract":"<div><p>This comprehensive survey explores diverse diabetes prediction models intending to enhance pattern comprehension and enable early diagnosis. The implications of this research extend to improved insulin treatment and reduced risks of associated complications. Notably, our investigation applies infrequently utilized machine learning (ML) models to diabetes datasets, yielding accuracy rates ranging from 65% to 97%. The recommendation emerges to leverage various algorithms for insulin diagnosis, with a particular emphasis on creating hybrid algorithms to enhance overall performance. A health classification technique rooted in ML is proposed for categorizing individuals into diabetic and non-diabetic groups. Our research spans multiple resources, contributing valuable insights for future studies on diverse diabetes prediction methods. This paper outlines the current landscape of ML algorithms for diabetic prediction, emphasizing the significance of employing varied algorithms. The study delves into an array of prediction techniques, focusing on the widely used Pima dataset. Among the 15 ML algorithms explored, methodologies such as Support Vector Machine (SVM) and Naive Bayes (NBs) are applied. The strategic use of these techniques conserves resources and yields more precise findings, contributing to anticipatory measures against diabetes. Readers are urged to explore this research, as it encapsulates the forefront of diabetic prediction through ML algorithms. The amalgamation of insights from diverse studies facilitates a nuanced understanding of the subject. This survey encapsulates key findings, emphasizing the importance of employing a spectrum of ML algorithms for optimal predictive outcomes.</p></div>","PeriodicalId":55473,"journal":{"name":"Archives of Computational Methods in Engineering","volume":"33 4","pages":"5187 - 5222"},"PeriodicalIF":12.1,"publicationDate":"2025-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147959041","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Systematic Review of Computational Methods for Protein Post-Translational Modification Site Prediction","authors":"Yuan-Yuan Li,&nbsp;Zi Liu,&nbsp;Xin Liu,&nbsp;Yi-Heng Zhu,&nbsp;Conghui Fang,&nbsp;Muhammad Arif,&nbsp;Wang-Ren Qiu","doi":"10.1007/s11831-025-10444-z","DOIUrl":"10.1007/s11831-025-10444-z","url":null,"abstract":"<div><p>Protein post-translational modifications (PTMs) are critical for regulating protein function and are closely linked to disease mechanisms. In-depth research and precise prediction of PTMs are vital for understanding life mechanisms, screening disease biomarkers, and identifying drug targets. Artificial intelligence (AI) approaches for PTM site prediction offer complementary advantages to traditional experimental methods, providing high-throughput and cost-effective screening that can prioritize candidate sites for further validation. This paper reviews advances in PTM site prediction since 2012, focusing on machine learning and deep learning techniques. It analyzes more than 500 relevant studies and categorizes 36 types of PTMs. Additionally, the paper briefly outlines core contents such as database resources related to PTMs, commonly used feature extraction methods, and major classification algorithms. In addition, 36 representative recent studies on PTMs have been carefully selected for in-depth analysis. The findings indicate that current machine learning-based PTM research employs multivariate feature extraction and construct composite models to enhance prediction performance. Finally, keyword visualization using CiteSpace identifies emerging research hotspots and future directions for PTM site prediction.</p></div>","PeriodicalId":55473,"journal":{"name":"Archives of Computational Methods in Engineering","volume":"33 3","pages":"4287 - 4307"},"PeriodicalIF":12.1,"publicationDate":"2025-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s11831-025-10444-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147665776","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AI and ML in Mitigating Membrane Fouling in Heavy Metal Wastewater Treatment: A Review on Recent Trends and Future Industrial Outlook","authors":"Gopirajan Punniyakotti Varadharajan,&nbsp;Kaladevi Ramar,&nbsp;Arun Jayaseelan,&nbsp;Priyadharsini Packiyadhas,&nbsp;Nirmala Narasiman,&nbsp;SundarRajan PanneerSelvam,&nbsp;Naveen Subbaiyan,&nbsp;Gopinath Kannappan Panchamoorthy,&nbsp;Sathish Kumar Palaniappan,&nbsp;Suchart Siengchin","doi":"10.1007/s11831-025-10434-1","DOIUrl":"10.1007/s11831-025-10434-1","url":null,"abstract":"<div><p>Implementing membrane fouling for the treatment of heavy metal–contaminated water is one of the critical challenges for ensuring efficient and sustainable purification. Increase in population, urbanization and industrialization leads to contamination of water sources. Chemicals, garbage, plastics and other pollution have suffocated the rivers, reservoirs, lakes and seas. Water contains both soluble/insoluble waste and it is very difficult to separate the soluble pollutants. Soluble heavy metals like mercury, lead and cadmium affect brain, heart, kidneys, lungs and immune system of people of all ages. Consumption of polluted water in India causes some of the deadly diseases like cholera, dysentery, diarrhea, tuberculosis, jaundice, etc. Not only human beings but also plants, animals, aquatic animals are affected by this pollution which leads to risks like toxicity, persistence in the environment and bio-accumulative nature. This article addresses AI/ML models, deep learning models and metaheuristic algorithms to predict fouling and optimize heavy metal removal through membrane filtration. These models enhance accuracy, decrease the load in experiments and optimize process conditions. Data quality, model interpretability and real-world generalizability still exist in essential gaps. This review suggests the significance of real-time monitoring and explains the need for integrating AI/ML models for wastewater treatment process.</p></div>","PeriodicalId":55473,"journal":{"name":"Archives of Computational Methods in Engineering","volume":"33 3","pages":"4217 - 4236"},"PeriodicalIF":12.1,"publicationDate":"2025-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147665630","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep LearnIng and Machine Learning for Brain Tumor Detection: A Review, Challenges, and Future Directions","authors":"Saeed Mohsen,&nbsp;Sarah Oraby,&nbsp;M. Abdel-Aziz","doi":"10.1007/s11831-025-10416-3","DOIUrl":"10.1007/s11831-025-10416-3","url":null,"abstract":"<div>\u0000 \u0000 <p>Brain tumor diagnosis and classification remain critical challenges in modern healthcare. Recent advancements in artificial intelligence (AI), particularly deep learning (DL) and machine learning (ML), have significantly enhanced medical image analysis, enabling automated and accurate detection of brain tumors. This paper reviews a wide range of ML and DL approaches for brain tumor detection, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), generative adversarial networks (GANs), hybrid architectures, vision transformers (ViTs), transfer learning (TL), attention mechanisms, and ensemble learning methods. Also, AI applications are highlighted across different medical imaging modalities, summarize commonly used publicly available datasets, and discuss preprocessing techniques adopted in recent studies. Furthermore, the paper addresses evaluation metrics, compares state-of-the-art DL approaches, and examines key challenges and limitations. Finally, future research directions are proposed to guide the development of more robust and clinically effective AI-based solutions. This review aims to provide researchers and clinicians with a comprehensive understanding of AIs potential in advancing brain tumor diagnosis and detection.</p>\u0000 </div>","PeriodicalId":55473,"journal":{"name":"Archives of Computational Methods in Engineering","volume":"33 3","pages":"3931 - 3955"},"PeriodicalIF":12.1,"publicationDate":"2025-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s11831-025-10416-3.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147665683","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Stress Detection Using Machine Learning and Deep Learning Techniques: A Systematic Review and Meta-Analysis","authors":"Logesh Kumar Kulanthaivel Lakshmanan,&nbsp;Kavisankar Leelasankar,&nbsp;Balasubramani Subbiyan","doi":"10.1007/s11831-025-10429-y","DOIUrl":"10.1007/s11831-025-10429-y","url":null,"abstract":"<div>\u0000 \u0000 <p>Stress has emerged as a major issue in today’s world, impacting people in numerous areas of their lives. It originates from multiple sources and can be classified into different types and categories. Physiological stress, exerts high pressure on the human body, disrupting daily activities and overall well-being. Prolonged exposure to elevated stress levels may lead to severe health complications, including cardiovascular diseases and other stress-induced disorders. To mitigate these risks, it is crucial to continuously monitor stress levels, enabling early detection and timely intervention. This systematic review explores various stress detection methodologies, including data-driven approaches that leverage online social networks (OSNs) and physiological signals, such as Electroencephalography (EEG) and Electrocardiography (ECG). Additionally, it explores stress indicators derived from wearable devices, including Galvanic Skin Response (GSR), and Skin Temperature (ST). The study further investigates the application of Machine Learning (ML) techniques and Deep Learning (DL) techniques in examining these signals to improve the accuracy of stress detection. Furthermore, this paper examines the utilization of stress detection models across multiple sectors, including the workplace, education, and the automotive industry. It also highlights key research aspects such as objectives, data sources, data analysis, ML or DL methodologies, and model performance evaluation. The review follows the PICO framework (Population, Intervention, Comparison, Outcome) to systematically identify relevant studies, define inclusion criteria, and evaluate the impact of computational models on stress detection outcomes. To provide a quantitative summary of performance across studies, this systematic review incorporates meta-analysis using forest plots and funnel plots. By reviewing existing methodologies, this review aims to identify research gaps and outline potential future directions in the field of stress monitoring, thereby supporting the development of more effective and dependable stress management solutions.</p>\u0000 </div>","PeriodicalId":55473,"journal":{"name":"Archives of Computational Methods in Engineering","volume":"33 3","pages":"4183 - 4215"},"PeriodicalIF":12.1,"publicationDate":"2025-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147665733","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Performance Evaluation of Existing Hybrid Models for Breast Cancer Diagnosis: A Review of Deep Learning and Fuzzy Clustering Approaches in Medical Imaging","authors":"Geeta Agarwal,&nbsp;Sonam Seth","doi":"10.1007/s11831-025-10449-8","DOIUrl":"10.1007/s11831-025-10449-8","url":null,"abstract":"<div><p>This review provides a systematic evaluation of recent advance in breast cancer detection using deep learning, fuzzy logic, and hybrid computational models, with a focus on ultrasound and mammography imaging. Unlike existing surveys that examine these methods in isolation, this work addresses a key gap by analyzing 42 studies published between 2019 and 2025, that integrate deep learning with fuzzy clustering and optimization techniques. Finding show that such hybrid approaches improve lesion segmentation and classification accuracy while mitigating limitations such as image noise, operator dependency, and low interpretability. The BUSI datasets emerged as most widely used benchmark for validation. However, persistent challenges remain, including limited datasets size, lack of external validation, and inconsistent performance reporting, which restrict clinical translation. By Comparing methods, performance and limitation, this review highlights the potential of hybrid AI frameworks to provide more reliable diagnostic support for early breast cancer detection. It concludes that future progress requires larger, multi-institutional datasets, standardized evaluation protocols, and an interpretable hybrid model to enable practical deployment in real-world applications.</p></div>","PeriodicalId":55473,"journal":{"name":"Archives of Computational Methods in Engineering","volume":"33 3","pages":"4309 - 4323"},"PeriodicalIF":12.1,"publicationDate":"2025-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147665734","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}