Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 22 3 2025 12 1 Improvement of SIFT matching algorithm for matching visible satellite images using Siamese deep neural network 3 18 FA Ahmadreza Zarei Master, Department of Electrical Engineering, Faculty of Engineering, University of Isfahan, Isfahan, Iran ahmadrezazarei92@gmail.com N Payman Moallem Professor, Department of Electrical Engineering, Faculty of Engineering, University of Isfahan, Isfahan, Iran p_moallem@eng.ui.ac.ir Y 10.66224/jsdp.22.3.3 Matching remote sensing images is a fundamental step in many image processing applications. Unlike regular images, remote sensing images often undergo complex and nonlinear background changes, making them difficult to match. They also pose challenges such as scale variations, rotation, and different viewing angles. One commonly used method for finding corresponding points between images is the Scale-Invariant Feature Transform (SIFT) algorithm; however, it often produces many incorrect matches when applied to such data. In contrast, deep learning-based approaches can extract and compare medium and high-level features for more accurate matching. Inspired by these advances, this work introduces a method that combines the SIFT algorithm with a Siamese deep neural network to improve the matching of remote sensing images. The proposed method modifies the conventional SIFT by adjusting its parameters to increase the proportion of correct to incorrect correspondences. After keypoints are extracted and described, initial correspondences are established. Then, for each matched point, a local patch is extracted based on the keypoint’s position, scale, and orientation. These patch pairs are input to a trained Siamese network that estimates the probability of a correct match. Matches with confidence below a threshold are rejected. This hybrid approach leverages the strengths of both traditional and deep learning-based techniques to enhance accuracy. The proposed approach introduces several key innovations, including optimized keypoint extraction to maximize true matches, patch-based feature representation aligned with local image geometry, and a neural network-based verification step to suppress incorrect matches. Based on experiments conducted on a dataset of 35 pairs of remote sensing images, and comparing the results with the SIFT algorithm and deep learning-based methods, the proposed approach achieved an accuracy of 0.849 by reducing false matches and increasing correct ones. Image Keywords: Image Matching, Remote Sensing, SITF Algorithm, Deep Learning, Siamese Convolutional Neural Networks Matching, Remote Sensing, SITF Algorithm, Deep Learning, Siamese Convolutional Neural Networks. http://jsdp.rcisp.ac.ir/article-1-1415-en.html http://jsdp.rcisp.ac.ir/article-1-1415-en.pdf

Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 22 3 2025 12 1 Analysis of the traffic structure of the roads of Chahar Mahal and Bakhtiari province using data-mining approaches 19 34 FA Vahideh Ahrari Assistant Professor, Department of Computer Science, Faculty of Mathematical Sciences, Shahrekord University, Shahrekord, Iran vahideh.ahrari@sku.ac.ir Y Robab Afshari Assistant Professor, Department of Statistics, Faculty of Sciences, University of Zanjan, Zanjan, Iran afshari@znu.ac.ir N 10.66224/jsdp.22.3.19 Road traffic management is a fundamental and multidimensional challenge within transportation systems, exerting a direct impact on public safety, economic efficiency, and environmental sustainability. The province of Chaharmahal and Bakhtiari, due to its strategic geographical location, plays a critical role in connecting various regions of Iran. Therefore, a precise analysis of the traffic structure of this province’s roadways is essential for improving the quality of data-driven planning and decision-making in the transportation sector. In this study, road traffic-counter data collected during September were utilized. These data include key variables such as the average number of vehicles in five different classes, instances of traffic violations (namely speeding, tailgating, and illegal overtaking), as well as the average speed on various road segments. The data were analyzed using three unsupervised learning algorithms: k-means clustering, hierarchical clustering, and autoencoder-based clustering. To assess the accuracy and performance of these algorithms in segment clustering, three well-established evaluation metrics—Silhouette score, Davies-Bouldin index, and Calinski-Harabasz index—were employed. The results demonstrate that the autoencoder and hierarchical clustering models offer a more accurate classification of road segments compared to the conventional k-means method, revealing latent traffic structures more effectively. Based on the findings, the roads in the province were categorized into two distinct clusters: the first cluster includes segments with the highest traffic volume and the highest rates of speeding and tailgating violations—indicative of risky driving behaviors and elevated accident risk. The second cluster encompasses segments characterized by safer traffic patterns. The primary contribution of this research lies in the integrated application of multiple advanced clustering algorithms alongside diverse performance evaluation metrics, which significantly enhance the precision and robustness of the analysis. Furthermore, the combination of technical and behavioral traffic variables within a unified data-driven framework enables the extraction of deeper insights into traffic behavior patterns. The proposed framework is generalizable to other regions and can serve as a novel model for the intelligent management of roadway networks.By providing accurate and actionable analytical tools, this study has the potential to support decision-makers in raising awareness, optimizing resource allocation, designing safety strategies, and ultimately reducing accident rates. To extend and enrich this line of research, future studies are encouraged to incorporate temporal (e.g., seasonal) analyses, integrate human and environmental variables, and develop hybrid predictive models in the transportation domain. Unsupervised learning, clustering, traffic structure analysis, autoencoder, speeding violation, tailgating, Chaharmahal and Bakhtiari Province. http://jsdp.rcisp.ac.ir/article-1-1435-en.html http://jsdp.rcisp.ac.ir/article-1-1435-en.pdf

Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 22 3 2025 12 1 A Novel Privacy-Preserving Distributed Data Publishing Protocol Based on Probabilistic Models 35 58 FA Elyas Mosayebi M.Sc. in Computer Engineering, University of Guilan, Rasht, Iran elyas.mosayebi@gmail.com N Reza Ebrahimi Atani Associated Professor, Department of Computer Engineering, University of Guilan, Rasht, Iran rebrahimi@guilan.ac.ir Y 10.66224/jsdp.22.3.35 In the era of digital transformation, government agencies and corporations increasingly rely on electronic services, generating vast volumes of sensitive data stored in distributed databases. While these records hold immense potential for knowledge discovery through data mining, their publication or sharing raises critical privacy concerns, particularly when sensitive individual information is at risk. Traditional Privacy-Preserving Distributed Data Publishing (PPDDP) methods rely heavily on Trusted Third-Party (TTP) intermediaries and Secure Multi-Party Computation (SMC), which introduce systemic vulnerabilities such as communication bottlenecks, synchronization failures, insider attacks, and inherent distrust in centralized entities. In healthcare analytics, hospitals leverage patient data to enhance diagnostic precision, optimize clinical workflows, and advance preventive and precision medicine. Yet, reliance on siloed datasets from individual institutions often restricts model generalizability and impedes comprehensive insights into health outcomes. Patient health is a multidimensional construct influenced not only by genetic and biological factors but also by behavioral patterns and socio-environmental determinants. Cross-institutional collaboration integrating diverse datasets from geographically distributed sources is essential to develop robust analytical models. However, such collaboration raises critical privacy concerns, as centralized aggregation of sensitive data risks exposure to breaches or misuse. Our probabilistic framework for privacy-preserving distributed data publishing directly addresses this challenge. By eliminating dependencies on trusted third parties and secure multi-party computation, our approach enables secure, decentralized integration of heterogeneous healthcare data. Through uncertainty-aware probabilistic anonymization and adaptive noise injection, the framework ensures compliance with stringent privacy regulations (e.g., GDPR, CPRA, HIPAA) while preserving the analytical utility required for accurate, actionable health outcome predictions. This balance of utility and privacy empowers researchers to harness the full potential of distributed datasets without compromising individual confidentiality, ultimately fostering innovation in precision medicine and population health management. This paper introduces a novel probabilistic framework for privacy preservation in distributed environments, eliminating dependencies on TTP and SMC. Unlike existing approaches, this method leverages uncertainty-aware probabilistic models to dynamically anonymize and perturb data across distributed nodes while preserving global data utility. First a survey of privacy preservation data publishing methods is presented in this paper and then we discuss about prose and cons of the techniques. After this we present the model and its implementation details. The results obtained by security evaluations shows that the presented method will balance out the privacy security and the accuracy of distributed data better, using the probability model without needing a Trusted Third-Party and Secure Multi-party Computation. Data Mining, Data Publishing, Data sharing, Privacy Preserving, Security. http://jsdp.rcisp.ac.ir/article-1-1467-en.html http://jsdp.rcisp.ac.ir/article-1-1467-en.pdf

Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 22 3 2025 12 1 Identification, detection and classification and of multiclass heterogeneous blood cell series based on the council algorithm and aggregation of tissue descriptors 59 76 FA omid eslamifar Ph.D. Candidate Department of Electrical Engineering, saveh Branch, Islamic Azad University, Saveh, Iran O.eslamifar@iau-saveh.ac.ir N Mohammadreza soltani Assistant Professor Department of Electrical Engineering Khomeinishahr Branch, Islamic Azad University, Isfahan , Iran mrsoltani@iautiran.ac.ir Y seyed Mohamadjalal Rastegar Fatemi Assistant Professor Department of Electrical Engineering, saveh Branch, Islamic Azad University, Saveh, Iran Jalal.pe77@gmail.com N 10.66224/jsdp.22.3.59 Understanding the structural and morphological characteristics of blood cells plays a crucial role in the early diagnosis and treatment of hematological disorders. Manual inspection of blood smears under a microscope is still the standard approach in many laboratories; however, this process is subjective, time-consuming, and highly dependent on the expertise of the hematologist. To overcome these limitations, the present study introduces an intelligent hybrid framework for multiclass classification of heterogeneous blood cells based on the integration of deep learning and metaheuristic optimization techniques. In the proposed approach, the wavelet coefficients of microscopic images are first extracted to capture discriminative frequency-domain features. These coefficients are then fed into a YOLO-based convolutional neural network to detect candidate cell regions and identify spatial characteristics. A customized CNN architecture is subsequently employed for hierarchical feature learning, while a Golden Eagle Optimization (GEO) algorithm is utilized to perform feature selection and dimensionality reduction by eliminating redundant and less informative attributes. To achieve robust decision-making, three classical classifiers Decision Tree (DT), Naïve Bayes (NB), and K-Nearest Neighbors (KNN) are combined through a weighted voting ensemble strategy. The model was trained and validated on a dataset consisting of microscopic images of five major white blood cell types: lymphocytes, monocytes, eosinophils, basophils, and neutrophils. Quantitative evaluation was performed using precision, recall, F1-score, and accuracy metrics. Experimental results demonstrate that the proposed CNN GEO ensemble model achieves an overall accuracy of 95.7% and an average F1-score of 94.9%, outperforming comparable state-of-the-art methods such as CNN+SVM, PSO+KNN, and VGG-16 in both accuracy and computational efficiency. The findings highlight the capability of the proposed system to accurately distinguish among multiple blood cell categories, thereby providing a reliable and automated decision-support tool for early hematological diagnosis. Future work will focus on expanding the dataset and integrating domain adaptation mechanisms to further enhance cross-laboratory generalization. blood cell type, classification, YOLO neural network, golden eagle optimization. http://jsdp.rcisp.ac.ir/article-1-1393-en.html http://jsdp.rcisp.ac.ir/article-1-1393-en.pdf

Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 22 3 2025 12 1 An application of the topological data analysis approach in Persian poetry classification 77 90 FA Naiereh Elyasi Assistant Professor, Faculty of Mathematical sciences and Computer, Kharazmi university, Tehran, Iran elyasi82@khu.ac.ir Y Mehdi Hosseini Moghadam NLP/LLM(Team lead), Oxford university, Oxford, UK m.h.moghadam1996@gmail.com N 10.66224/jsdp.22.3.77 This research delves into authorship attribution through an avant-garde lens, employing Topological Data Analysis (TDA) as a potent instrument to unravel intricate patterns within classical Persian poetry. The focal point of this study is the distinguished works of Ferdowsi and Hafez, two preeminent Persian poets, exploring the latent structures in their verses through the lenses of Persistent Homology and Mapper a pair of TDA methodologies. The discernment between Non-Semantic and Semantic authorship attribution methodologies lays the groundwork, elucidating the significance of capturing structural nuances in textual data. The main focus of this investigation revolves around the deployment of Persistent Homology a cutting-edge technique that transcends traditional text analysis methodologies. It operates in high-dimensional spaces, extracting topological features, and rendering them comprehensible through persistent diagrams. This paper meticulously unpacks the mathematical underpinnings of Persistent Homology, providing a stepwise exposition of its application, focusing on Homology, Simplicial Complex, and Group Theory. These foundational elements converge to empower extracting meaningful topological signatures from the poetic corpus. In tandem, Mapper, another TDA tool, unfolds as a pivotal player in this explorative journey. This algorithmic entity facilitates dimensionality reduction and simplicial complex construction to portray an accurate depiction of the intrinsic topological architecture residing in the dataset. The intricacies of Mapper's workflow from filter function selection to binning and clustering are meticulously detailed, forming a coherent narrative of its operational dynamics. Transitioning from theoretical discourse to practical implementation, this research adopts a case study approach, weaving Ferdowsi and Hafez's poetic masterpieces into the TDA tapestry. Beyond the mere application of algorithms, the study delves into the realm of accuracy assessments, subjecting the Mapper algorithm to rigorous tests, and gauging the precision of its poem classifications within identified clusters. An additional layer of complexity unfolds as the research embraces semantic clustering, elucidating thematic resonances embedded within the verses. The results borne out of this meticulous exploration not only underscore the efficacy of TDA methodologies in unveiling the intricate structures of Persian poetry but also offer a nuanced perspective on their interpretability and utility in the realm of authorship attribution. The poetic narrative, with its semantic richness and structural subtleties, emerges as a fertile ground for the application of TDA, pushing the boundaries of text classification methodologies. This research, therefore, contributes significantly to the evolving discourse on the intersection of literature and data science, offering a profound understanding of how TDA can be wielded as a transformative lens to decipher the profound threads of authorial expression. Topological data analysis, Persistent Homology, Mapper, Persian poems. http://jsdp.rcisp.ac.ir/article-1-1395-en.html http://jsdp.rcisp.ac.ir/article-1-1395-en.pdf

Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 22 3 2025 12 1 Using Majority Voting in Graph Neural Networks for Aspect-Based Sentiment Analysis 91 102 FA Ali Balouchi M.Sc in Computer Engineering, Ilam University, Ilam, Iran a.balouchi@ilam.ac.ir Y Mozafar Bagmohmmadi Associate Professor of Faculty of Computer Engineering, Ilam University, Ilam, Iran mozafar@ilam.ac.ir N Mojtaba Karmi Assistant Professor of Faculty of Computer Engineering, Ilam University, Ilam, Iran m.karami@ilam.ac.ir N 10.66224/jsdp.22.3.91 Aspect-Based Sentiment Analysis (ABSA) is a detailed subdomain of sentiment analysis that focuses on detecting sentiments toward specific aspects of entities, such as product features or service attributes, rather than providing a general sentiment polarity. This granular understanding is essential in domains such as customer feedback evaluation, social media opinion mining, and intelligent recommendation systems. However, capturing the syntactic and semantic dependencies required for accurate ABSA remains a challenge for conventional models. In this study, we propose an ensemble-based approach utilizing Graph Convolutional Networks (GCNs), which are particularly effective in learning structural relationships from sentence-level dependency trees. Our methodology involves the integration of four advanced GCN-based models: DualGCN, RDGCN, SSEGCN, and R-GAT. Each model offers distinct strengths, ranging from dual-graph encoding and reinforcement-driven attention mechanisms to syntax-aware semantic enhancements. These models are trained individually and then aggregated through a majority voting mechanism to create a robust ensemble capable of improved sentiment prediction at the aspect level. The models were evaluated on benchmark datasets including SemEval-2014 (Rest14 and Laptops subsets) and Twitter, covering both formal and informal texts. Extensive preprocessing was conducted to standardize input formats and ensure fair comparison across models. Moreover, training was performed using both GLoVE and BERT embeddings, allowing the ensemble to benefit from a diverse range of semantic features. The proposed majority voting strategy aggregates the predictions of individual models and determines the final sentiment class based on the most frequent output. In case of a tie, the model with the highest validation accuracy takes precedence. This strategy effectively combines the complementary capabilities of multiple GCN variants, leading to improved performance and stability across diverse datasets. Experimental results show that the proposed ensemble method significantly outperforms both baseline models and recent state-of-the-art methods. On the Rest14 dataset, the ensemble achieved an accuracy of 88.47%, improving upon the best recent model (SAGCN + BERT) by +1.34%. On the Laptops dataset, it attained 85.44%, exceeding SAGCN’s 85.12% by +0.32%. Similarly, on the Twitter dataset, our model reached 82.12%, surpassing SAGCN’s 81.45% by +0.67%. Additionally, compared to individual baseline models, the proposed method improved accuracy by 2.15% and F1-score by 2.8% on Rest14, 9.2% and 11.74% on Laptops, and 7.8% and 8.7% on Twitter, respectively. These improvements highlight the robustness of the ensemble in handling varying linguistic structures and domains. We also explored alternative ensemble strategies including weighted voting, neural fusion, and combined embedding approaches, yet none outperformed the majority voting strategy in consistency or accuracy. This further reinforces the effectiveness and simplicity of our proposed method .In conclusion, this research introduces a novel and practical ensemble technique for ABSA using multiple GCN models and a majority voting strategy. The method achieves state-of-the-art accuracy across multiple benchmarks and demonstrates strong generalization, making it a valuable contribution to aspect-level sentiment analysis. Future work may extend this approach to multilingual and domain-specific contexts or integrate large pretrained language models such as RoBERTa or GPT to further enhance contextual understanding. Natural Language Processing, Deep Learning, Graph Neural Networks, Aspect Level Sentiment Analysis, Ensemble Learning. http://jsdp.rcisp.ac.ir/article-1-1431-en.html http://jsdp.rcisp.ac.ir/article-1-1431-en.pdf