Transactions on Machine Intelligence2821-16932420191201Optimal Placement of Protective Relays and Distributed Generation Units in Distribution Networks for Enhancing System Reliability Using the PSO Algorithm20521220610910.47176/TMI.2019.205ENM.KhoddamAssistant Professor, Department of Electrical Engineering, Khomein Branch, Islamic Azad University, Khomein, Iran0000-0003-0070-1643Journal Article20190710The integration of Distributed Generation (DG) into modern distribution networks introduces several operational challenges, particularly in terms of network protection and reliability. One of the key challenges for distribution network operators is the transition from a traditional radial structure to a more complex meshed configuration due to the presence of DG sources. This structural change alters the short-circuit levels, rendering conventional fixed protection settings ineffective. As a result, the placement and coordination of protective devices such as overcurrent relays, reclosers, and fuses must be carefully reconsidered to ensure effective fault detection and system reliability. This paper proposes an optimization-based algorithm for determining the optimal locations for installing protective devices and DG units within a distribution network. The proposed methodology aims to enhance system reliability by strategically placing protection equipment while considering the impact of DG integration. The objective function in this study is to minimize key reliability indices, such as the System Average Interruption Duration Index (SAIDI) and the System Average Interruption Frequency Index (SAIFI), while also reducing network losses. The effectiveness of the proposed approach is evaluated using simulation studies conducted on standard test distribution systems. The results demonstrate that the optimized placement of DG and protection devices significantly improves network resilience, enhances fault detection capabilities, and reduces overall power losses. This study provides a comprehensive framework for improving the reliability and efficiency of distribution networks in the presence of DG, contributing to the development of smarter and more adaptive power systems.https://www.tmachineintelligence.ir/article_206109_1963348bbb92d0242177b46372a39a0e.pdfTransactions on Machine Intelligence2821-16932420191201Breast Cancer Diagnosis Using Scattering Wavelet Transform and Hierarchical Multilayer Perceptron Neural Network21322022076710.47176/TMI.2019.213ENM.NouriFaculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Isfahan, IranN.NematbakhshFaculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Isfahan, IranS.FarrokhiFaculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Isfahan, IranJournal Article20250513Breast cancer has been one of the leading causes of mortality among women in the past decade. Although this type of cancer cannot be prevented due to the unknown nature of its primary causes, early diagnosis can significantly improve a patient's chances of full recovery. Mammography is a well-established tool that aids in the early detection of this disease. Various studies have been conducted to develop breast cancer detection methods; however, these efforts have often failed to achieve sufficient accuracy due to the lack of an effective feature extraction method capable of capturing essential texture characteristics and the absence of a robust classifier. In this study, scattering wavelet transform is employed to extract texture-based features from medical images. The use of multiple features increases the dimensionality of input data for the classifier, necessitating an effective dimensionality reduction approach. To address this, Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) have been applied. Finally, a hierarchical multilayer perceptron (MLP) neural network is utilized as the classifier for cancer detection. To evaluate the proposed method, the Mini-MIAS dataset has been used, achieving an accuracy of 97.57%.https://www.tmachineintelligence.ir/article_220767_5b82807fea535595433cc4dd03e76b4c.pdfTransactions on Machine Intelligence2821-16932420191201Brain Tumor Segmentation in MRI Images Using Transform Domain Methods22123422084910.47176/TMI.2019.221ENA.AshamiPourDepartment of Computer Science, Buin Zahra Branch, Islamic Azad University, Buin Zahra, IranH.YarahmadiAssistant Professor, Department of Computer Science, Buin Zahra Branch, Islamic Azad University, Buin Zahra, IranJ.MohammadzadehAssistant Professor, Department of Computer Science, Buin Zahra Branch, Islamic Azad University, Buin Zahra, IranJournal Article20190804Image segmentation techniques are widely used in medical imaging to isolate homogeneous regions. To date, no complete image segmentation method has been presented that can yield satisfactory results for imaging applications such as brain MRI, brain cancer detection, and others. This study presents a method for brain tumor segmentation in MRI images using transform domain methods and contourlets. First, a multi-resolution representation of the input image is created using the contourlet transform. Then, an 8-dimensional feature vector is extracted for each pixel using inter-resolution and intra-resolution data. The final feature vector's dimensions are reduced using Principal Component Analysis (PCA). Finally, the feature vectors are grouped into discrete clusters for segmentation. The proposed method was implemented on brain images using MATLAB software. The algorithm is computationally simple, yet efficient for brain tumor segmentation in MRI images. Using the eight subbands, as opposed to the conventional wavelet transform that extracts coefficients in only three directions, helps us better identify directional details in the image. Additionally, the use of eight directional features for each pixel allows the extracted details in different subbands to be enhanced in accordance with the direction, making maximal use of the subband correlation. The proposed method does not suffer from the region overlapping problem of active contour methods and can detect the internal areas of large regions. Overall, the proposed algorithm improves performance by six percent compared to the active contour method and by one percent compared to the two-dimensional feature extraction method using wavelet transform.https://www.tmachineintelligence.ir/article_220849_69db1c22b087a2e9ab06db852d93ec05.pdfTransactions on Machine Intelligence2821-16932420191201Comparative Prediction of Epileptic Seizures Based on Phase Synchronization in EEG23524522085210.47176/TMI.2019.235ENS. H.Shokouh-AlaeiFaculty of Engineering, Islamic Azad University, Mashhad branch, Mashhad, IranM. A.KhalilzadehFaculty of Engineering, Islamic Azad University, Mashhad branch, Mashhad, IranJournal Article20190620Epileptic seizure prediction has garnered significant attention in recent years due to its potential to improve patient outcomes and reduce the burden of epilepsy. The integration of advanced machine learning techniques, particularly deep learning, into seizure prediction models has opened new avenues for research. Currently, epilepsy patients who have not achieved complete seizure control face the challenge of sudden and unpredictable epileptic seizures. A method capable of predicting seizure onset could significantly enhance the quality of life for these individuals. The fundamental basis of seizure prediction lies in distinguishing the preictal phase dynamics from other phases. In this study, phase synchronization is employed as an indicator for analyzing interactions between different brain regions and as a reliable metric for identifying the preictal period. Furthermore, seizure prediction horizon and seizure onset time are two critical factors in evaluating seizure prediction methods. Although previous studies have primarily used these temporal parameters for assessment, this paper incorporates them into a neuro-fuzzy model, allowing for adaptive seizure prediction based on patient feedback. The implementation of the proposed model on intracranial EEG signals demonstrated that, across various time window values, sensitivity and specificity exceeded 70%.https://www.tmachineintelligence.ir/article_220852_e9d53751ae42c2e9bea9d0c2b0a2c79c.pdfTransactions on Machine Intelligence2821-16932420191201Classification of Abdominal Electromyogram Signals for Detecting Pregnancy Contractions Using Support Vector Machine in the Wavelet Packet Domain24625222090410.47176/TMI.2019.246ENS. S.HasaniDepartment of Biomedical Engineering, Faculty of Electrical and Computer Engineering, Babol Noshirvani University of Technology, Babol, IranH.Montazery KordyAssistant Professor, Department of Biomedical Engineering, Babol Noshirvani University of Technology, Babol, Iran0000-0002-2010-4945Journal Article20190502One of the early signs of natural labor is the occurrence of contractions in the abdominal region of a pregnant woman. However, the presence of contractions in the uterus alone is not a definitive indicator of the onset of natural labor. One of the recent research topics has been the processing of abdominal electromyogram signals from pregnant women to detect preterm labor. The objective of this paper is to classify abdominal electromyogram signals into two classes: labor contractions and pregnancy contractions, in order to detect preterm labor. Due to the differences in the energy distribution of abdominal electromyogram signals throughout pregnancy, the signals are decomposed by a three-level wavelet packet transform. The energy of the wavelet packets at the final decomposition level is then calculated and used for signal classification. The results show that the support vector machine is capable of distinguishing pregnancy contractions from labor-induced pain with a classification accuracy of 86%, sensitivity of 88%, and specificity of 83%, based on the energy features of the wavelet packet.https://www.tmachineintelligence.ir/article_220904_e5d7ae8c3cc79573e22d3ab10ee8a706.pdfTransactions on Machine Intelligence2821-16932420191201An Improved Algorithm Based on Super-Resolution Techniques in the Frequency Domain for Video Image Processing25326022090510.47176/TMI.2019.253ENH.Mehtari TaheriDepartment of Electronics, Faculty of Electrical Engineering, Shahid Beheshti University, Tehran, IranK.JafariAssistant Professor, Department of Electronics, Shahid Beheshti University, Tehran, IranJournal Article20190416In this paper, a novel method for image quality enhancement based on super-resolution algorithms in the frequency domain is presented. The proposed algorithm improves image quality by amplifying high-frequency components, which are crucial for preserving fine details and sharp edges. Unlike many existing super-resolution techniques that rely on multiple image frames, the proposed approach operates efficiently using only a single input frame. This characteristic not only simplifies implementation but also makes the method applicable in scenarios where acquiring multiple frames is impractical. A significant advantage of the proposed method is its reduced computational complexity compared to traditional super-resolution techniques, which often involve iterative optimization processes or deep learning models requiring extensive training datasets. By leveraging the frequency domain for enhancement, the algorithm achieves superior processing efficiency, making it particularly suitable for real-time applications such as video image processing. In such applications, computational speed is a critical factor, and the ability to enhance image quality without introducing excessive processing delays is highly desirable. To evaluate the effectiveness of the proposed method, extensive experiments were conducted on various image datasets, and the results demonstrate that the algorithm successfully enhances image sharpness while maintaining computational efficiency. The promising outcomes suggest potential applications in medical imaging, surveillance, and satellite image processing, where high-quality image reconstruction is essential.https://www.tmachineintelligence.ir/article_220905_a3d723460d9b5d1f60cac526518eb8c5.pdf