Abdolhossein Fathi

  Abstract Multiple sclerosis (MS), as a chronic and disabling disease of the central nervous system, has created many challenges in the field of diagnosis and treatment for doctors and health systems. Rapid and accurate identification of lesions caused by this disease in MRI images due to structural similarities with other brain tissues requires the use of accurate and advanced image processing and machine learning methods. In this study, an optimized model called DE-CNN-Gray is presented for automatic diagnosis of MS from gray-scale MRI images. In this model, a convolutional neural network is first designed and then the network structure including the number of layers and effective parameters is optimized using the Differential Evolution algorithm. The main goal of this optimization was to increase the classification accuracy and reduce the computational complexity of the model. Model evaluation using 5-Fold validation showed that the proposed method performed very well in identifying MS patients with an accuracy of 99.40%, sensitivity of 98.89%, positive accuracy of 99.90%, and F1 score of 99.33%. The results show that the DE-CNN-Gray method, using gray images and meta-heuristic algorithms, can be used as an accurate, fast, and low-cost tool for developing MS diagnosis systems and play an effective role in improving the treatment process and reducing treatment costs . Keywords: MS, Deep Learning, Convolutional Neural Network, Differential Evolution Algorithm, MRI, DE-CNN-Gray,Medical Diagnosis, Image Processing.

   With the rapid development of smart cities and the growing need for accurate and real-time analysis of road infrastructure, the design of AI-based systems capable of perceiving, analyzing, and recording environmental data has become increasingly crucial. In this regard, the present study focuses on the design and implementation of an innovative system for the automatic detection, tracking, and localization of traffic signs. This system not only pushes technical boundaries but also makes a significant contribution to the localization of traffic-related data. The advanced YOLOv9 model is employed for precise traffic sign detection, while the powerful ByteTrack algorithm ensures continuous tracking. What truly distinguishes this research is the novel application of federated learning using the FedAvg algorithm—implemented for the first time in the domain of traffic sign recognition. This method enables the training of models on heterogeneous datasets, including two distinct subsets, DFG and Mapillary, without requiring physical data aggregation. This approach not only preserves data privacy but also significantly enhances the generalization capability of the model. On the data side, the study introduces a rich and unprecedented dataset comprising 14,111 images and over 19,000 traffic sign instances across 118 classes. The data was collected over two years in varying temporal conditions (morning, noon, evening, night) and all four seasons, spanning urban, rural, and interurban areas across the country using mobile phone cameras. The images were meticulously annotated using the Makesense tool in both YOLO (.txt) and Pascal VOC (.xml) formats. The system’s performance, evaluated through 6-Fold Cross Validation, demonstrates its high accuracy, achieving a remarkable mAP50 of 95.66%. This not only reflects the model's robustness in real-world conditions but also shows a clear advantage over various versions of YOLO from v5 to v11. The initial idea for this project stemmed from a proposal by Tehran Municipality to design a digital map of traffic signs. However, the outcomes of this research go far beyond a municipal application and offer valuable tools for navigation systems, intelligent vehicles, spatial analytics, and the development of a national traffic sign map in Iran. This work presents a seamless integration of cutting-edge technology, locally-driven data, and modern AI architectures to pave the way for a smarter and safer future on the country’s roads.

Sleep is one of the fundamental human needs that significantly impacts physical and mental health. Stress during sleep can lead to sleep disorders and related health issues, making accurate prediction of sleep stress particularly important. This thesis explores the detection of sleep stress using deep learning, specifically focusing on LSTM and GRU recurrent neural network models, as well as a hybrid model combining the two. The aim of this research is to provide an efficient and accurate model for predicting and detecting sleep stress based on the SaYoPillow dataset. We used 10-fold Cross Validation. Various models were evaluated, and the results showed that the hybrid Bidirectional LSTM-GRU model achieved the best performance with an accuracy of 1.00, precision of 1.00, recall of 1.00, and an F1 score of 1.00, outperforming individual LSTM and GRU, and MLP models in detecting all 5 levels of sleep stress. The use of a confusion matrix and evaluation metrics such as accuracy, precision, recall, and F1 score demonstrated that the hybrid model not only has high accuracy in detecting positive cases but also reduces errors related to identifying negative cases. This research highlights that deep learning models, particularly the hybrid Bidirectional LSTM-GRU model, can be effective tools for detecting sleep stress, thereby contributing to improved sleep quality and overall health. The development of these models can assist healthcare professionals in providing appropriate preventive and therapeutic strategies for managing sleep stress.

   Epilepsy is a common neurological disorder characterized by recurrent seizures. Research indicates that approximately 30% of epilepsy patients are resistant to pharmaceutical treatments or surgical interventions. Abnormal brain activity, known as the pre-ictal state, typically begins a few minutes before a seizure occurs. Electroencephalography (EEG) is a practical technique for recording brain electrical activity and aiding in the diagnosis of epilepsy. Seizure prediction and assistance for epilepsy patients remain significant challenges in preventing seizure-related complications and improving the quality of life for individuals affected by this condition. Accurate prediction of the onset of the pre-ictal state can help reduce the adverse effects of seizures for patients and their caregivers by providing timely care. The objective of this thesis is to develop a system that enhances evaluation metrics for seizure prediction using deep learning methods. In this study, the CHB-MIT dataset, comprising scalp EEG signals, has been utilized, and the proposed method was evaluated on 24 patients from this dataset. To predict seizures, deep learning-based models and knowledge distillation techniques were employed for model compression, aiming to reduce time and hardware costs and enable real-time application of the network. The teacher model, designed as a patient-independent framework with 22 channels and preprocessed mel-spectrogram inputs, employs a 3D convolutional neural network. This model achieved an accuracy of 87.52%, sensitivity of 88.82%, specificity of 85.97%, and an F1 score of 86.56%. Subsequently, the knowledge distillation technique was applied. By utilizing this approach and employing a single electrode, we identified two electrodes (Electrode 20 and Electrode 22) with superior performance compared to others. The proposed method, for Electrode 20, achieved accuracy84.56%, sensitivity86.76%, specificity82.77%, and F1-score values of 83.63%, and for Electrode 22, achieved accuracy84.30%, sensitivity86.45%, specificity82.93%, and F1-score values of 83.35%, enabling seizure prediction 30 minutes before onset. The results obtained from our proposed method were compared with advanced seizure prediction techniques. The proposed method demonstrated superior performance in terms of accuracy, sensitivity, specificity, and F1 score, highlighting its effectiveness in seizure prediction.

Abstract Today, with the growing complexity of digital circuits and the increasing compactness of manufacturing technologies, the likelihood of failure during both the manufacturing process and the operation of digital circuits has risen. As a result, these products require testing both during production and operation to ensure proper performance. Therefore, producing a high-quality test with a minimal number of test vectors and in the shortest time is crucial. In this thesis, a simulation-based test generation method for combinational circuits is proposed. This method utilizes an approximate criterion called approximate critical path tracing with parallel patterns to evaluate the effectiveness of candidate test vectors. This criterion is based on the traditional critical path tracing method, but introduces an approximate approach for backtracking in fan-outs. By allowing some inaccuracies in the results, this method reduces the complexity of fault simulation compared to traditional methods. The parallel pattern critical path tracing method leverages both fault-level and pattern-level parallelism, resulting in significantly higher simulation speed compared to traditional methods for determining fault coverage. Applying the parallel patterns approximate critical path tracing algorithm to ISCAS85, ISCAS89, and ITC99 benchmark circuits shows that, in over 98% of the circuits, the results strongly correlate with the exact fault coverage index. Additionally, compared to the parallel pattern fault simulation method, this approach is more than 500 times faster. The proposed test generation method, which utilizes the criterion derived from the parallel pattern approximate critical path tracing algorithm to identify effective test vectors, is capable of producing high-quality test vectors quickly. Evaluations on benchmark circuits demonstrate that this method is over 15 times faster than the one using parallel pattern fault simulation, with only a 1% average increase in the number of test vectors.    Key words: Digital circuits testing, Fault coverage, Simulation-based test-pattern generation, Approximate fault coverage index, Critical path tracing.   

   Parkinson's disease is a brain disorder caused by damage to dopamine producing cells in the brain. People with Parkinson's disease have symptoms such as tremors and slowness of movement, which makes it difficult for these people to control their movements. Parkinson's is usually diagnosed based on tests done by a neurologist. Actions such as; Analysis of the patient's medical history, examination of symptoms, neurological and physical examination. Therefore, the identification of Parkinson's disease is a long-term process that always requires the availability of all the patient's information (history) and their careful study in each session. Therefore, according to the conditions and problems that exist in this field, misdiagnosis is among the possibilities according to its risks. One of the solutions used today to prevent such mistakes is the use of automatic machine learning detection systems. Considering the mentioned issues and problems, this study tests a two-way LSTM model with two activation functions, Softsign and Tanh, for the automatic diagnosis of Parkinson's disease based on the gait analysis of PD people. The raw data of VGRF signals obtained from the Physionet database were tested in the proposed model to classify PD and healthy subjects. Experiments show the high efficiency of the proposed method in diagnosing Parkinson's disease based on the analysis of movement signals related to people's walking. The proposed algorithm achieved 97.1% accuracy. Among the methods investigated in this study, the presented method has obtained the best performance in the diagnosis of Parkinson's disease using movement signals related to walking. These results show that this model can learn efficient features from existing data that can be useful in clinical diagnosis.

     Medical imaging plays an important role in clinical diagnosis and treatment. Medical imaging is a way to show the anatomical structures of the body with the help of X-rays, which is obtained from computed tomography and magnetic resonance imaging. But often this type of photography is more suitable for physiological function than anatomy. With the development of computer and imaging technology, medical imaging has greatly affected the medical field. Since the quality of medical imaging has had a great impact on disease diagnosis, medical image processing has become one of the most important clinical applications that store and retrieve images for the future, which are prerequisites for accurate storage of these images. Bones are solid organs in the human body that protect many important organs such as the brain, heart, lungs and other internal organs. The human body has 206 bones of different shapes, sizes and sizes. The largest bone is the femur, and the smallest is the ossicle. A common problem in humans is "bone fracture". A bone fracture can be caused by an accident or any other case where a lot of pressure is applied to the bone. There are different types of bone fractures: oblique, compound, comminuted, spiral, girrin's stick1 and transverse. Compared to other methods, X-ray imaging provides precise details of bones and less details of tissue and muscle, which makes it easier to detect fractures

Verifying a person's identity using handwritten signatures is challenging in the presence of a skilled forger, where the forger has access to the person's signature and deliberately tries to imitate it. In offline (static) signature verification, the dynamic information of the linear signature process is lost, and it is difficult to design good feature extractors that can distinguish between genuine signatures and skilled forgeries. A signature is a handwriting of people that has special features and makes each person's signature unique, so a system can be designed to recognize people's signatures and authenticate their identity by means of signatures. One of the machine learning methods that has the appropriate accuracy to detect such projects is convolutional neural networks. In this research, we combined the deep convolutional network model with the federated learning approach, which provides proper accuracy in signature detection. This model recognizes professional forgery signatures with an accuracy of more than 91% and random forgeries with an accuracy of about 97%.  

تشخيص وسيله­­نقليه، يك بخش مهم در حمل و نقل و هوش مصنوعي است. وسايل­نقليه مي­توانند در قسمت­هاي مختلف تصاوير قرار بگيرند. پيشرفت هاي اخير در روش هاي تشخيص، منجر به طيف وسيعي از تكنيك هاي مختلف شده كه مي­تواند براي شناسايي و تشخيص وسايل­نقليه مورد استفاده قرار گيرد. يادگيري عميق، در سال‌هاي اخير است كه كاربردهاي قابل توجهي در روش­هاي تشخيص وسايل نقليه دارد. باتوجه به اهميت تشخيص وسيله­نقليه در سيستم­هاي حمل و نقل هوشمند، در اين پايان­نامه به بررسي و تشريح روش­هاي تشخيص وسايل­نقليه مختلف از تصاوير دوربين­هاي ترافيكي پرداخته­ و نيز از معماري قدرتمندي به نام يولو[1] براي تشخيص وسايل­نقليه روي ديتاست BVMMR استفاده مي­كنيم.­ به دليل تغييرپذيري در محيط‌هاي رانندگي، تشخيص خودرو ممكن است با مشكلات و چالش‌هاي متفاوتي مواجه شود، مثلا ظاهر وسايل­نقليه در اندازه، شكل و رنگ متفاوت، روشنايي خاص، شرايط آب و هوا و.. است. معماري، YOLOv5 شامل چهار بخش اصلي ورودي، backbone ، neck   و خروجي است. ترمينال ورودي عمدتاً شامل پيش پردازش داده ها است، از جمله افزايش داده موزاييك و پر كردن تطبيقي تصوير. شبكه backbone   عمدتاً از يك شبكه جزئي چند مرحله­اي (CSP) براي كاهش مقدار محاسبات و افزايش سرعت استنتاج و   ادغام هرمي فضايي (  ) براي استخراج feature map   با اندازه­هاي مختلف از ورودي تصوير با هدف بهبود دقت تشخيص با كانولوشن چندگانه و pooling استفاده مي­كند. در شبكه neck، از ساختارهاي هرمي ويژگي FPN و PAN استفاده مي­شود. با استفاده از معماري يولو نسخه پنجم[2] آموزش داده شده، موقعيت خودروها و نوع و دسته­ي آن­ها را نيز مشخص كرده­ايم و به   98.88% و دقت مجموع 99.73% و نيز سرعت 0.03 ثانيه براي تشخيص اشيا موجود در يك تصوير دست مي­يابيم كه خود گواهي بر مناسب بودن اين روش براي كاربردهاي بلادرنگ[3] مي­باشد. كلمات كليدي: تشخيص وسايل­نقليه، تشخيص اشيا، يولو، يادگيري عميق، سرعت و دقت بالا در تشخيص اشيا، شناسايي نوع و مدل وسايل نقليه [4]. [1]. you only look once (YOLO) [2] YOLOv5 [3] real-time

Abstract Due to the progress of science and technology in various cultural, social and economic fields, the need to receive data from various databases based on extracting information patterns to achieve life-giving research results is increasing day by day. On the other hand, maintaining the security of private data of individuals and organizations is a key issue in this field, ignoring it will lead to incorrect information and customer dissatisfaction, and will ultimately lead to incorrect research results. One of the most important information data in this regard is medical data, which even according to medical regulations, respecting people's privacy and keeping information confidential is essential. With the world entering the channel of the global epidemic of the corona virus, controlling the epidemic in the first place and finding its cure in the second place has become the challenge of the scientists and doctors of the world. In this regard, the computer science community has offered its role to control the corona virus epidemic to the world. Using deep learning to detect covid-19 from X-ray images becomes a fast method to diagnose patients and manage care services for patients. To achieve better results, we need a lot of data from different information sources, and data privacy as a barrier in this way will prevent engineers from achieving this important goal. Therefore, by introducing federated learning, which is a nascent leap towards creativity and better results, we will introduce its advantages and challenges and simulate it in the diagnosis of Covid-19 and try to take a small step to achieve more accurate results with more data. And of course, more organized to train neural network models, in the meantime, by presenting a strong aggregate approach, we were able to surpass our competitors with an accuracy of 97.04.      Keywords: Federated learning, covid19, x-ray, deep learning

  The

  امروزه با افزايش ارتباط بين فناوريهاي رايانهاي و حوزه پزشكي، واسطهاي مغز و كامپيوترتأثير مهمي در زمينههاي مختلف از جمله تشخيص فعاليت تصور حركت، بازشناسي احساسات، تشخيص بيماري صرع، امتياز بندي سطح خواب و باركاري ذهني دارند. تشخيص تصور حركت يكي از تكنيكهاي مبتني بر واسط مغز و كامپيوتر است. اين تكنولوژي با پردازش سيگنالهاي مغز و استخراج الگو از يكي از مهمترين سيگنالها در تشخيص اين نوع EEG تأثير به سزايي بر مطالعه ذهن وكاركردهاي آن دارد. سيگنال ،MI سيگنالهاي فعاليت است. اين پژوهش به طراحي، پياده سازي و ارزيابي يك روش جديد براي تشخيص تصور حزكت دست انسان مي پردازد. مغز انسان اين قابليت را دارد كه از طريق ارتباط بين نواحي و اثرگذاري بر يكديگر منجر به فعاليت هاي شناختي شود. به عبارت ديگر مغز از نواحي مختلفي تشكيل شده است كه هر كدام از آن ها يا به طور جداگانه يا تعاملي منجر به اجراي وظايف مختلف توسط انسان مي شود. اين روش جديد، از قابليت ذكر شده جهت ارزيابي عملكرد مغز و تشخيص تصور حركت دست انسان استفاده كرده است. در اين تحقيق تلاش شده است كه اطلاعات دقيق تر و كامل تري جهت ارزيابي مدل پيشنهادي، استفاده شود. به اين منظور براي تحليل سيگنال هاي حاصل از تصور حركت، از رويكرد مسئله معكوس به كار برده شد كه به اطلاعات آناتوميكي مغز حين تصور حركت دست EEG دسترسي دارد. در اين راستا از بازسازي منبع سه بعدي كه شامل مراحل مدل سازي فضاي منبع، مدل پيشرو و مسئله معكوس است، به كار برده شد. باتوجه به اطلاعات به دست آمده، از اتصال مؤثر (يكي از سه نوع اتصال بين نواحي مغز) مبتني بر مدل سازي علّي پويا استفاده گرديد كه گراف مربوط به نواحي مرتبط با تصور حركت طراحي و پياده سازي شود. نواحي مؤثر به كمك بازسازي منبع به دست آمده است. اين نوع مدل سازي بهتر مي تواند اتصالات جهت دار و علّي بين نواحي مغز و نقش مؤثر فعاليت نورون هاي قشر مغز را در ايجاد و اجراي تصور حركت تفسير كند. به دليل اينكه اطلاعات حاصل از گراف مدل سازي علّي پويا يك ماتريس مجاورت از مقادير اتصال مؤثر بين نواحي، ناشي از تعامل و اثر گذاري قشرهاي مغز است براي سهولت در استخراج ويژگي هاي سطح بالاي تصور حركت، از تكنيك شبكه عصبي كانولوشن گراف گونه جهت طبقه بندي نوع تصور حركت به كار برده شد. اين شبكه عصبي از طريق ماتريس مجاورت، جهت بين نواحي را تشخيص مي دهد و اطلاعات يال و رئوس گراف را براي استخراج ويژگي به كار مي برد. نتايج اين روش دقت بالاتري را در مقايسه با 5 و 10 لايه جهت تشخيص (GCN_ نشان داده است كه اجراي شبكه عصبي با 15 لايه ي كانولوشني ( 15 0% است. در مقايسه با / 0% و 99 / نوع تصور حركت را داشته است. دقت حاصل براي تصور حركت دست راست و چپ به ترتيب 95 پژوهش هاي پيشين نيز، روش پيشنهادي توانسته است دقت تشخيص را افزايش دهد.

Advances in Blockchain and distributed ledger technologies are driving the rise of incentivized social media platforms over Blockchains. Blockchain-based online social media is decentralized social media that uses blockchain technology to reward users' social activities and store information. In order to protect the privacy of users and expose fake news.    In this study, presents an empirical analysis of Steemit, a key representative of these emerging platforms, to understand and evaluate the actual level of decentralization in these modern social media platforms. Similar to Bitcoin, Steemit is operated by a decentralized community, where 21 members are periodically elected to cooperatively operate the platform through the Delegated Proof-of-Stake (DPoS) consensus protocol.    Our study performed on 539 million operations performed by 1.12 million Steemit users during the period 2016/03 to 2018/08 reveals that the actual level of decentralization in Steemit is far lower than the ideal level, indicating that the DPoS consensus protocol may not be a desirable approach for establishing a highly decentralized social media platform. For this reason, in this dissertation, we tried to provide a solution to the problem of decentralization of the consensus algorithm used in Steemit social media. Our solution to this problem is to replace its consensus algorithm with a more advanced consensus algorithm called the Algorand, which can form a committee without elections involving user interaction. Algorand is a new cryptography that proposes a new Byzantine agreement algorithm that allows choices to be made by randomly validated cryptographic functions rather than by users.    Using the simulator design as well as the API published by Algorand's team, we explored its three main aspects of decentralization, high scalability and security, and show that Algrand can be a good alternative to the DPOS algorithm. Be in Steemit.      

سرطان ريه شايع ترين سرطان در دنيا است. در اين پژوهش با استفاده از مدل يادگيري عميق توانستيم سرطان ريه را به دو دسته تومور و سالم طبقه بندي كنيم.

  Abstract Today, the security debate is considered an important and challenging issue. Older tools such as usernames and passwords alone are not responsive and reliable. That is why, day by day, in many areas, we need tools to identify individuals based on vital signs. With the advent of biometric knowledge, common methods of authentication in biometric systems have changed. Recently, the use of electrical brain signals (EEG) in biometric systems has been considered by researchers as an attractive and practical branch of research because it has two main advantages: First, this signal must be recorded from a living person in a normal mental state. Second, the EEG signal, unlike many other biometrics, is the result of a set of internal and cortical events in the brain that make it impossible to mimic. In this study, a data set with two different stimuli (relaxation and concentration) has been used that in the first period of time people are in a state of relaxation and in the second period of time people are in a state of concentration. An electrode is used to process and record EEG signals, then the analog signals are converted into digital signals. In this research, EEG data set with 109 topics has been used. In order to improve the performance of the authentication system in this study, instead of extracting features and selecting optimal features, deep features have been used. The results of our experiments on Albasri database with 99% accuracy indicate that using deep features and neural network algorithm Convolution using the genetic algorithm (GPCNN) is significantly improved over other electrical signal-based authentication systems of the brain, and shows a clear vision of the practical and commercial use of brain electrical signals in future authentication systems.

AbstractThe increasing number of private cars on the tra  ortation routes causes a heavy traffic load. In many countries, high occupancy vehicles (HOVs) have been developed to reduce the traffic load on special lines. Also, only buses, police vehicles, fire trucks, emergency vehicles, and personal vehicles with capacities to carry more than one passenger are allowed to use these lines. Another issue in monitoring the tra  ortation and traffic of vehicles is the observance of driving rules within the vehicle compartment. These rules include the drivers' use of seat belts while driving, and the accurate and automatic detection of these rules is of particular importance. In this paper, we propose a method based on deep learning models for simultaneous detection of the occupants and the status of driver's seat belt. In this method, first, the windshield is detected using the YOLOv5s network. Then, we determine the presence of a person in the passenger compartment using the front seat passenger detector model. Finally, using the deep learning-based image >Keywords: Car occupant detection, Seat belt status detection, Automated tra  ort images analysis, deep learning, transfer learning, YOLOv5, ResNet34, TPP,   , PMT   

   With advances in microelectromechanical systems (MEMS), researchers have now become interested in the systems operating based on inertial signals. In fact, inertial signals have proven useful in different areas due to advances in their manufacturing technology, availability, and inexpensiveness as well as the development of powerful processing methods such as deep learning techniques. Handwritten character recognition (HCR) is among such areas. This paper aimed to design, implement, and evaluate a novel system for the recognition of handwritten Farsi characters extracted from an inertial pen. For this purpose, a wireless inertial pen was designed. Its motion trajectory was then determined by combining the signals of its angular velocity and acceleration and using the concepts of navigation systems such as quaternion in order to estimate the position signals of characters. A convolutional neural network (CNN) was also employed to facilitate the extraction of high-level features and classification of characters. The position signal was also extracted as an image used for model learning to enhance the classifier efficiency. The experimental results indicated the CNN-6 architecture outperformed the other CNN-n architectures in terms of character classification accuracy. According to the evaluation of the proposed method through test data, character recognition accuracies of Farsi letters and numbers were reported 91.06% and 94.52%, respectively. In comparison with the previous systems, the proposed method managed to improve the recognition of handwritten Farsi characters.

In the last century, remote sensing imagery has been a major source of information in many applications, such as land cover detection, resource management, and monitoring. These images include a variety of aerial and satellite imagery, the use of which has expanded dramatically by placing various cameras and sensors on airplanes and other aircrafts. Detection and extraction of surface water is one of the main applications of remote sensing images that play a key role in controlling resources and preventing floods and crises such as drought. To date, various methods such as image thresholding, index detection, edge-based detection, and machine learning methods such as support vector machine have been used to improve the quality of water detection in images; However, the main application of these methods has been in problems where the water areas are not very dispersed and the water body has smoother boundaries and in images that include challenges such as the existence of water dispersed areas or narrow rivers, almost none - did not provide acceptable accuracy. Despite these issues, deep neural networks have obtained the state of the art results in the field of remote sensing image segmentation. In this research, a hybrid architecture called "stacked ensemble model" is presented to pixelwise >Based on the obtained results, the stacked ensemble method proposed in this thesis has succeeded in receiving the best result and also achieving the first rank among the participants of AIcrowd LNDST water body segmentation challenge which was held in August 2020.[1]    Key Words: Remote Sensing, Satellite Imagery, Water Body, Surface Water, Semantic Segmentation, Deep Learning [1] https://www.aicrowd.com/challenges/ai-for-good-ai-blitz-3/problems/lndst/leaderboards   

چكيده تشخيص سن استخوان روشي است كه به طور مكرر براي ارزيابي ناهنجاري رشد و تشخيص و درمان اختلالات غدد درون‌ريز و سندرم­هاي كودكان بيمار انجام مي‌شود. چندين دهه است كه تعيين سن استخواني با ارزيابي بصري از رشد اسكلت دست چپ انجام مي­شود و معمولاً از روش مرجع G&am   استفاده مي­شود. با پيدايش تصويربرداري ديجيتال، تلاش­هاي زيادي براي ايجاد روش­هاي پردازش تصوير انجام شده است كه به طور خودكار ويژگي­هاي اصلي مراحل تشكيل استخوان را براي ارزيابي مؤثر و دقيق­تر سن استخواني استخراج مي­كند. بااين‌حال ماهيت ذهني روش­هاي دستي، تعداد زياد مراكز استخوان در دست و تغييرات گسترده در مراحل استخوان‌سازي سبب پيچيدگي ارزيابي سن استخواني شده است و يك چالش براي طراحي الگوريتم­هاي كامپيوتري تشخيص خودكار در اين حوزه است. هدف: اين مطالعه با هدف ارائه يك روش جديد براي كاهش خطاي روش­هاي ذهني و بهبود روش­هاي اتوماتيك موجود در تخمين سن انجام شده است. روش: اين مدل روي 1400 تصوير از كودكان سالمِ صفر تا هجده سال از چهار قاره پياده‌سازي شده است. با استفاده از تكنيك­هاي پردازش تصوير در محيط برنامه‌نويسي متلب شش ناحيه در دست استخراج شدند؛ تجزيه‌وتحليل مراكز استخوان و محاسبه سن در هركدام از اين ناحيه­ها توسط تكنيك­هاي يادگيري عميق در محيط برنامه‌نويسي پايتون انجام شده است. دسته‌بندي نهايي نيز بر مبناي ميانگين رأي‌گيري صورت‌گرفته است. نتيجه: در مدل ارائه شده تمام سنين رشد و چهار نژاد آسيايي، آفريقايي، اروپايي و آمريكايي در نظر گرفته شده است. در قسمت پيش­پردازش تمام انگشت­هاي دست و مچ دست به‌درستي استخراج شده­اند. براي تشخيص نهايي سن از چند شبكه عصبي پيچشي و يك Ensemble بين آنها استفاده شده است. روش پيشنهادي به طور ميانگين 81 درصد دقت در تشخيص داشته است. اين دلايل نشان­دهنده برتري مدل پيشنهادي در مقايسه با ديگر مدل­هاي ارائه شده است. كلمات كليدي: اختلالات رشد، سن استخواني، روش Greulich and Pyle، روش Tanner-Whitehouse، مناطق اوليه رشد (ديافيز)، مناطق ثانويه رشد (اپيفيزها)، استخوان­هاي مچ (Carpal)، تصاوير ديجيتال (x-ray Image)، يادگيري عميق، شبكه­هاي عصبي پيچشي (CNN)، Ensemble، ميانگين رأي‌گيري (Average Voting).  

طبقه­بندي كاني­ها بخش جدايي­ناپذيري از زمين­شناسي است. به‌صورت سنتي براي مطالعه­ كاني­هاي موجود در مقاطع نازك، مرز بين كاني­ها به‌صورت دستي جداشده، هر ناحيه برچسب­گذاري و درصد هر كاني محاسبه مي­شود. اين روش نيازمند دانش، تخصص و تجربه­ بالايي است. از سوي ديگر خطاي انساني ناشي از خستگي و بي­دقتي موجب كاهش دقت طبقه­بندي مي­شود. بنابراين به­كارگيري يك سامانه مبتني بر پردازش تصوير براي تشخيص خودكاركاني­هاي موجود در سنگ­ها امري ضروري است. ارائه چنين سامانه­اي مي­تواند باعث افزايش دقت، كاهش خطاهاي انساني، كاهش هزينه و كاهش زمان جهت تشخيص نوع كاني­ها مي­شود؛ بنابراين، هدف اين پژوهش، پيشنهاد يك سامانه تشخيص خودكار كاني است كه با استفاده از پردازش تصوير، كاني­هاي موجود در سنگ را شناسايي و طبقه­بندي كند. مرحله اول در انجام اين پژوهش ايجاد يك پايگاه داده از تصاوير مقاطع نازك سنگ است. اين مرحله يكي از چالش­برانگيزترين مراحل اين پژوهش بود، زيرا ايجاد يك پايگاه داده مناسب از تصاوير مقاطع نازك، فرآيندي سخت و وقت­گير است. از سوي ديگر، پايگاه داده مشترك و استانداردي در اين حوزه وجود ندارد و هر پژوهشي از پايگاه داده متفاوتي استفاده مي­كند. پس از ايجاد پايگاه داده و برچسب­گذاري تصاوير مقاطع نازك، چند روش­ قطعه­بندي بررسي و الگوريتم JSEG براي قطعه­بندي انتخاب شده است. پس از انجام قطعه­بندي، ويژگي­هاي مبتني بر رنگ و بافت از هر ناحيه استخراج شده­اند. ويژگي­هاي رنگي از هر دو فضاي رنگي RGB و HSI استخراج شده­اند. هم­چنين به دليل اينكه برخي كاني­هاي متفاوت داراي رنگ­هاي مشابه هستند، ويژگي­هاي بافت نيز از هر ناحيه استخراج شده­اند. ويژگي­هاي استخراج‌شده از هر ناحيه، براي طبقه­بندي به طبقه­بند فرستاده شده و طبقه­بند هر ناحيه را به‌عنوان يك كاني برچسب­گذاري كرده است. در اين پژوهش كارايي شش طبقه­بند Linear Discriminant، Su  ace Discriminant، Boosted Tree، Bagged Tree، Linear SVM و Weighted KNN بر اساس معيارهاي مختلف مورد ارزيابي قرار گرفته است. بر اساس نتايج تجربي به‌دست‌آمده، طبقه­بند Bagged Tree داراي بالاترين دقت به ميزان 5253/95 و همچنين كمترين ميزان خطاي MAE برابر با 0447/0 و خطاي RMSE برابر با 2115/0 مي­باشد. همچنين همه طبقه­بندها داراي دقت قابل قبول بالاي 93% هستند. اين نتايج نشان مي­دهد كه روش پيشنهادي داراي قابليت مناسبي جهت شناسايي خودكار كاني­هاست.  

   With advances in Internet technology and easy access to Internet networks, digital images that are available to the public can be modified, manipulated, and copied illegally without any qualitative loss. Without respecting the ownership of this data, the desired content can be reproduced and distributed on a large scale. Digital watermarking appears as a branch of data concealment science and protects the ownership and copyright of these images. In this dissertation, the idea of designing and implementing an invisible digital image watermarking method in the field of Lifting Wavelet Transform and Singular Value Decomposition in order to improve image Robustness and increase the reliability of image watermarking system using The Three Module Redundancy Technique is proposed. First, the Lifting Wavelet Transform is applied to the host image, and then its three high frequency su  ands divided into 8×8 non overlapping blocks. In each frequency subbands, the Lifting Wavelet Transform is converted to each block, and four frequency subbands ll, lh, hl and hh are obtained for each block. Then the lh frequency subband is selected from each block and the singular value Decomposition is applied to the selected frequency subband. In order to increase the security of the hidden image against alteration and manipulation, the hidden image is first encrypted using the improved Arnold Transform. Following the technique of the Three Module Redundancy, the Singular Values of the watermark image are inserted into the host image by adding the Singular Values of each frequency subbands. The three optimal scale factors used by the ant bee colony algorithm to determine the balance between Robustness and imperceptability are used to insert the watermark image. Evaluations show that the average hidden Robustness of different images against signal processing and geometric attacks is more than 96% and the average imperceptability of the hidden images is more than 50 dB. According to the assessments, the proposed method is more Robustness and imperceptability than the existing works, and the proposed system is a fault tolerance system that can work properly under different circumstances. Also, by using the improved Arnold conversion to encrypt the watermark image, attempt has been made to increase the image security.

Abstract You've definitely seen a variety of movies and animations in the cinema that make for spectacular special effects. These special effects are quite similar to the real world, and the movements of the characters are similar to those in the real world. With the advent of 3D motion capture technology and its move to computers, movies, computer games and especially animations have entered a new world. When movies started using 3D models, the goal was for them to create real motion and speed up the workflow so that the motion was not manually animated. The solution is to capture the movements of an actor in 3D and apply them to 3D computer models. The purpose of this technology is to allow us to create more effective and realistic characters and effects that we were not able to do before. The purpose of this dissertation is to design and implement an actor's face recording system using digital image processing techniques and machine learning algorithms, which is performed without the use of a specific hardware system. In this design, the camera is first captured by a computer webcam and then detected the face in the image and then the key points of the face are detected, then the two-dimensional points are identified. The camera parameters and two-dimensional mapping algorithms and their combination with facial feature points are mapped to the 3D coordinate space and a three-dimensional model of the face is created. This three-dimensional model is independent of head rotation and a particular face. Finally, the data obtained from the previous steps are transferred to a 3D virtual character in Maya 3D software by connecting to a TCP / IP socket.  

  رمبندگي خاك به ريزش و كاهش حجم ناگهاني خاك در موقع اشباع شدن آن اطلاق مي­شود. خاك­هاي رمبنده مناطق وسيعي معادل 15 درصد توده‌هاي قاره‌اي جهان را شامل شده است. خاك­هاي رمبنده در مناطق خشك و نيمه­خشك جغرافيايي از جمله ايران مشكلات بيشتري ايجاد مي­كنند. پديد? رمبندگي زماني اهميت مي­يابد كه نفوذ آب­هاي سطحي، نشت آب از لوله­ها، بالا آمدن سطح آب زير­  hy;زميني، باعث اشباع شدن خاك رمبنده شود و تغيير حجم ناگهاني خاك براي ابنيه احداثي مجاور آن خطر آفرين باشد. آلاينده­  hy  hy  hy;هاي زيست محيطي فراواني مانندNa2SO4   و   H2 SO4   وجود دارند كه مي­توانند با نفوذ در خاك، به مرور باعث تغييرات عمد? خصوصيات فيزيكي و شيميايي خاك­ از جمله شاخص رمبندگي آن   شوند.بر اساس استاندارد ASTM D5333-03 شاخص رمبندگي ، شاخصي است كه به­ منظور تعيين بزرگي رمبندگي خاك حين اشباع شدن نمونه در تنش سربار 200 كيلوپاسكال در دستگاه ادومتر ارزيابي مي­شود. هدف از انجام اين پژوهش، بررسي تأثير جابه­جايي ترتيب لايه­هاي رسي و ماسه­اي بر تغييرات شاخص رمبندگي خاك رمبنده طبيعي در حضور نمك سولفات سديم و سولفوريك اسيد است.در اين پژوهش خاك ماسه‌اي و رسي از محل فروچاله عظيم روستاي كِردآباد همدان تهيه شد. پس از تفكيك ريزدانه و درشت دانه خاك‌ها، نمونه‌ها در سه حالت تك لايه ماسه‌اي، دو لايه و سه لايه با چينش‌هاي مختلف لايه‌اي با وزن‌هاي مخصوص 3/1 و 5/1 و 7/1 گرم بر سانتي‌مترمربع درون رينگ‌هاي پنج سانتي‌متري دستگاه ادومتري تهيه شد و طبق استاندارد ASTMD5333-03 مورد ارزيابي قرار گرفت. سپس به مقدار 4 و 8   درصد وزني به نمونه‌ها نمك سولفات سديم اضافه شد و مجدداً طبق روال قبل، شاخص رمبندگي نمونه‌ها اندازه‌گيري شد. براي اشباع سازي نمونه‌ها در تنش 200 كيلو پاسكال هم از آب خالص و هم از اسيد سولفوريك رقيق با غلظت يك مولار استفاده شد.با انجام تعداد قابل توجهي آزمايش تعيين شاخص رمبندگي مشخص شد وجود نمك سولفات سديم در نمون? تك لايه خاك ماسه طبيعي، شاخص رمبندگي را به شدت افزايش مي‌دهد. به طوري‌كه بيشترين شاخص رمبندگي در بين تمام نمونه‌هاي اين پژوهش، در خاك ماسه‌اي تك لايه حاوي 8 درصد وزني سولفات سديم با شاخص رمبندگي 80/24 درصد اتفاق افتاده است. همچنين وجود يك لاي? رسي حاوي آهك به علاوه 4 درصد تا 8 درصد وزني نمك سولفات سديم بين لاي? ماسه حاوي سولفات سديم، شاخص رمبندگي را   به مقدار قابل توجهي كاهش مي‌دهد.در مورد   خاك‌هاي چند لايه شامل ماسه و رس حاوي آهك، در حالتي‌كه در نمون? دو لايه، رس بالاي ماسه قرار گيرد (چينش C-S) ودر نمونه سه لايه در حالتي‌كه ماسه مايبن دو لاي? رس قرار گيرد (چينش C-S-C)، جود نمك سولفات سديم باعث كاهش شاخص رمبندگي مي‌شود. اما در حالت چينش سه لايه در حالتي‌كه رس بين دو لايه ماسه قرار مي‌گيرد (چينش (S-C-S، وجود نمك سولفات سديم در نمونه‌ها منجر به افزايش شاخص رمبندگي مي‌شود.كلمات كليدي: شاخص رمبندگي، سيستم چند لايه خاك، وزن مخصوص خشك خاك، نمك سولفات سديم، سولفوريك اسيد، آهك

an independent unified section.  

This thesis aims at extracting the influential components for identifying architectural spaces, in particular, residential architecture and subsequently designing a residential complex in Nurabad, Lorestan. The lack of identity in contemporary architectural spaces and residential environments, and the resulting semantic vacuum, have been the main motivation for this research to come to an identity-based approach to residential architecture and architecture; the main question that is being explored during the literature review; Identity maker in architecture. Hence, using a descriptive-analytical strategy, gathering various internal and external sources in the field of architecture and architectural identity, drawing on the views of experts and scholars on architectural identity, and using content analysis of these ideas and opinions, It has extracted and deduced the identity component in architecture. These components of identity are: "adherence to culture", meaning that architecture is rooted in the set of values, beliefs, norms, and cultural expectations of society. "Adapting to human needs", meeting the material and spiritual needs of residents of all ages, genders and conditions. "Adaptation to materials" means the use of native materials consistent with the environment of the architectural establishment. "Flexibility" means the ability of the environment to personalize it as well as flexibility against time changes. "Adaptation to climate" to design architecture based on environmental identities and climate forces and consider the necessary arrangements for coexistence with environmental conditions. "Continuity of Architecture" to maintain and maintain timeless principles, values ??and patterns and associative elements. "Innovation" means moving towards the dynamic, technology and continuous evolution of achievements and values. 'Adaptation to performance' means the coordination of the architectural form with the expected performance as well as the current human activity in it. "Adaptation to time" means birth and bio-architecture with all its elements in its time and day, and harmony with the spirit of time. "Adaptation to the landscape and nature" means architectural attention to the environment and the natural context of its construction, from terrain and features to landscapes, landscapes, and the surrounding nature of architecture.In the course of his research, he has adapted the principles of Iranian architecture that have emerged from the viewpoints of the experts with the extracted components of identity, and has shown that these components are in conformity with the principles of Iranian architecture, on the one hand, and on the other, the identity of the past architecture. It has deduced the validity of the extracted components, and also provided a model of the relevance of the extracted components of research in relation to the distinct dimensions of location (body, meaning, and activity) to better understand the broad dimensions of place identity. At the end, these ten extracted components have been used as the main conceptual basis for designing the aforementioned residential complex and have been the criteria for designing the project.

   Due to the complexity of medical decisions, the application ofinformation systems to support these decisions has increased. The presence oflarge and unknown variables means more complexity of decision-making. Given thevariables’ frequency and interference in the medical field, physicians cantime on the assessment of the decision more. In order to design medical expertdecide faster and more efficiently through using expert systems and spend theiror clinical guidelines and entered into the knowledge base. I.e. the knowledgesystems, specialized knowledge is extracted from the experts in the given areaand experience of the professionals in different fields can be entered into theto recommendations at any time and place will increase with these systems. Thisdecisions of different people, and eventually, the speed of analysis and accessis very important for medical decisions. According to the above, it isconcepts, the use of fuzzy logic in this area is effective; since fuzzy systemsrecognized that we are faced with serious problems in the process of medicaldiagnosis and performance of physicians that necessitates a collective wisdomFurthermore, due to the inherent ambiguity in the definitions of medicalto improve the quality of treatment with the help of expert systems.can play a valuable role in the diagnosis of diseases [24, 31, 32, and 33].diagnose the individual’s disease, if any, through using the fuzzy system andHence, in this study, through using fuzzy system, we are aimed at creating aset of fuzzy rules in order to achieve a higher-level series of information byanalyzing the initial data in blood test results such as red blood cells,tests.hematocrit, white blood cells, hemoglobin, platelets, etc. Our goal is tomedical rules as well as the high-level information obtained from the blood

Human Identification Based on Ear Biometric Employing a Hybrid Approach  

Recognizing the emotional states using matching point  

هدف اين تحقيق، ايجاد يك سيستم كامپيوتري مؤثر سيستماتيك آموزشي براي طراحي اشكال ساده براي افراد نابينا است (پروتوتايپ)، بنابراين ما يك سيستم آموزشي براي كودك عرب ايجاد كرديم كه در همان زمان، قرآن كريم را نجات داد. طراحي ELMS بايد به حداكثر رساندن نتايج آموزشي براي كانديد نامزد ما در اينجا، روش برنامه نويسي پويا با توجه به مجموعه اي از حروف شناخته شده است. صداي نزديك به كلمه واژگاني با مقايسه كلمه كليدي با تمام الگوها در كتابخانه گوگل و انتخاب آن كه داراي حداقل فاصله (مشابه) با بستر مطلق است و سرعت پاسخ به سرعت بستگي دارد از اينترنت يك حالت كد در صدا و كد ديگري در متن است. تبديل كد در نرم افزار python انجام شده است كه با كد API گوگل كار مي كند. تعويض بين (Speech to command) (متن به گفتار) چندين مرحله دارد. اما در دسته كلي، دو مدل اصلي كه عبارتند از: 1-متن به گفتار 2- گفتار به فرمان. اولين گام ورود به متن عربي به كامپيوتر و شناسايي متن و تبديل متن به فايل صوتي است. بيشتر خطا در مرحله دوم به دليل اين ماژول بسته به دستگاه ورودي (ميكروفون)، سرعت اينترنت و سر و صدا در اطراف فرد و كيفيت صدا رخ مي دهد، تمام دستورات در سيستم عمل مي كنند (بازي سوره قرآن) و غيره تا زماني كه تمام سوره هايي كه در LMS ما ذخيره مي شوند، اين 10 قرآن سورات كوتاه (سوره القطار، سوره فلاع، سوره النس، سوره الطوف، سوره الفيل، سوره النشره، سوره آل -Asr، سوره القرية، القادر و سوره الاخلا) پس از آن درصد خطا در ماژول STC در بيست افراد عرب 10? براي سوره القرية، 5? سوره الاراي، 40? براي سوره النشره، 5? براي سورات الفيل، 0? براي سوره آل نوجوان، 0? براي سوره النس، 0? براي القاد، 15? براي سوره القوث، 25? براي سوره الاخلا و 5? براي سوره الفلق، ما الگوريتم اين ELMS را ساختيم، اما مشكل بزرگي كه من با آن مواجه شدم، معرفي زبان عربي به برنامه بود. دليل اين امر اين است كه زبان عربي زود هنگام در جهان vo تشخيص يخ در مقايسه با زبان انگليسي.  

  در اين پژوهش، فرآيند درج واترمارك شامل اعمال روش بهينه سازي هوشمند DE بر روي تصاوير ميزبان و واترمارك براي يافتن مكان مناسب هر بلوك از تصوير واترمارك در تصوير ميزبان است. سپس جهت بازيابي موفق،‌ خروجي برنامه بهينه سازي در تصوير ميزبان تحت حوزه فركانسي جاسازي مي­شود.   همچنين ضرايب مورد استفاده در جاسازي تصاوير به شكل بهينه بدست آمده است تا بيشترين مقدار   R را بدست دهد. در اين روش، يك بهينه سازي چند هدفه با استفاده از الگوريتم تفاضلي انجام شده است كه در آن مقدار   R در مرحله جاسازي براي تصوير واترمارك و در مرحله استخراج براي تصوير واترمارك بازيابي شده، بسيار مناسب است. در فرآيند درج و استخراج واترمارك، تعبيه و آشكارسازي واترمارك مهمترين بخش مي­باشند چرا كه مقاوم بودن طرح واترماركينگ به بخش تعبيه واترمارك مربوط مي­باشد. سپس مقاوم بودن طرح واترماركينگ در بخش نتايج تجربي مورد ارزيابي قرار مي گيرد و در بخش نتايج تجربي تصوير واترمارك شده را تحت حملاتي از قبيل فشرده سازي تصوير، نويز گوسي و غيره مورد آزمايش قرار داده   و صحت درستي وجود واترمارك مورد ارزيابي قرار خواهد گرفت.

  Proper user interface design is one of the most important issues in software production. The user understands only the interface to which it relates, and recognizes it as software. Therefore, user interfaces play an important role in the acceptance of the software. Developing a software without designing a suitable user interface will result in its disapproval. The acceptance of a UI is not limited to technical factors because of the user interface is used by the human, and human decisions are depended on psychological factors. As a result, one of the issues that involved in accepting or rejecting the user interface is the human-psychological factor. On the other hand, designing an appropriate user interface often requires the right feedback from the evaluators of that user interface and then applying the correct changes to the final software product.by investigating the majority of papers and computer-implemented studies, and especially the design of the user interface, In this field, we find that in most of these papers, the evaluators are only selected based on their expertise and academical degrees to evaluate the user interface of a system, and the importance of psychological factors and personality traits on the quality of examining different aspects of a system by evaluators is not significant. Therefore, it can be presented that in evaluating the user interface for a software product, the condition of expertise is not sufficient. Depending on the psychological nature of the study, Evaluators should have the reasonable level of emotional intelligence.In this research, we investigate emotional intelligence as one of the most influential factors of human-computer interaction. According to the effect of emotional intelligence on evaluating the user interface of a software by the human, we will provide a method for selecting proper persons to evaluate the user interface.Therefore, we used the Bar-On's emotional intelligence questionnaire to measure the emotional intelligence of the subjects. The Nielsen's criteria questionnaire is used for evaluating the effectiveness of the user interface, of the Shagerdaneh learning management system. Investigating psychological factors (emotional intelligence) and proving their impacts on the quality of analyzing the user interface and finally filtering the evaluators who have a normal rating in these emotional intelligence features are part of the objectives of this study.The statistical population consists of 35 software professionals. The system used for this research is the user interface of the Shagerdaneh learning management system, which was previously designed by the author of this research. The required information will be collected from the questionnaire. For determining the feasibility of the project, normalization and correlation of data will be analyzed by    software. In order to investigate the effect of evaluators' emotional intelligence on how to evaluate the user interface and providing a prediction model for input data, we use multiple regression methods in genetic programming using the GPTIPS version 2 toolkit in MATLAB software version 2017. We also use clustering methods to evaluate the work's accuracy. All methods and tools are described in Chapter three.The results of this research prove the effect of emotional intelligence on the way that evaluators investigate the user interface. After filtering, seven persons were selected from 35 samples as the appropriate evaluators. In order to verify the accuracy of the proposed method in this study, the scores of the seven selected users were compared with the expert views in this field, and more than 71% of them was close to expert Opinions.Increasing the quality of the web content by examining the emotional intelligence of content providers and teaching emotional intelligence to them, is a suggestion for future studies.

  The database used in this study is a 10000 sample medical database from a cohort study conducted in ravansar city, IRAN. This database consist variables like biochemical tests, CBC tests, anthropometric data and lifestyle variables. The main reason of creating this database was to have a baseline source for storing ravansar city residents’ information to follow up in a 15 year prospective cohort study. Therefore in this study the effort was to extract interesting relationships between these variables and chronic diseases using genetic programming. In contrast to majority of data mining researches that take performance and creating a novel model as the main purpose, this study has chosen knowledge extraction. At first by using the feature selection ability of genetic programming we have tried to reveal the most related variable for each disease. Rule mining is done by association rule mining algorithm. An evaluation process of the GP results is conducted by the extracted rules to check the correctness of GP’s results. Features selection and rule generation are done separately for biochemical test and CBC tests based on their different definition. From the diseases we have chosen diabetes and hypertension as the goal diseases because of the quality of the data they have. For diabetes ALP, TR, GGT, BU are considered as the most related variables while for hypertension FBS, ALP, LDL, TR are the most important ones. Results considering CBC variables were not mentionable.

بررسي آزمايشگاهي بتن ساخته شده از مصالح بازيافتي پلي اتيلن سنگين به عنوان سنگدانه

چكيدهبا توجه به افزايش ترافيك و نياز به پاسخ گويي سريع به درخواست­ها دسته­بندي بسته­ها به يك تكنولوژي مهم و يك چالش در عملكرد مسيرياب­ها تبديل شده است، بخصوص در زمان همگام سازي تصميم گيري خود با سرعت تبادل داده­ها اين موضوع بيشتر نمود پيدا مي كند، يعني سرعت جست و جوي فيلدها با سرعت لينك­هاي انتقال برابر باشد و تا زماني كه سرعت شبكه­ها ثابت نشود كار روي دسته­بندي بسته­ها اهميت خود را حفظ مي­كند. افزايش روزافزون داده­هاي انتقالي و پويا بودن آنها باعث شده راه حل­ها و معماري­هاي سخت­افزاري يا نرم­افزاري متعددي براي اين موضوع ارائه شود. الگوريتم­هاي نرم­افزاري با وجود توسعه­پذيري بالايي كه فراهم مي­كنند اما از سرعت پائيني برخوردارند از طرف ديگر راه­حل­هاي سخت­افزاري سرعت خوبي دارند ولي هزينه بالا و قابليت توسعه­پذيري كمي دارند. از اين رو ارائه روشي براي ايجاد مصالحه بين سخت­افزار و نرم­افزار مورد توجه محققان قرار گرفته است. طبقه بندي بسته­ها يك جستجوي چند فيلدي با سرعت لينك ا انجام مي­دهد.در اين تحقيق ، به منظور رفع مشكلاتي كه در بالا ذكر شد از دو فيلتر بلوم و خارج قسمت استفاده شد و به منظور انطباق روش جستجو با بسته هاي ارسالي در تعداد فيلدهاي موجود در معماري نوين SDN، اين تعداد به 15 فيلد سرايند افزايش يافت. در نهايت با استفاده از ابزارهاي در دسترس از جمله   Intel Platform Power Estimation Tool (IPPET)   معيارهاي مورد نظر براي بررسي قابليت هاي روش ارائه شده استفاده گرديد و از نتايج حاصل از دو فيلتر برتري فيلتر بلوم نسبت به فيلتر خارج قسمت دربرخي معيارها اثبات گرديد به اين صورت كه در مورد زمان مصرفي، سرعت انجام الگوريتم، توان عملياتي و انرژي مصرفي فيلتر بلوم عملكرد بهتري داشته ولي در موارد حافظه مصرفي و نرخ خطاي مثبت فيلتر خارج قسمت عملكرد بهتري دارد.

پيش بيني سيل با استفاده از تلفيق تصاوير ماهواره اي و مدل بارش-رواناب در مناطق فاقد آمار

  Energy storage systems (E  ) play a major role in power systems planning and operation. Regarding the continuing evolution of the storage technologies, the E   will be highly regarded in the future power networks, and also their various applications will be increased. The optimal E   allocation problems mean reducing the investment costs (or initial costs) as well as reducing the networks expected operation cost. By increasing the ESS capacity, the investment costs are increased, but the network operation cost is reduced; therefore, selecting inappropriate size and site for the E   would result in undesirable costs in the system. Among several advantages of the E  , improvement of the power system costs and voltage profile can be considered as the most prominent characteristics of the E  . Besides, increasing the renewable energy resources penetration, such as wind power, for reduction of the environmental pollution, postponed the construction of large and concentrated fossil-fuel power plants as well as transmission lines, increased complexity and harder distribution and its subsequent significant costs, improved power quality and increased reliability, and supply of the subscribers power demand at load peak times are some of other advantages of the use of E  . By using the E  , the generation power shortage resulted from disconnection of the existing units or separation of the renewable resources can be controlled; thus, the microgrids reliability criterion, especially in cases of high penetration of the renewable resources, is met. Simultaneous determination of size and site of the E   is a non-deterministic non-convex problem, which should be modeled in presence of the real constraints governing the power system. On this basis, in the present dissertation, for the first time, the optimal allocation problem of E   was investigated considering the practical constraints including generation and consumption balance, units generation power limitation, prohibited performance zones, ram   rate, as well as simultaneous reduction of three different and incompatible objective functions of cost, voltage deviation, and air Emission. Due to complexity of the problem, the presence of various constraints and incompatible objective functions, as well as possibility of involvement of the classical optimization methods in local points, from among the evolutionary optimization methods, two hybrid multi-objective algorithms known as MOGSA and MOPSO-NSGA-II were proposed. The modeling and formulation of the problem was performed in MATLAB software. In order to take the uncertainty into account, wind generation power was discretized using the five-point estimation method (5PEM), and the IEEE 30-bus standard system was selected for simulation. Furthermore, the multi-criteria decision-making techniques were used in order to increase accuracy and make sure of selection of the best solution from among the optimal solutions. The simulation results clearly show efficiency and effectiveness of the proposed method.

  The issue of   unit commitment is to determine the situation of switching on or off the available power generation units in a power grid. By solving this problem, the quantities of productive powers are determined economically, according to the existing constraints. One of the important criteria in solving the problem of orbiting unit commitment is to minimize the cost of electric power supply for the power system over a given time period. In recent years, the use of renewable energy for electrical energy has been given particular attention due to rising fossil fuel costs and also environmental problems caused by these fuels. On the other hand, , there are also some problems of the participation of these units in supplying demand for the network with the increased penetration of renewable energy sources in power systems. One of these problems is the uncertainty in the electrical energy produced by these units, and also the unavailability of these units in the hours of the day. The uncertain nature of the energy produced by renewable energy sources has led to a difference in the actual and projected output power of renewable unit commitment. Therefore, the combination of active units in planning in the circuit of the unit commitment in real time may be subject to changes that will remove the system from the performance at the optimal point. To overcome this problem, it is necessary to determine how the renewable energy sources participate in supplying the demanded load of network through the planning of unit commitment. Therefore, by reducing the uncertainty of the capacity of the production of renewable energy sources, the system can be brought closer to the optimal work point. Therefore, in order to compensate for the nature of periodic and incidence of renewable energy sources, electrical energy storage devices have been used in power systems. In fact, the purpose of using energy storage systems in power systems is to maximize the use of renewable energy sources. In this paper, the problem of the installation of unit commitment has been investigated, due to the many problems associated with the use of fossil fuels, with regard to the cost of pollution in the presence of renewable energy sources and energy storages. The MATLAB software, the Combination of the Genetic Algorithm and the Priority List Procedure with new use have been used to solve the problem in unit commitment. In this method, the initial population is produced in such a way that the supply and demand equivalence requirements and the minimum time down, the minimum time up of the units the constraints of the ramp rate, and the reduction of electrical power of units and the reserve of the spinining reserve erformed on the standard 10 units and 38 units. The results indicate that this process can help reduce the cost of the problem in the circuit of the unit commitment.

<  gt;بهبود پايداري ريز شبكه هاي شامل توربوژنراتورهاي بادي با استفاده از اينرسي مجازي وكنترل مقاوم</P>

هدف از شناسايي گوينده ، تمايز قائل شدن بين افراد از طريق تفاوت در ويژگيهاي گفتار آنهاست. به اين معني افراد نه تنها در ويژگيهايي مانند اثر انگشت و برخي ويژگيهاي شناخته شده از هم قابل تفكيك هستند، بلكه مي­توان از تفاوتهاي ديگري مانند، شكل دستگاه صوتي و ويژگيهايي مثل لحن، لهجه، طرز بيان و ... نيز بهره برد. روشهاي زيادي براي مدل كردن سيگنال صوتي، بصورتي قابل تحليل بوجود آمده­اند. از جمله­ي اين روشها مي­توان به روش مدل مخلوط گوسي و مدل پس­زمينه جهاني استفاده كرد. از اين مدل براي تشكيل ابربردارهاي گوسي استفاده شده است. ابربردارهاي گوسي بردارهايي با بعد ثابت هستند كه از سال 2006، توسط كمپبل تعريف شده­اند. و در سيستمهاي شناسايي گوينده مورد استفاده قرار گرفته­اند. مشكل اين ابربردارها، بعد بالاي آنهاست كه موجب افزايش پيچيدگي محاسباتي شده است. براي مقابله با اين مشكل، از روشهاي كاهش بعد مانند بدست آوردن بردار i-vector مربوط به هرگوينده استفاده شده است. در اين تحقيق مؤلفه­هاي گوسي كه براي مدل كردن i-vectorها استفاده شده اند با توجه به مقدار آماره باوم ولچ مرتبه صفر آنها به دو دسته مؤلفه­هاي كم اهميت و مؤلفه­هاي مؤثر دسته­بندي شده­اند. از هركدام از اين مجموعه­ها عناصري بصورت تصادفي حذف مي­گردد كه تعداد اين عناصر حذفي در دو مجموعه متفاوت است. براي ارزيابي عملكرد سيستم از پايگاه داده TIMIT استفاده شده است. ميانگين خطاي EER روش پيشنهادي نسبت به كمترين مقدار خطاي EER در ساير روشها 56درصد كاهش داشته است.كلمات كليدي: ابربردار، i-vector، نمايش تنك، ماتريس نگاشت، شناسايي گوينده، مدل مخلوط گوسي، مدل پس زمينه جهاني  

Abstract- The need to private and sensitive information security on multi-touch devices like smartphones and tablets is one of the main problems in information security. Methods that are commonly used passwords and tokens that have a lot of obstacles and challenges. Biometric authentication methods, these methods are a good alternative to overcome the problems. The introduction of biometric based smartphone touchscreen for user authentication is based on finger touch and movement. The purpose of this Study is to examine method of authentication using biometric behavior based on specific gesture for unlocking the device based in existing designs is safe. In this study, by extracting a large number of features and using Distance learning with Genetic Programming, With high accuracy in authentication based on finger multi-touch touch screen to unlock the device achieved.

In recent years iris recognition has attracted the attention of many researchers and it also has been used in many real-world applications. In iris recognition, segmentation phase always has been the one of the challenging problems and it always consumes significant processing time. On the other hand features in a classification task play a major role and as the selected features are good the performance of the classifier can be improved. Particle swarm optimization algorithm is an evolutionary algorithm and it successfully has been used in many optimization problems. We have used this algorithm to select the most appropriate features in an iris recognition task and in this way we have learned a near optimum distance metric. In addition, in this study we have provided an effective and simple method to detect the iris area that could greatly improve iris area detection process speed. To evaluate the proposed method two data sets CASIA Interval and IITD have been tested and the results have been very promising.