Pub Date : 2026-01-19DOI: 10.1109/OJCS.2026.3653583
{"title":"2025 Reviewers List*","authors":"","doi":"10.1109/OJCS.2026.3653583","DOIUrl":"https://doi.org/10.1109/OJCS.2026.3653583","url":null,"abstract":"","PeriodicalId":13205,"journal":{"name":"IEEE Open Journal of the Computer Society","volume":"7 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11358645","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146026556","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-15DOI: 10.1109/OJCS.2026.3654171
Jamil Ahmad;Mustaqeem Khan;Wail Guiaeab;Abdulmotaleb Elsaddik;Giulia De Masi;Fakhri Karray
Zero-shot fine-grained recognition is challenging due to high visual similarities between classes and the inferior encoding of fine-grained features in embedding models. In this work, we present an attribute-guided Contrastive Language-Image Pre-training (AG-CLIP) model with an additional attribute encoder. Our approach first identifies relevant visual attributes from the textual class descriptions using an attribute mining module leveraging a large language model (LLM) GPT-4o. The attributes are then used to construct prompts for an open vocabulary object/region detector to extract relevant corresponding image regions. The attribute text, along with focused regions of the input, then guides the CLIP model to focus on these discriminative attributes during fine-tuning through a context-attribute fusion module. Our attribute-guided attention mechanism allows CLIP to effectively disambiguate fine-grained classes by highlighting their distinctive attributes without requiring fine-tuning or additional training data on unseen classes. We evaluate our approach on the CUB-200-2011 and plant disease datasets, achieving 73.3% and 84.6% accuracy, respectively. Our method achieves state-of-the-art zero-shot performance, outperforming prior methods that rely on external knowledge bases or complex meta-learning strategies. The strong results demonstratethe effectiveness of injecting generic attribute awareness into powerful vision-language models like CLIP for tackling fine-grained recognition in a zero-shot manner.
{"title":"AG-CLIP: Attribute-Guided CLIP for Zero-Shot Fine-Grained Recognition","authors":"Jamil Ahmad;Mustaqeem Khan;Wail Guiaeab;Abdulmotaleb Elsaddik;Giulia De Masi;Fakhri Karray","doi":"10.1109/OJCS.2026.3654171","DOIUrl":"https://doi.org/10.1109/OJCS.2026.3654171","url":null,"abstract":"Zero-shot fine-grained recognition is challenging due to high visual similarities between classes and the inferior encoding of fine-grained features in embedding models. In this work, we present an attribute-guided Contrastive Language-Image Pre-training (AG-CLIP) model with an additional attribute encoder. Our approach first identifies relevant visual attributes from the textual class descriptions using an attribute mining module leveraging a large language model (LLM) GPT-4o. The attributes are then used to construct prompts for an open vocabulary object/region detector to extract relevant corresponding image regions. The attribute text, along with focused regions of the input, then guides the CLIP model to focus on these discriminative attributes during fine-tuning through a context-attribute fusion module. Our attribute-guided attention mechanism allows CLIP to effectively disambiguate fine-grained classes by highlighting their distinctive attributes without requiring fine-tuning or additional training data on unseen classes. We evaluate our approach on the CUB-200-2011 and plant disease datasets, achieving 73.3% and 84.6% accuracy, respectively. Our method achieves state-of-the-art zero-shot performance, outperforming prior methods that rely on external knowledge bases or complex meta-learning strategies. The strong results demonstratethe effectiveness of injecting generic attribute awareness into powerful vision-language models like CLIP for tackling fine-grained recognition in a zero-shot manner.","PeriodicalId":13205,"journal":{"name":"IEEE Open Journal of the Computer Society","volume":"7 ","pages":"365-375"},"PeriodicalIF":0.0,"publicationDate":"2026-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11352798","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146082298","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-14DOI: 10.1109/OJCS.2026.3654173
Gayathri Ramasamy;Tripty Singh;Xiaohui Yuan;Ganesh R Naik
The methylation status of the $O^{6}$-methylguanine-DNA methyltransferase (MGMT) promoter is an established prognostic and predictive biomarker in glioma, particularly for estimating response to alkylating chemotherapy such as temozolomide. However, many existing radiogenomic methods remain constrained by invasive biopsy dependence, slice-wise 2D modelling, limited use of multi-modal MRI, and insufficient interpretability, which collectively impede clinical translation. We propose MM-3DAttNet, a multi-modal 3D attention network for noninvasive prediction of MGMT promoter methylation status from pre-operative multiparametric brain MRI. The model employs four modality-specific 3D CNN encoder branches (T1, T1ce, T2, and FLAIR) and integrates them using a cross-modality attention fusion module to capture complementary diagnostic cues. MM-3DAttNet was trained and evaluated on the BraTS 2021 cohort comprising 585 glioma cases with MGMT labels, achieving an average accuracy of 91.6%, $_{1}$-score of 89.9%, and AUC of 0.925 under five-fold cross-validation. Interpretability was supported using Grad-CAM saliency maps, which consistently emphasized clinically relevant regions such as enhancing tumour boundaries and peritumoural oedema. Ablation experiments verified the importance of multi-modal learning and attention-based fusion, with the most pronounced performance reductions observed when excluding T1ce or FLAIR. Overall, MM-3DAttNet provides an accurate and interpretable radiogenomic framework for MGMT methylation assessment and supports future validation in multi-centre settings and integration into MRI-based decision-support workflows for glioma management.
{"title":"MM-3DAttNet: Multi-Modal 3D Attention Network for MGMT Methylation Prediction","authors":"Gayathri Ramasamy;Tripty Singh;Xiaohui Yuan;Ganesh R Naik","doi":"10.1109/OJCS.2026.3654173","DOIUrl":"https://doi.org/10.1109/OJCS.2026.3654173","url":null,"abstract":"The methylation status of the <inline-formula><tex-math>$O^{6}$</tex-math></inline-formula>-methylguanine-DNA methyltransferase (MGMT) promoter is an established prognostic and predictive biomarker in glioma, particularly for estimating response to alkylating chemotherapy such as temozolomide. However, many existing radiogenomic methods remain constrained by invasive biopsy dependence, slice-wise 2D modelling, limited use of multi-modal MRI, and insufficient interpretability, which collectively impede clinical translation. We propose MM-3DAttNet, a multi-modal 3D attention network for noninvasive prediction of MGMT promoter methylation status from pre-operative multiparametric brain MRI. The model employs four modality-specific 3D CNN encoder branches (T1, T1ce, T2, and FLAIR) and integrates them using a cross-modality attention fusion module to capture complementary diagnostic cues. MM-3DAttNet was trained and evaluated on the BraTS 2021 cohort comprising 585 glioma cases with MGMT labels, achieving an average accuracy of 91.6%, <inline-formula><tex-math>$_{1}$</tex-math></inline-formula>-score of 89.9%, and AUC of 0.925 under five-fold cross-validation. Interpretability was supported using Grad-CAM saliency maps, which consistently emphasized clinically relevant regions such as enhancing tumour boundaries and peritumoural oedema. Ablation experiments verified the importance of multi-modal learning and attention-based fusion, with the most pronounced performance reductions observed when excluding T1ce or FLAIR. Overall, MM-3DAttNet provides an accurate and interpretable radiogenomic framework for MGMT methylation assessment and supports future validation in multi-centre settings and integration into MRI-based decision-support workflows for glioma management.","PeriodicalId":13205,"journal":{"name":"IEEE Open Journal of the Computer Society","volume":"7 ","pages":"343-353"},"PeriodicalIF":0.0,"publicationDate":"2026-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11352853","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146082297","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Electroencephalography (EEG)-based emotion recognition holds promise for real-time mental health monitoring, adaptive interfaces, and affective computing. However, accurate prediction across individuals remains challenging due to inter-subject variability and the non-stationary nature of EEG signals. To address this, we propose EMO-CARE, a lightweight deep learning framework that integrates multi-scale temporal convolutional networks with feature-level self-attention operating on multi-scale temporal representations. This architecture captures emotional patterns across diverse neural timescales while adaptively weighting multi-scale temporal features based on their relevance. Evaluated under the rigorous Leave-One-Subject-Out (LOSO) protocol on three benchmark datasets: SEED, SEED-V, and DREAMER, EMO-CARE achieves state-of-the-art accuracy with low inference latency. Extensive ablation experiments demonstrate the contribution of each architectural component, and the learned attention patterns align with known emotion-related neural activity. These findings collectively highlight EMO-CARE’s effectiveness in achieving subject-independent generalization and real-time applicability for EEG-based emotion recognition.
{"title":"EMO-CARE: EEG Multi-Scale Temporal Modeling With Channel-Aware Feature Attention for Robust Subject-Independent Emotion Recognition","authors":"Yeganeh Abdollahinejad;Ahmad Mousavi;Petros Siaplaouras;Zois Boukouvalas;Roberto Corizzo","doi":"10.1109/OJCS.2026.3653766","DOIUrl":"https://doi.org/10.1109/OJCS.2026.3653766","url":null,"abstract":"Electroencephalography (EEG)-based emotion recognition holds promise for real-time mental health monitoring, adaptive interfaces, and affective computing. However, accurate prediction across individuals remains challenging due to inter-subject variability and the non-stationary nature of EEG signals. To address this, we propose EMO-CARE, a lightweight deep learning framework that integrates multi-scale temporal convolutional networks with feature-level self-attention operating on multi-scale temporal representations. This architecture captures emotional patterns across diverse neural timescales while adaptively weighting multi-scale temporal features based on their relevance. Evaluated under the rigorous Leave-One-Subject-Out (LOSO) protocol on three benchmark datasets: SEED, SEED-V, and DREAMER, EMO-CARE achieves state-of-the-art accuracy with low inference latency. Extensive ablation experiments demonstrate the contribution of each architectural component, and the learned attention patterns align with known emotion-related neural activity. These findings collectively highlight EMO-CARE’s effectiveness in achieving subject-independent generalization and real-time applicability for EEG-based emotion recognition.","PeriodicalId":13205,"journal":{"name":"IEEE Open Journal of the Computer Society","volume":"7 ","pages":"354-364"},"PeriodicalIF":0.0,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11348088","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146082296","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Images have become a strategic digital asset that powers creative industries, e–commerce, and data–driven services. However, modern editing tools and large–scale sharing platforms have made copyright infringement, unauthorized redistribution, and covert manipulation easier to perpetrate and harder to detect. These risks lead to financial losses and weaken trust in digital ecosystems, creating an urgent need for technical protections that complement legal remedies. This paper presents a comprehensive survey of technologies and approaches for image copyright protection, with a particular emphasis on digital watermarking, deep learning-based methods, and blockchain-enabled frameworks. We systematically examine the principles, mechanisms, and applications of these techniques, evaluating their strengths, limitations, and potential synergies. In addition, we explore how these technologies can be effectively integrated into practical systems for secure, reliable, and scalable copyright protection of images. Finally, we identify existing challenges and propose promising future research directions to advance the state of the art in image copyright protection.
{"title":"Image Copyright Protection: A Comprehensive Survey of Digital Watermarking, Deep Learning, and Blockchain Approaches","authors":"Phuc Nguyen;Tan Hanh;Truong Duy Dinh;Trong Thua Huynh","doi":"10.1109/OJCS.2026.3651292","DOIUrl":"https://doi.org/10.1109/OJCS.2026.3651292","url":null,"abstract":"Images have become a strategic digital asset that powers creative industries, e–commerce, and data–driven services. However, modern editing tools and large–scale sharing platforms have made copyright infringement, unauthorized redistribution, and covert manipulation easier to perpetrate and harder to detect. These risks lead to financial losses and weaken trust in digital ecosystems, creating an urgent need for technical protections that complement legal remedies. This paper presents a comprehensive survey of technologies and approaches for image copyright protection, with a particular emphasis on digital watermarking, deep learning-based methods, and blockchain-enabled frameworks. We systematically examine the principles, mechanisms, and applications of these techniques, evaluating their strengths, limitations, and potential synergies. In addition, we explore how these technologies can be effectively integrated into practical systems for secure, reliable, and scalable copyright protection of images. Finally, we identify existing challenges and propose promising future research directions to advance the state of the art in image copyright protection.","PeriodicalId":13205,"journal":{"name":"IEEE Open Journal of the Computer Society","volume":"7 ","pages":"244-263"},"PeriodicalIF":0.0,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11339926","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145982319","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-12DOI: 10.1109/OJCS.2026.3651226
Xiuyuan Zhao;Jingyi Liu;Ying Wang;Jiyuan Wang
Cryptocurrency markets exhibit complex microstructural dynamics characterized by high-frequency volatility bursts, rapid regime switching, and long-range temporal dependencies, which expose several limitations of existing volatility forecasting approaches. In particular, attention-based models suffer from prohibitive quadratic computational cost on long high-frequency sequences, while many recurrent architectures struggle to adapt to regime transitions, asymmetric volatility responses, and risk-aware uncertainty estimation. To address these gaps, this paper proposes CryptoMamba-SSM, a novel volatility prediction framework built upon Mamba-based state space models with linear computational complexity. CryptoMamba-SSM integrates selective memory mechanisms with structured state space representations to effectively capture critical market microstructure signals arising from liquidity shocks and sentiment transitions, while dynamically adjusting memory retention across different volatility regimes. This design enables efficient modeling of long-sequence dependencies inherent in cryptocurrency price movements without incurring the computational bottlenecks of traditional attention-based architectures. Through comprehensive experiments on Bitcoin historical data spanning multiple market regimes, we demonstrate that CryptoMamba-SSM consistently outperforms conventional LSTM, GRU, and Transformer baselines, achieving up to a 23.7% reduction in Mean Absolute Error and a 31.2% improvement in directional accuracy. The selective memory mechanism effectively captures regime-switching behaviors and microstructural anomalies, leading to more reliable short-term volatility risk quantification. Moreover, the linear-time complexity of CryptoMamba-SSM enables real-time processing of high-frequency trading data while maintaining strong generalization across diverse market conditions.
{"title":"CryptoMamba-SSM: Linear Complexity State Space Models for Cryptocurrency Volatility Prediction","authors":"Xiuyuan Zhao;Jingyi Liu;Ying Wang;Jiyuan Wang","doi":"10.1109/OJCS.2026.3651226","DOIUrl":"https://doi.org/10.1109/OJCS.2026.3651226","url":null,"abstract":"Cryptocurrency markets exhibit complex microstructural dynamics characterized by high-frequency volatility bursts, rapid regime switching, and long-range temporal dependencies, which expose several limitations of existing volatility forecasting approaches. In particular, attention-based models suffer from prohibitive quadratic computational cost on long high-frequency sequences, while many recurrent architectures struggle to adapt to regime transitions, asymmetric volatility responses, and risk-aware uncertainty estimation. To address these gaps, this paper proposes <bold>CryptoMamba-SSM</b>, a novel volatility prediction framework built upon Mamba-based state space models with linear computational complexity. CryptoMamba-SSM integrates selective memory mechanisms with structured state space representations to effectively capture critical market microstructure signals arising from liquidity shocks and sentiment transitions, while dynamically adjusting memory retention across different volatility regimes. This design enables efficient modeling of long-sequence dependencies inherent in cryptocurrency price movements without incurring the computational bottlenecks of traditional attention-based architectures. Through comprehensive experiments on Bitcoin historical data spanning multiple market regimes, we demonstrate that CryptoMamba-SSM consistently outperforms conventional LSTM, GRU, and Transformer baselines, achieving up to a 23.7% reduction in Mean Absolute Error and a 31.2% improvement in directional accuracy. The selective memory mechanism effectively captures regime-switching behaviors and microstructural anomalies, leading to more reliable short-term volatility risk quantification. Moreover, the linear-time complexity of CryptoMamba-SSM enables real-time processing of high-frequency trading data while maintaining strong generalization across diverse market conditions.","PeriodicalId":13205,"journal":{"name":"IEEE Open Journal of the Computer Society","volume":"7 ","pages":"226-243"},"PeriodicalIF":0.0,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11339936","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145982383","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-12DOI: 10.1109/OJCS.2026.3651756
Mohammed Saad Javeed;Jannatul Maua;Rahomotul Islam;Mumtahina Ahmed;M. F. Mridha;Md. Jakir Hossen
Effective incident management in modern IT systems requires timely interpretation and routing of alerts generated from diverse sources such as SNMP Traps, Syslog messages, and xMatters notifications. However, conventional frameworks often lack unified processing and intelligent automation, resulting in delayed response and SLA violations. This paper presents an AI-enhanced unified alerting and incident management framework that integrates heterogeneous alert streams via the ServiceNow platform. Leveraging two real-world datasets comprising over 140,000 event records and 24,000 unique incidents, we implement a multi-task deep neural network to jointly predict resolution time, incident priority, and responsible assignment group. The proposed method incorporates temporal feature engineering, trainable embeddings for categorical data, and variational autoencoders for dimensionality reduction. A synthetic alert-source simulation is introduced to mimic real-world alert diversity within the data pipeline. Experimental results demonstrate superior performance over baseline models in all key metrics, validating the effectiveness of the proposed architecture. The framework sets the stage for scalable, automated, and context-aware incident triaging in enterprise IT environments.
{"title":"A Multi-Task Neural Framework for Unified Alert Processing and Incident Prediction in Enterprise IT Systems","authors":"Mohammed Saad Javeed;Jannatul Maua;Rahomotul Islam;Mumtahina Ahmed;M. F. Mridha;Md. Jakir Hossen","doi":"10.1109/OJCS.2026.3651756","DOIUrl":"https://doi.org/10.1109/OJCS.2026.3651756","url":null,"abstract":"Effective incident management in modern IT systems requires timely interpretation and routing of alerts generated from diverse sources such as SNMP Traps, Syslog messages, and xMatters notifications. However, conventional frameworks often lack unified processing and intelligent automation, resulting in delayed response and SLA violations. This paper presents an AI-enhanced unified alerting and incident management framework that integrates heterogeneous alert streams via the ServiceNow platform. Leveraging two real-world datasets comprising over 140,000 event records and 24,000 unique incidents, we implement a multi-task deep neural network to jointly predict resolution time, incident priority, and responsible assignment group. The proposed method incorporates temporal feature engineering, trainable embeddings for categorical data, and variational autoencoders for dimensionality reduction. A synthetic alert-source simulation is introduced to mimic real-world alert diversity within the data pipeline. Experimental results demonstrate superior performance over baseline models in all key metrics, validating the effectiveness of the proposed architecture. The framework sets the stage for scalable, automated, and context-aware incident triaging in enterprise IT environments.","PeriodicalId":13205,"journal":{"name":"IEEE Open Journal of the Computer Society","volume":"7 ","pages":"264-275"},"PeriodicalIF":0.0,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11339966","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145982318","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Accurate and objective assessment of rehabilitation exercises is critical for ensuring correct execution and maximizing patient recovery, particularly in unsupervised or home-based settings. Existing deep learning approaches frequently rely on graph-based skeletal representations with predefined topologies, which constrain the discovery of long-range or task-specific joint dependencies and limit adaptability across datasets with varying skeletal definitions. To address these limitations, we propose a Spatial–Temporal Transformer framework that directly models 3D joint position data without requiring an explicit adjacency matrix. The framework incorporates a joint-wise feature encoding and structure embedding mechanism to provide unique representations for each joint, thereby mitigating ambiguities arising from symmetry or overlapping movements. Furthermore, a stochastic time-warping augmentation strategy is introduced to simulate execution speed variations, enhancing robustness to diverse patient movement patterns. By applying small, randomized temporal scaling to local segments while consistently interpolating spatial coordinates within temporal boundaries, this stochastic variation enriches the dataset significantly while preserving the biomechanical patterns. Experimental results on the KIMORE dataset demonstrate that the proposed method reduces mean absolute deviation (MAD) by 67.4 % relative to the current state of the art, while also maintaining strong generalization on the UI-PRMD dataset. The approach is compatible with multiple pose estimation algorithms and acquisition modalities, making it suitable for deployment in real-world telerehabilitation and clinical monitoring applications.
{"title":"Spatial–Temporal Transformers With Stochastic Time-Warping and Joint-Wise Encoding for Rehabilitation Exercise Assessment","authors":"Tanawat Matangkasombut;Wuttipong Kumwilaisak;Chatchawarn Hansakunbuntheung;Nattanun Thatphithakkul","doi":"10.1109/OJCS.2025.3650355","DOIUrl":"https://doi.org/10.1109/OJCS.2025.3650355","url":null,"abstract":"Accurate and objective assessment of rehabilitation exercises is critical for ensuring correct execution and maximizing patient recovery, particularly in unsupervised or home-based settings. Existing deep learning approaches frequently rely on graph-based skeletal representations with predefined topologies, which constrain the discovery of long-range or task-specific joint dependencies and limit adaptability across datasets with varying skeletal definitions. To address these limitations, we propose a Spatial–Temporal Transformer framework that directly models 3D joint position data without requiring an explicit adjacency matrix. The framework incorporates a joint-wise feature encoding and structure embedding mechanism to provide unique representations for each joint, thereby mitigating ambiguities arising from symmetry or overlapping movements. Furthermore, a stochastic time-warping augmentation strategy is introduced to simulate execution speed variations, enhancing robustness to diverse patient movement patterns. By applying small, randomized temporal scaling to local segments while consistently interpolating spatial coordinates within temporal boundaries, this stochastic variation enriches the dataset significantly while preserving the biomechanical patterns. Experimental results on the KIMORE dataset demonstrate that the proposed method reduces mean absolute deviation (MAD) by 67.4 % relative to the current state of the art, while also maintaining strong generalization on the UI-PRMD dataset. The approach is compatible with multiple pose estimation algorithms and acquisition modalities, making it suitable for deployment in real-world telerehabilitation and clinical monitoring applications.","PeriodicalId":13205,"journal":{"name":"IEEE Open Journal of the Computer Society","volume":"7 ","pages":"190-201"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11321303","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145982323","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-26DOI: 10.1109/OJCS.2025.3649157
Iman H. Meskini;Cristina Alcaraz;Rodrigo Roman Castro;Javier Lopez
Anomaly Detection Systems (ADS) are essential in Industrial and Internet of Things (IIoT) environments by identifying equipment failures, environmental anomalies, operational irregularities, and cyberattacks. However, the increasing reliance on Machine Learning and Deep Learning (DL) exposes ADS to adversarial attacks, particularly transferable evasion attacks, where Adversarial Examples (AE) crafted for one model can deceive others. Despite their importance, limited research has examined the transferability of adversarial attacks in industrial and IoT contexts or the effectiveness of defense strategies against them. This work systematically evaluates the transferability of adversarial evasion attacks across six ADS models, including both tree-based and neural network architectures, trained on industrial and IIoT scenarios datasets. We also analyze multiple adversarial detection methods, measuring not only their performance, but also their computational efficiency in terms of execution time, processor utilization, and energy consumption. Our results show that most ADS are vulnerable to transferable evasion attacks and that existing detection methods fail in model- and attack-agnostic settings. We further demonstrate that incorporating adversarial learning with a small set of low-perturbation examples significantly improves detection while maintaining low computational overhead, enabling practical and efficient real-time deployment.
{"title":"Analysis of Transferable Adversarial Evasion Attack Detection in IoT and Industrial ADS","authors":"Iman H. Meskini;Cristina Alcaraz;Rodrigo Roman Castro;Javier Lopez","doi":"10.1109/OJCS.2025.3649157","DOIUrl":"https://doi.org/10.1109/OJCS.2025.3649157","url":null,"abstract":"Anomaly Detection Systems (ADS) are essential in Industrial and Internet of Things (IIoT) environments by identifying equipment failures, environmental anomalies, operational irregularities, and cyberattacks. However, the increasing reliance on Machine Learning and Deep Learning (DL) exposes ADS to adversarial attacks, particularly transferable evasion attacks, where Adversarial Examples (AE) crafted for one model can deceive others. Despite their importance, limited research has examined the transferability of adversarial attacks in industrial and IoT contexts or the effectiveness of defense strategies against them. This work systematically evaluates the transferability of <italic>adversarial evasion attacks</i> across six ADS models, including both tree-based and neural network architectures, trained on industrial and IIoT scenarios datasets. We also analyze multiple adversarial detection methods, measuring not only their performance, but also their computational efficiency in terms of execution time, processor utilization, and energy consumption. Our results show that most ADS are vulnerable to transferable evasion attacks and that existing detection methods fail in model- and attack-agnostic settings. We further demonstrate that incorporating adversarial learning with a small set of low-perturbation examples significantly improves detection while maintaining low computational overhead, enabling practical and efficient real-time deployment.","PeriodicalId":13205,"journal":{"name":"IEEE Open Journal of the Computer Society","volume":"7 ","pages":"142-153"},"PeriodicalIF":0.0,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11316607","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145929483","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ransomware poses a growing threat to healthcare systems, compromising patient safety, operational continuity, and financial stability. Although machine learning techniques have been widely used for intrusion detection, most approaches do not support real-time, cost-sensitive response planning. In this paper, we propose a hybrid framework that integrates deep learning with quantum optimization to both predict the severity of ransomware infection and recommend optimal recovery strategies. The system employs a multilayer perceptron (MLP) trained on structured ransomware incident data to forecast infection rates, followed by a quantum post-processing module using the Quantum Approximate Optimization Algorithm (QAOA) to minimize operational and financial costs through discrete response decisions. We evaluated the framework on a realistic healthcare ransomware dataset consisting of 5,000 simulated attack scenarios. Our approach achieves a root mean squared error (RMSE) of 0.073 in the prediction of infection rates and demonstrates up to 25% cost savings over classical heuristics in recovery decisions. Extensive experiments confirm the generalizability of the model to unseen attack types, the scalability across data volumes, and the stability of the decision in risk contexts. The proposed method represents a step toward intelligent real-time ransomware mitigation systems for high-risk environments such as healthcare.
{"title":"A Hybrid Deep Learning and Quantum Optimization Framework for Ransomware Response in Healthcare","authors":"Ahsan Ahmed;Md Aktarujjaman;Mohammad Moniruzzaman;Md Shahab Uddin;Arifa Akter Eva;M. F. Mridha;Kyoungyee Kim;Jungpil Shin","doi":"10.1109/OJCS.2025.3648741","DOIUrl":"https://doi.org/10.1109/OJCS.2025.3648741","url":null,"abstract":"Ransomware poses a growing threat to healthcare systems, compromising patient safety, operational continuity, and financial stability. Although machine learning techniques have been widely used for intrusion detection, most approaches do not support real-time, cost-sensitive response planning. In this paper, we propose a hybrid framework that integrates deep learning with quantum optimization to both predict the severity of ransomware infection and recommend optimal recovery strategies. The system employs a multilayer perceptron (MLP) trained on structured ransomware incident data to forecast infection rates, followed by a quantum post-processing module using the Quantum Approximate Optimization Algorithm (QAOA) to minimize operational and financial costs through discrete response decisions. We evaluated the framework on a realistic healthcare ransomware dataset consisting of 5,000 simulated attack scenarios. Our approach achieves a root mean squared error (RMSE) of 0.073 in the prediction of infection rates and demonstrates up to 25% cost savings over classical heuristics in recovery decisions. Extensive experiments confirm the generalizability of the model to unseen attack types, the scalability across data volumes, and the stability of the decision in risk contexts. The proposed method represents a step toward intelligent real-time ransomware mitigation systems for high-risk environments such as healthcare.","PeriodicalId":13205,"journal":{"name":"IEEE Open Journal of the Computer Society","volume":"7 ","pages":"154-165"},"PeriodicalIF":0.0,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11316401","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145929480","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：