Shahid Beheshti UniversityJournal of Innovations in Computer Science and Engineering (JICSE)2981-21353120250601Malfustection: Obfuscated Malware Detection and Malware Classification with Data Shortage by Combining Semi-Supervised and Contrastive Learning12110606610.48308/jicse.2025.237344.1041ENMohammad MahdiMaghouliFaculty of Computer Science and Engineering, Shahid Beheshti University, Tehran, Iran.MohamadrezaFereydooniFaculty of Computer Science and Engineering, Shahid Beheshti University, Tehran, IranMojtabaVahidi AslFaculty of Computer Science and Engineering, Shahid Beheshti University, Tehran, IranMonirehAbdoosShahid Beheshti University Faculty of Computer Science and Engineering
artificial intelligenceJournal Article20241103With the advent of new technologies, using various formats of digital gadgets is becoming widespread. In today's world, where everyday tasks are inevitable without technology, this extensive use of computers paves the way for malicious activity. Malware is one of the well-known and widely used means utilized for doing destructive activities by malicious attackers. Producing malware from scratch is somewhat difficult, so attackers tend to obfuscate existing malware and prepare it to become an unrecognizable program. Since creating new malware from an old one using obfuscation is a creative task, there are some drawbacks to identifying obfuscated malwares. In this research, we propose a solution to overcome this problem by converting the code to an image in the first step and then using a semi-supervised approach combined with contrastive learning. In this case, an obfuscation in the malware bytecode corresponds to an augmentation in the image. Hence, by utilizing meaningful augmentations, which simulate some obfuscation changes and combine them to generate complex ambiguity procedures, our proposed solution is able to construct, learn, and detect a wide range of obfuscations. This work addresses two issues: 1) malware classification despite the data deficiency and 2) obfuscated malware detection by training on non-obfuscated malwares. According to the results, the proposed method overcomes the data shortage problem in malware classification, as its accuracy is 90.1% when just 10% of data is used for training the model. Moreover, training on basic malwares without obfuscation achieved 96.21 percent accuracy in detecting obfuscated malware.https://jicse.sbu.ac.ir/article_106066_ed3f53134a28809195cf74216365f581.pdfShahid Beheshti UniversityJournal of Innovations in Computer Science and Engineering (JICSE)2981-21353120250601Automatic Multi-Class Cardiovascular Magnetic Resonance Image Quality Assessment using Unsupervised Domain Adaptation in Spatial and Frequency Domains223510620110.48308/jicse.2025.239094.1048ENShahabedinNabaviShahid Beheshti UniversityHosseinSimchiVelenjak, Sb University, Dept. Of Computer EngineeringMohsenEbrahimi MoghaddamShahid Beheshti UniversityAhmad AliAbinShahid Beheshti UniversityAlejandro F.FrangiThe University of ManchesterJournal Article20250309Population imaging studies rely on good-quality medical imagery before quantifying downstream images. This study provides an automated approach for image quality assessment (IQA) from cardiovascular magnetic resonance (CMR) imaging at scale. We identify four common CMR imaging artefacts: respiratory motion, cardiac motion, Gibbs ringing, and aliasing. Four datasets, including UK Biobank, York University (YU), the Universidad Carlos III (UCIII) and CMR-Tehran, were used to perform the experiments. This study proposes two deep-learning models for CMR IQA in spatial and frequency domains. The presented spatial-domain model also has domain adaptation. The accuracies of supervised 4-fold cross-validation experiments for UK Biobank, YU, UCIII and CMR-Tehran datasets are 99.41%, 75.78%, 89.46% and 67.87% for the spatial-domain and 87.46%, 63.76%, 80.25% and 58.48% for the frequency-domain. Domain adaptation results, considering UK Biobank as the source set and YU, UCIII and CMR-Tehran as the target sets, show the domain shift gap coverage between the datasets to the extent of +11.91%, +3.93% and +16.57%, respectively. Besides, by training and testing the spatial-domain model on 30,125 images from the UK Biobank, an accuracy of 89.56% was obtained in a training time of 394.80 seconds. Meanwhile, the frequency-domain model with training and testing on 180,750 images achieves an accuracy of 87.99% in a training time of 255.04 seconds. Thus, the frequency-domain model can achieve almost the same accuracy yet 1.548 times faster than the spatial model. The proposed models can detect four common CMR imaging artefacts by receiving images or the corresponding k-spaces.https://jicse.sbu.ac.ir/article_106201_2ea4fbc1cde04dde16b760ff0749f7fd.pdfShahid Beheshti UniversityJournal of Innovations in Computer Science and Engineering (JICSE)2981-21353120250601Automated Recognition of Marine Thermal Patterns Using Deep Learning364210620210.48308/jicse.2025.239850.1079ENAlirezaSharifiDepartment of Surveying Engineering, Faculty of Civil, Water and Environmental Engineering, Shahid Beheshti University, Tehran, IranAlirezaVafaeinejadDepartment of Surveying Engineering, Faculty of Civil, Water and Environmental Engineering, Shahid Beheshti University, Tehran, IranJournal Article20250507Abstract— Sea Surface Temperature (SST) data reveal the temporal and spatial distribution of warm anticyclonic eddies and cold cyclonic eddies, impacting ocean behavior. The SST combined products attained adequate decisions to permit the recognition of mesoscale eddies with the introduction of altimeter operations and the availability of two or more altimeters at the same time. Climate change impacts ocean circulation and atmospheric anomalies linked to SST variations. Ocean eddies, vital for material and energy transport, require precise identification to advance oceanography. This study uses SST data from CMEMS in the Atlantic to introduce EddyNet, a deep-learning model for automatic eddy detection and classification. EddyNet's encoder-decoder architecture includes a pixel-wise classification layer, labeling each pixel as "0" (non-eddy), "1" (anticyclonic), or "2" (cyclonic). The high-resolution feature representation outperforms existing models, marking a significant leap in eddy detection accuracy and reliability. This study introduces EddyNet, a deep-learning model based on the U-Net architecture for automatic eddy detection and classification using Sea Surface Temperature (SST) data. The model was trained and evaluated on satellite imagery from the Copernicus Marine Environment Monitoring Service (CMEMS), achieving a training accuracy of 78.55%, a Dice score of 31.99%, and a precision of 0.9259. The recall values for different classes indicate that the model correctly identifies 99.51% of non-eddy pixels, 51.57% of anticyclonic eddy pixels, and 57.19% of cyclonic eddy pixels. These results demonstrate the effectiveness of deep learning in mesoscale eddy detection and highlight the potential for further optimization in classifying eddy structures with higher precision.Shahid Beheshti UniversityJournal of Innovations in Computer Science and Engineering (JICSE)2981-21353120250601Unlocking individual motor signatures using feature-based clustering of a graphomotor task434710626710.48308/jicse.2025.239593.1084ENZinatZarandiINSERM UMR1093-CAPS, UFR des Sciences du Sport, Université Bourgogne Franche-Comté, Dijon, France.Journal Article20250922Abstract—Understanding individual motor signatures (IMS) is essential for personalized treatment and performance optimization. This study investigates the effectiveness of Fuzzy C-Means (FCM) clustering for identifying individual motor signatures from graphomotor tasks. We analyze various kinematic and geometric features, such as movement duration, velocity, and trajectory length, to reveal which aspects of motor behavior are most effective in distinguishing individuals. The results show that features like length of movement are particularly discriminative, while others, such as beta and velocity, offer weaker clustering outcomes. <br />Understanding individual motor signatures (IMS) is essential for personalized treatment and performance optimization. This study investigates the effectiveness of Fuzzy C-Means (FCM) clustering for identifying individual motor signatures from graphomotor tasks. We analyze various kinematic and geometric features, such as movement duration, velocity, and trajectory length, to reveal which aspects of motor behavior are most effective in distinguishing individuals. The results show that features like length of movement are particularly discriminative, while others, such as beta and velocity, offer weaker clustering outcomes.https://jicse.sbu.ac.ir/article_106267_7791995e2015364845929d63d832d4b9.pdfShahid Beheshti UniversityJournal of Innovations in Computer Science and Engineering (JICSE)2981-21353120250601Deep Learning Frailty Model for Heart Failure Survival Prediction485210626610.48308/jicse.2025.241654.1085ENSolmazNorouziUniversity of Qazvin, Zanjan University of Medical SciencesHosseinKhormaeiDepartment of Electrical Engineering, National University of Skills (NUS),
Tehran, Iran.EbrahimHajizadehDepartment of Biostatistics, Faculty of Medical Sciences, Tarbiat Modares University,
Tehran, Iran.NasimNaderiRajaie Cardiovascular Medical and Research Center, Iran University of Medical Sciences,
Tehran, Iran.Journal Article20250922Abstract—The study employed Deep Learning Frailty (DLF), a compelling neural modeling framework for predicting heart failure patient survival. The DLF embeds a notion of multiplicative frailty from classical survival analysis that deals with unobserved heterogeneity while exploiting the neural structure's strong capabilities in approximating any non-linear covariate relationship. The results showed that Incorporating frailty leads to significant improvements, and the DLF model performs better on average. <br /><br />Abstract—The study employed Deep Learning Frailty (DLF), a compelling neural modeling framework for predicting heart failure patient survival. The DLF embeds a notion of multiplicative frailty from classical survival analysis that deals with unobserved heterogeneity while exploiting the neural structure's strong capabilities in approximating any non-linear covariate relationship. The results showed that Incorporating frailty leads to significant improvements, and the DLF model performs better on average. <br /><br />Abstract—The study employed Deep Learning Frailty (DLF), a compelling neural modeling framework for predicting heart failure patient survival. The DLF embeds a notion of multiplicative frailty from classical survival analysis that deals with unobserved heterogeneity while exploiting the neural structure's strong capabilities in approximating any non-linear covariate relationship. The results showed that Incorporating frailty leads to significant improvements, and the DLF model performs better on average.https://jicse.sbu.ac.ir/article_106266_14250901d59083600d859963c0d5d587.pdfShahid Beheshti UniversityJournal of Innovations in Computer Science and Engineering (JICSE)2981-21353120250601A Novel Fixed-Parameter Activation Function for Neural Networks: Enhanced Accuracy and Convergence on MNIST535810629610.48308/jicse.2025.241737.1086ENNajmehHosseinipour-MahaniDepartment of Applied Mathematics,
Graduate University of Advanced Technology,
Kerman, IranAmirrezaJahantabDepartment of Computer science,
Shahid Bahonar University of Kerman
Kerman, IranJournal Article20250928Abstract— Activation functions are essential for extracting meaningful relationships from real-world data in deep learning models. The design of activation functions is critical, as they directly influence the performance of these models. Nonlinear activation functions are commonly preferred since linear functions can limit a model’s learning capacity. Nonlinear activation functions can either have fixed parameters, which are predefined before training, or adjustable ones that modify during training. Fixed-parameter activation functions require the user to set the parameter values prior to model training. However, finding suitable parameters can be time-consuming and may slow down the convergence of the model. In this study, a novel fixed-parameter activation function is proposed and its performance is evaluated using benchmark MNIST datasets, demonstrating improvements in both accuracy and convergence speed.<br /><br />Abstract— Activation functions are essential for extracting meaningful relationships from real-world data in deep learning models. The design of activation functions is critical, as they directly influence the performance of these models. Nonlinear activation functions are commonly preferred since linear functions can limit a model’s learning capacity. Nonlinear activation functions can either have fixed parameters, which are predefined before training, or adjustable ones that modify during training. Fixed-parameter activation functions require the user to set the parameter values prior to model training. However, finding suitable parameters can be time-consuming and may slow down the convergence of the model. In this study, a novel fixed-parameter activation function is proposed and its performance is evaluated using benchmark MNIST datasets, demonstrating improvements in both accuracy and convergence speed.https://jicse.sbu.ac.ir/article_106296_18a39d5e9db0dfef18891ecfe5f2c41c.pdfShahid Beheshti UniversityJournal of Innovations in Computer Science and Engineering (JICSE)2981-21353120250601Persian Intelligent Assistant in Healthcare Domain596410629710.48308/jicse.2025.241738.1087ENSarinaChitsazFaculty of Computer Science and Engineering
Shahid Beheshti University
Tehran, IranMehrnoushShamsfardFaculty of Computer Science and Engineering
Shahid Beheshti University
Tehran, IranJournal Article20250929Abstract—Nowadays, advances in technology and medical science have led to significant changes in the field of healthcare. Consequently, an effort has been taken to develop an intelligent health assistant in the Persian language, focusing on the emergency department. To achieve this goal, a labeled dataset was prepared. Subsequently, an intelligent assistant architecture was developed, utilizing slot filling and speech act classification for natural language understanding. A dialogue manager was designed to address negation in patient statements, resulting in the classification of triage patients. Evaluation revealed that the assistant's performance matched that of emergency staff in 83% of cases.<br /><br />Abstract—Nowadays, advances in technology and medical science have led to significant changes in the field of healthcare. Consequently, an effort has been taken to develop an intelligent health assistant in the Persian language, focusing on the emergency department. To achieve this goal, a labeled dataset was prepared. Subsequently, an intelligent assistant architecture was developed, utilizing slot filling and speech act classification for natural language understanding. A dialogue manager was designed to address negation in patient statements, resulting in the classification of triage patients. Evaluation revealed that the assistant's performance matched that of emergency staff in 83% of cases.https://jicse.sbu.ac.ir/article_106297_04b752ec5fc4c7161f0f8de87dbb0800.pdfShahid Beheshti UniversityJournal of Innovations in Computer Science and Engineering (JICSE)2981-21353120250601Improving Predicted Answer Accuracy in Visual Question and Answer Systems using Attention Mechanisms and Neural Networks657610653010.48308/jicse.2025.241646.1083ENSoheilaKarbasiGolestan UniversityFatemehRezaeiGolestan UniversityJournal Article20250922In recent years, one of the most widely studied areas in computer vision and natural language processing (NLP) is the interdisciplinary problem of Visual Question Answering (VQA), which involves the integration of computer vision and NLP.<br /><br />Important challenges in this field include the need for large and suitable datasets as well as powerful hardware for training the model. Key factors to improve the performance of these models include selecting the appropriate neural network for processing the inputs, selecting the appropriate dataset, and the method of combining the features extracted from the inputs. Also, using different attention mechanisms can improve the overall performance of the system. Furthermore, incorporating various attention mechanisms into the model can significantly enhance the overall performance of VQA systems. In these systems, different neural networks are employed to process inputs: convolutional neural networks (CNNs) with various architectures are used for image processing, and different types of recurrent neural networks (RNNs) are used for text processing.<br /><br />In this research, the architecture of the convolutional neural network is changed and the self-attention mechanism is used in text processing and the Skipgram language model is used for embedding the input text. The performance of the proposed model is evaluated on two datasets, VQA 1.0 and VQA 2.0. The results show that the proposed model has been able to increase the overall accuracy in the VQA 1.0 dataset to 67.25% and in the VQA 2.0 dataset to 61.57%, which show a significant improvement over the baseline models.https://jicse.sbu.ac.ir/article_106530_0e5b26a97e7c176b5a667596b84877f4.pdfShahid Beheshti UniversityJournal of Innovations in Computer Science and Engineering (JICSE)2981-21353120250601Analyzing Answer-Type Preferences Among Expertise Shapes on Stack Overflow778710694210.48308/jicse.2026.242359.1092ENOmidMohammadi KiaShahid Beheshti Faculty of Computer Science and Engineering, Shahid Beheshti University:G.CMahmoodNeshatiShahid Beheshti Faculty of Computer Science and Engineering, Shahid Beheshti University, Tehran, IranJournal Article20251106Developer expertise is a critical component of community-based question-answering platforms like Stack Overflow. However, expertise is not monolithic. Developers exhibit different "shapes" of expertise, such as deep specialists (I-shaped) or broad generalists with one specialty (T-shaped). While prior research has examined developer expertise and answer quality independently, how different expertise structures influence preferences for specific answer characteristics remains insufficiently understood. This paper investigates the relationship between a developer's expertise shape and their preference for specific answer characteristics.<br /><br />We present a large-scale empirical study of over 48,000 Stack Overflow users, classifying them into I-shaped, T-shaped, Pi-shaped, and Comb-shaped profiles based on the distribution of tag-level reputation, following established expertise-shape modeling approaches. We then analyze the types of answers these user profiles tend to upvote and accept as solutions, focusing on characteristics such as answer length, inclusion of code snippets, use of images, and citation of external references.<br /><br />Using separate analyses for community-level (upvotes) and task-resolution-level (accepted answers) preference signals, our findings reveal distinct and systematic differences across expertise shapes. I-shaped specialists favor technically deep, code-heavy answers, while T-shaped and Comb-shaped experts show a preference for more summarized, conceptual answers that include diagrams or references. These patterns are consistent across robustness checks and sensitivity analyses.<br /><br />The results highlight that answer usefulness is user-dependent rather than universal, and they can help improve expertise-aware answer recommendation systems and foster more effective knowledge sharing on collaborative platforms.Shahid Beheshti UniversityJournal of Innovations in Computer Science and Engineering (JICSE)2981-21353120250601Improving the Accuracy of Spectrum-Based Fault Localization Techniques by Focusing on the Program’s Changes and Dependencies8810810694110.48308/jicse.2026.242589.1096ENFaezeAghazade-ParFaculty of Computer Science and Engineering, Shahid Beheshti UniversityMojtabaVahidi-AslFaculty of Computer Science and Engineering, Shahid Beheshti UniversityJournal Article20251123Spectrum-Based Fault Localization (SBFL) techniques are widely used and studied. These techniques are straightforward, low-cost, and fast in comparison to other fault localization techniques, such as Mutation-Based and Learning-Based ones. However, the accuracy of these techniques is always controversial. The main criticism of these techniques refers to the mere use of coverage and spectra information. Hence, this study seeks to improve SBFL techniques by enhancing their accuracy while maintaining simplicity, aiming to make them applicable for regression faults. To achieve this, the first step involves integrating SBFL techniques with two features: (1) Change and (2) Test Case Weight, which utilizes the same coverage and spectrum data in a novel way. Additionally, the study examines the impact of incorporating Data and Control Dependency into the formula for fault localization accuracy. It also considers the suspiciousness scores generated by SBFL formulae as a feature. <br /><br />Three SBFL techniques—Tarantula, Ochiai, and Jaccard—are applied in this study both as a feature and as a baseline for result comparison. The findings reveal that combining SBFL suspiciousness scores with Change and Test Case Weight significantly enhances performance by identifying at least 27% more faults at the Top-3 ranking. Furthermore, integrating these features with Data Dependency leads to even greater improvements, locating minimum 45% more faults at the Top-3 ranking compared to aforementioned SBFL formulae. Overall, this study emphasizes the limitations of existing SBFL formulae while highlighting the advantages of augmenting SBFL with additional features and repurposing spectral information to achieve greater accuracy in fault localization.