﻿<?xml version="1.0" encoding="utf-8"?><records><record><language>per</language><publisher>  Iranian Research Institute for Electrical Engineering</publisher><journalTitle>فصلنامه مهندسی برق و مهندسی کامپيوتر ايران</journalTitle><issn>16823745</issn><eissn>16823745</eissn><publicationDate>2026-11</publicationDate><volume>24</volume><issue>2</issue><startPage>69</startPage><endPage>82</endPage><documentType>article</documentType><title language="eng">A Chaos-Based Image Encryption Scheme with Simultaneous Permutation-Substitution for Resource-Constrained Platforms</title><authors><author><name>E. Zarei Zefreh</name><email>e.zarei@khc.ui.ac.ir</email><affiliationId>1</affiliationId></author><author><name>M. Kiani</name><email>m.kiani@khc.ui.ac.ir</email><affiliationId>2</affiliationId></author></authors><affiliationsList><affiliationName affiliationId="1">Dept. of Comp. Sc., Khansar Campus, University of Isfahan, Isfahan, Iran</affiliationName><affiliationName affiliationId="2">Dept. of Comp. Sci., Khansar Campus, University of Isfahan, Isfahan, Iran</affiliationName></affiliationsList><abstract language="eng">&lt;p style="direction: ltr;"&gt;Multimedia data such as images are considered a valuable and sensitive resource. Therefore, their safe transmission over public networks formed, such as those composed of internet of things (IoT) devices, is vital. Image encryption techniques play a crucial role in protecting image data from unauthorized access and ensuring privacy. Given the limited computing power, storage capacity, and scarce energy resources of IoT devices, it is vital to develop image encryption algorithms that are computationally affordable for such devices. This paper proposes a chaos-based image encryption scheme that enforces simultaneous permutation-substitution and is designed for resource-constrained IoT devices. In the proposed method, circular shift operation for pixel reorganization and a substitution operation for modifying pixel values in four directions are applied simultaneously. The operations are accelerated through vectorization, achieving both low execution time and a high security level. The algorithm demonstrates significant robustness against various security analyses, achieving a key space of 2&lt;sup&gt;256&lt;/sup&gt;, an entropy greater than 7.99, UACI greater than 33.46, NPCR greater than 99.60, and correlation values close to zero. Furthermore, the proposed algorithm encrypts a 1024&amp;times;1024 image in 0.0362345 seconds on a Raspberry Pi 5 board.&lt;/p&gt;</abstract><fullTextUrl>http://ijece.org/Article/51413</fullTextUrl><keywords><keyword>Image encryption</keyword><keyword> chaotic functions</keyword><keyword> permutation</keyword><keyword> substitution</keyword><keyword> IoT (Internet of Things)</keyword></keywords></record><record><language>per</language><publisher>  Iranian Research Institute for Electrical Engineering</publisher><journalTitle>فصلنامه مهندسی برق و مهندسی کامپيوتر ايران</journalTitle><issn>16823745</issn><eissn>16823745</eissn><publicationDate>2026-11</publicationDate><volume>24</volume><issue>2</issue><startPage>83</startPage><endPage>96</endPage><documentType>article</documentType><title language="eng">HMCNet: A Novel Framework for Multi-View Clustering under Incomplete View Conditions</title><authors><author><name>S. Bahadori</name><email>sondos.bahadori@iau.ac.ir</email><affiliationId>1</affiliationId></author></authors><affiliationsList><affiliationName affiliationId="1" /></affiliationsList><abstract language="eng">&lt;p style="direction: ltr;"&gt;Clustering multi-view data in the presence of incomplete views is one of the major challenges in machine learning and data mining. Existing methods often suffer from performance degradation when faced with high rates of incomplete views and are unable to extract common and consistent features between views. Considering the statistics and objective examples presented, which indicate the high prevalence of incomplete views in real applications (with rates between 10% and 50%) and the severe performance degradation of traditional methods in these conditions, the HMCNet model is introduced in this study. By utilizing the combination of attention module, nonlinear UMAP dimensionality reduction, and probability-based composite loss function, this model is able to preserve the local structures of the data and enhance the consistency between views even in the presence of high incompleteness rates (up to 50%). Extensive experiments were conducted on datasets for evaluation. The results showed that the proposed model consistently outperforms state-of-the-art methods in ACC, NMI, and Purity metrics. Also, the analyses of deletion experiments confirmed the importance of each component in improving the model performance. Overall, the findings of this research show that HMCNet can be used as a general and powerful framework for clustering multi-view data under incomplete view conditions and provide new avenues for real-world applications in various scientific and industrial fields.&lt;/p&gt;</abstract><fullTextUrl>http://ijece.org/Article/52865</fullTextUrl><keywords><keyword>Probabilistic loss function</keyword><keyword> multi-view clustering</keyword><keyword> attention module</keyword><keyword> incomplete views</keyword><keyword> embedding based on nonlinear dimensionality reduction.</keyword></keywords></record><record><language>per</language><publisher>  Iranian Research Institute for Electrical Engineering</publisher><journalTitle>فصلنامه مهندسی برق و مهندسی کامپيوتر ايران</journalTitle><issn>16823745</issn><eissn>16823745</eissn><publicationDate>2026-11</publicationDate><volume>24</volume><issue>2</issue><startPage>97</startPage><endPage>108</endPage><documentType>article</documentType><title language="eng">proposing a new method based on LSTM neural network for anomaly detection in ECG time series data</title><authors><author><name>N.  Ashrafi-Payaman</name><email>ashrafi@khu.ac.ir</email><affiliationId>1</affiliationId></author><author><name>B. Mazaheri</name><email>std_behradmazaheri@khu.ac.ir</email><affiliationId>2</affiliationId></author></authors><affiliationsList><affiliationName affiliationId="1">Engineering School, Kharazmi University, Tehran, Iran</affiliationName><affiliationName affiliationId="2">Dept. of Elec. and Comp. Eng., Faculty of Engineering, Kharazmi University, Tehran, Iran</affiliationName></affiliationsList><abstract language="eng">&lt;p style="text-align: left;"&gt;In this study, a novel method for detecting abnormalities in electrocardiogram (ECG) signals is presented, which is based on deep learning and LSTM neural networks. In this method, first, ECG signals received from the MIT-BIH dataset are preprocessed. Preprocessing includes noise removal, normalization, and segmentation of the signals. Then, key features from each period of cardiac electrical activity, such as P, Q, R, S, T waves and the time intervals between them, are extracted. Subsequently, these features are given as input to the LSTM neural network so that it can learn the temporal dependencies between them. The results show that the proposed model has been able to distinguish normal and abnormal rhythms with high accuracy. This method can be used in medical applications and real-time monitoring of heart health. The proposed model is able to identify 19 arrhythmia classes with 97.33% accuracy using only 9 features and an LSTM model. While other methods using 60 features and more complex models have been less accurate.&lt;/p&gt;</abstract><fullTextUrl>http://ijece.org/Article/52877</fullTextUrl><keywords><keyword>anomaly detection</keyword><keyword> Long Short-Term Memory</keyword><keyword> ٍژۀ</keyword></keywords></record><record><language>per</language><publisher>  Iranian Research Institute for Electrical Engineering</publisher><journalTitle>فصلنامه مهندسی برق و مهندسی کامپيوتر ايران</journalTitle><issn>16823745</issn><eissn>16823745</eissn><publicationDate>2026-11</publicationDate><volume>24</volume><issue>2</issue><startPage>109</startPage><endPage>118</endPage><documentType>article</documentType><title language="eng">Generating Minimal Test Suite by Combining Gravitational Search and Harris Hawks Optimization Algorithms</title><authors><author><name>S. Esfandyari</name><email>esfandyari@malayeru.ac.ir</email><affiliationId>1</affiliationId></author><author><name>D. Giveki</name><email>davar.giveki@malayer.ac.ir</email><affiliationId>2</affiliationId></author><author><name>L. Yousofvand</name><email>l.yousofvand@gmail.com</email><affiliationId>3</affiliationId></author><author><name>P. Einollah</name><email>pira@azaruniv.ac.ir</email><affiliationId>4</affiliationId></author></authors><affiliationsList><affiliationName affiliationId="1">Dept. of Comp. Eng., Malayer University, Malayer, Iran</affiliationName><affiliationName affiliationId="2">Dept. of Comp. Eng., Malayer University, Malayer, Iran</affiliationName><affiliationName affiliationId="3">Dept. of Comp. Eng., Faculty of Engineering, Lorestan University, Korramabad, Iran</affiliationName><affiliationName affiliationId="4">Faculty of Info. Tech. and Comp. Eng., Azarbaijan Shahid Madani University, Tabriz, 5375171379, Iran</affiliationName></affiliationsList><abstract language="eng">&lt;p style="direction: ltr;"&gt;Nowadays, combinatorial testing has become essential for the development of modern software systems. Combinatorial testing, also known as t-way testing, is recognized as an efficient approach to reducing the size of a test suite, enabling software testing with a smaller number of test cases. Covering arrays are a crucial component of combinatorial testing, and various approaches have been proposed for their generation. These methods often achieve acceptable performance in terms of efficiency (generation speed) and effectiveness (array size); however, there remains a need for an approach that simultaneously achieves high efficiency and effectiveness. In this study, a hybrid approach is proposed by combining the Gravitational Search Algorithm (GSA) with the Harris Hawks Optimization (HHO) algorithm to generate an optimized test suite with a minimum number of test cases and high computational efficiency. Furthermore, the proposed approach is capable of generating test suites for configurations with high interaction strengths. The evaluation results demonstrate that the proposed method performs effectively and efficiently, offering a powerful solution for combinatorial test generation.&lt;/p&gt;</abstract><fullTextUrl>http://ijece.org/Article/46218</fullTextUrl><keywords><keyword>Combinatorial testing</keyword><keyword> covering array</keyword><keyword> gravitational search algorithm</keyword><keyword> Harris hawks optimization olgorithm.</keyword></keywords></record><record><language>per</language><publisher>  Iranian Research Institute for Electrical Engineering</publisher><journalTitle>فصلنامه مهندسی برق و مهندسی کامپيوتر ايران</journalTitle><issn>16823745</issn><eissn>16823745</eissn><publicationDate>2026-11</publicationDate><volume>24</volume><issue>2</issue><startPage>119</startPage><endPage>127</endPage><documentType>article</documentType><title language="eng">Adaptive Frame-Level Quantization Parameter Adjustment Based on Human Visual System Characteristics in Video Coding</title><authors><author><name>M. Kalantari</name><email>sarakalantari1994@gmail.com</email><affiliationId>1</affiliationId></author><author><name>M. Khademi</name><email>khademi@um.ac.ir</email><affiliationId>2</affiliationId></author><author><name>A. Ebrahimi Moghadam</name><email>a.ebrahimi@um.ac.ir</email><affiliationId>3</affiliationId></author></authors><affiliationsList><affiliationName affiliationId="1">Dept. of Elec. Eng., Faculty of Engineering, Ferdowsi University of Mashhad, Mashhad, Iran</affiliationName><affiliationName affiliationId="2">Dept. of Elec. Eng., Faculty of Engineering, Ferdowsi University of Mashhad, Mashhad, Iran</affiliationName><affiliationName affiliationId="3">Dept. of Elec. Eng., Faculty of Engineering, Ferdowsi University of Mashhad, Mashhad, Iran</affiliationName></affiliationsList><abstract language="eng">&lt;p&gt;&lt;br data-start="332" data-end="335" /&gt; Video coding plays a key role in reducing data volume and optimizing the transmission and storage of multimedia content. In recent years, perceptual video coding has emerged as an effective approach for reducing bitrate without introducing perceptible visual degradation by exploiting characteristics of the HVS. In this paper, a perceptual method for adaptive adjustment of the quantization parameter in video coding with a hierarchical temporal structure is proposed. The proposed method employs a just noticeable difference model to estimate the perceptual sensitivity of each frame and adjusts the &amp;Delta;QP according to content characteristics and temporal importance, such that frames with lower perceptual sensitivity are more strongly compressed while perceptually sensitive frames preserve their visual quality.&lt;br data-start="1178" data-end="1181" /&gt; Experimental results on standard video sequences using the HEVC encoder show that the proposed method achieves an average bitrate reduction of 2.5% together with a 1.67% reduction in encoding time. In addition, subjective evaluation based on the MOS demonstrates an average improvement of 0.17 compared to the reference encoder, with equal or better visual quality achieved in most sequences. These results indicate that the proposed JND-based adaptive &amp;Delta;QP strategy is an efficient, low-complexity, and HEVC-compatible solution for practical video coding applications.&lt;/p&gt;</abstract><fullTextUrl>http://ijece.org/Article/52661</fullTextUrl><keywords><keyword>Video Coding</keyword><keyword> Perceptual Video Coding</keyword><keyword> Quantization Parameter (QP)</keyword><keyword> Just Noticeable Difference (JND)</keyword><keyword> Hierarchical Temporal Structure</keyword><keyword> High Efficiency Video Coding (HEVC).</keyword></keywords></record><record><language>per</language><publisher>  Iranian Research Institute for Electrical Engineering</publisher><journalTitle>فصلنامه مهندسی برق و مهندسی کامپيوتر ايران</journalTitle><issn>16823745</issn><eissn>16823745</eissn><publicationDate>2026-11</publicationDate><volume>24</volume><issue>2</issue><startPage>128</startPage><endPage>134</endPage><documentType>article</documentType><title language="eng"> Increasing Security in Internet of Things Networks by Transfer Learning</title><authors><author><name>A. Samiee</name><email>arman_samiee@elec.iust.ac.ir</email><affiliationId>1</affiliationId></author><author><name>E. Khedersolh</name><email>erfan_khedersolh@elec.iust.ac.ir</email><affiliationId>2</affiliationId></author><author><name>H .S. Shahhoseini</name><email>shahhoseini@iust.ac.ir</email><affiliationId>3</affiliationId></author></authors><affiliationsList><affiliationName affiliationId="1">School of Electrical Engineering, Iran University of Science and Technology, Tehran, Iran</affiliationName><affiliationName affiliationId="2">School of Electrical Engineering, Iran University of Science and Technology, Tehran, Iran</affiliationName><affiliationName affiliationId="3">School of Electrical Engineering, Iran University of Science and Technology, Tehran, Iran</affiliationName></affiliationsList><abstract language="eng">&lt;p&gt;Intrusion Detection Systems (IDS) play a fundamental role in defending computer networks by identifying unauthorized access and investigating potential security breaches. Traditional IDS solutions struggle to counter modern, sophisticated security attacks. Consequently, there is a growing need for automated intrusion detection solutions capable of identifying malicious activities with greater accuracy. In this paper, we propose a novel intrusion detection model for Internet of Things (IoT) networks, utilizing a neural network based on the pre-trained VGG16 architecture. The model employs a two-stage process: first, we leverage a VGG16 network&amp;mdash;pre-trained on the ImageNet dataset&amp;mdash;by transferring its learned weights; second, we perform fine-tuning specifically tailored for intrusion detection within IoT networks. This is necessary because IoT networks possess unique vulnerabilities&amp;mdash;stemming from resource constraints, specialized protocols, and device heterogeneity&amp;mdash;that require dedicated intrusion detection approaches. The model was evaluated using the UNSW-NB15 and NSL-KDD datasets. The results demonstrate that the proposed method effectively learns to detect intrusions, yielding strong performance in terms of Accuracy, Precision, Recall, and F1-score.&lt;/p&gt;</abstract><fullTextUrl>http://ijece.org/Article/48089</fullTextUrl><keywords><keyword>Internet of things</keyword><keyword> intrusion detection systems</keyword><keyword> transfer learning</keyword><keyword> deep learning.</keyword></keywords></record></records>