Transactions on Machine Intelligence2821-16932220190601Speckle Noise Reduction in Optical Coherence Tomography Images Using a Combination of Edge-Preserving Filters and Discrete Wavelet Transform597520602210.47176/TMI.2019.59ENA.Fahmi JafarqulukhaniDepartment of Biomedical Engineering-Bioelectric, Faculty of Biomedical Engineering, Sahand University of Technology, Tabriz, Iran0000-0002-1661-4475M.GhiyamiAssitant Professor, Department of Electrical Engineering, Ardabil Branch, Islamic Azad University, Ardabil, Iran0000-0001-8274-6743S.Jafari NaminDepartment of Electrical Engineering, Ardabil Branch, Islamic Azad University, Ardabil, IranJournal Article20190410Imaging techniques are repeatedly employed by radiologists to detect internal structural details and body function. Optical coherence tomography (OCT) is a significant non-invasive medical imaging modality used to examine the microstructure of biological tissues. In this study, images from 500 patients who visited Noor Clinic in Ardabil for OCT imaging of the retina were collected. Since OCT images suffer from speckle noise, edge-preserving filters were utilized for noise reduction. By subtracting the output of the filtered image from the original image, some of the original image information and noise remained in the output. To restore the remaining information, the output resulting from subtracting the original image from the filtered image was decomposed using the discrete wavelet transform and soft thresholding, and the remaining image information was added to the filtered image. Among the edge-preserving filters, the guided filter demonstrated the best performance with MSE, PSNR, and IQI values of 79.949, 28.41, and 0.984, respectively, whereas the proposed method achieved values of 19.825, 33.78, and 0.989, respectively. After extracting the image information and restoring it to the output of the edge-preserving filters, significant changes in quantitative noise reduction metrics were observed. Consequently, in this study, by employing guided, bilateral, and db8 wavelet filters, we achieved better performance in preserving the edges of OCT images compared to edge-preserving filters.https://www.tmachineintelligence.ir/article_206022_4f2a297e3cc44ddd8f661304834ec3c4.pdfTransactions on Machine Intelligence2821-16932220190601Novel Three-port Buck-Boost DC-DC Converter for Hybrid Vehicle Powertrains Capable of Returning Regenerative braking Energy768720603510.47176/TMI.2019.76ENH.Soltani GohariDepartment of Electrical Engineering, K.N.Toosi University of Technology, Tehran, IranK.AbbaszadehProfessor, Department of Electrical Engineering, K.N.Toosi University of Technology, Tehran, Iran0000-0002-5388-8280Journal Article20190309The paper introduces a novel three-port buck-boost DC-DC converter designed for application in hybrid electric vehicle (HEV) powertrains. Notably, the converter's unique feature lies in its ability to function as a buck-boost converter in two directions, facilitating the recovery of regenerative braking energy during diverse driving cycles characterized by varying braking durations. The operating modes of the converter encompass charging, propulsion, and regenerative braking. While the battery can be charged using any DC power source such as a photovoltaic (PV) panel or rectified grid voltage, the paper specifically assumes a PV panel as the power source. This emphasis underscores the converter's versatility in accommodating a broad range of voltages. In propulsion mode, the converter can operate with one or two power sources in accordance with commands from the energy management system. To validate the system analysis, the proposed converter undergoes simulation in MATLAB software, covering different conditions within each operating mode. The simulation results are meticulously analyzed, providing a comprehensive understanding of the converter's performance under varied scenarios. This research contributes to advancing the capabilities of HEV powertrains, offering a versatile and efficient solution for energy management and regenerative braking in diverse driving conditions.https://www.tmachineintelligence.ir/article_206035_aed8a4da83f2ecd3aed1d52d7a9e916c.pdfTransactions on Machine Intelligence2821-16932220190601Short-Term Load Forecasting of Distribution Networks Using a Hybrid Method of Wavelet Transform and Neural Networks Based on Bacterial Foraging Optimization Algorithm889820603610.47176/TMI.2019.88ENA.Zarif LouloueiDepartment of Electrical Engineering, Pardis Branch, Islamic Azad University, Tehran, IranN.KuchmeshkiDepartment of Computer Engineering, Arak Branch, Islamic Azad University, Arak, IranJournal Article20190307The effective application of consumption management in electrical distribution systems is expected to result in uniform load curves in the future, promoting the efficient utilization of resources. A critical aspect of consumption management is the precise prediction of electrical load in local networks, which consist of a diverse mix of residential, commercial, and industrial consumers. This paper proposes a hybrid approach combining neural networks and wavelet transform to accurately forecast the load of distribution networks. The model utilizes electrical load data from the Qom province distribution network to train and evaluate the prediction system. The wavelet transform is employed for multi-resolution analysis, enabling the model to capture both short-term and long-term patterns in the load data. The neural network's parameters, including the filter type, window length (the number of past data points used for forecasting), and the number of hidden layers, are optimized using the E. coli bacterial foraging algorithm. This optimization technique helps minimize forecasting errors by identifying the most effective configuration for the neural network model. The proposed hybrid model aims to improve forecasting accuracy compared to traditional methods by effectively addressing the complexities of load prediction in distribution networks. The results demonstrate the potential of this integrated approach for enhancing load forecasting and supporting more efficient consumption management in local electrical grids.https://www.tmachineintelligence.ir/article_206036_999f7619e44409564b262358ed9019e7.pdfTransactions on Machine Intelligence2821-16932220190601A Low-Cost 16×16 and 32×32 DCT Architectures for HEVC Application Using Configurable Constant Multipliers9910720603710.47176/TMI.2019.99ENR.YounesiDepartment of Electrical Engineering, Saveh Branch, Islamic Azad University, Saveh, IranS. M. J.Rastegar FatemiDepartment of Electrical Engineering, Saveh Branch, Islamic Azad University, Saveh, IranM.RastegarpourDepartment of Computer Engineering, Saveh Branch, Islamic Azad University, Saveh, IranJournal Article20190202This paper introduces a novel area-efficient Discrete Cosine Transform (DCT) architecture designed for the High Efficiency Video Coding (HEVC) standard, a recently introduced international video compression standard. The DCT architecture is capable of performing 16 and 32-point DCT computations, which are essential for HEVC applications. Given the high complexity associated with larger transform sizes in the HEVC standard, the focus of this paper is primarily on developing efficient solutions for these larger point DCT transform sizes. The key innovation lies in leveraging commonality in constant multiplications through the use of configurable constant multipliers, aiming to reduce the area cost and hardware utilization. Consequently, the proposed architecture exhibits a significant reduction in the number of adder and shift blocks compared to existing architectures. Experimental results, conducted for the 90-nm technology node, indicate a remarkable 42% reduction in area consumption compared to other architectures. Furthermore, the proposed DCT architecture is demonstrated to support real-time 4K video resolution.https://www.tmachineintelligence.ir/article_206037_4697756f2f3c2945e6cf2a5e194183bf.pdfTransactions on Machine Intelligence2821-16932220190601Analysis of the Impact of Electricity Price Variations on Self-Healing Improvement in Smart Distribution Networks Considering Consumer Behavior10811920603910.47176/TMI.2019.108ENS. Karimi TalkhounchehS. Karimi TalkhounchehMSc Student, Faculty of Electrical Engineering, K. N. Toosi University of Technology, Tehran, IranA.FereidunianAssistant Professor, Faculty of Electrical Engineering, K. N. Toosi University of Technology, Tehran, Iran0000-0002-3879-9853A.MoshariAssistant Professor, Department of Power System Planning and Operation, Niroo Research Institute, Tehran, Iran0000-0002-1507-9124Journal Article20190203Self-healing is a capability of smart electricity distribution networks that enables the automatic restoration of the network in the event of a persistent fault. To restore the power supply to loads downstream of the fault location, adjacent feeders can be utilized. However, the use of backup feeders is subject to network technical constraints such as bus voltage levels and permissible line currents. One method to prevent the violation of network constraints during load restoration via backup feeders is the implementation of demand response programs. This paper examines the impact of Critical Peak Pricing (CPP), a price-based program, on the self-healing of smart distribution networks. Two models, exponential and linear, have been considered for modeling the CPP program. Additionally, the impact of different price surges on self-healing improvement has been analyzed. To make the results more realistic, various consumer participation rates in demand response programs have been taken into account. The proposed model has been evaluated using bus number 4 of the Roy-Billinton test system.https://www.tmachineintelligence.ir/article_206039_e2a7ec7f0295916cd5aa4eb9dd436a67.pdfTransactions on Machine Intelligence2821-16932220190601Design of a New Triple-Band Frequency Selective Surface for Filtering WiMAX, WLAN, and X Bands12012720604010.47176/TMI.2019.120ENM.BashiriAssistant Professor, Department of Electrical Engineering, Faculty of Engineering, Islamic Azad University, Miandoab Branch, Miandoab, IranJournal Article20190308This paper proposes a novel single-layer Frequency Selective Surface (FSS) designed on an FR4 substrate with a thickness of 1.6 mm for triple-band filtering applications. The proposed FSS structure consists of unit cells with dimensions of 10×10 mm². The design incorporates a square ring at the top layer, accompanied by two symmetrical rectangular rings at the sides and a vertical rectangular branch placed in the center of the rectangular rings on the bottom layer. The conductive elements are made from a perfect conductor to achieve optimal performance. The structure operates across three distinct frequency bands: 3.2 to 3.81 GHz, 4 to 5.82 GHz, and 8.21 to 12.2 GHz, which correspond to WiMAX, WLAN, and X bands, respectively. This triple-band functionality makes the proposed FSS suitable for applications in modern wireless communication systems. The frequency response of the FSS remains stable across various incidence angles for both Transverse Electric (TE) and Transverse Magnetic (TM) polarizations, confirming the robustness and reliability of the structure. Key advantages of the proposed FSS include its smaller physical dimensions compared to previous designs, the simplicity of its conductive elements, and its alignment with widely used frequency bands. These features collectively enhance the potential of the FSS for integration in compact and efficient filtering systems, offering significant improvements over traditional multi-band filters.https://www.tmachineintelligence.ir/article_206040_ec2322d93f2e80f56ef5770ff001d445.pdf