Search

Fake news detection is a challenging problem in online social media, with considerable social and political impacts. Several methods have already been proposed for the automatic detection of fake news, which are often based on the statistical features of the content or context of news. In this paper, we propose a novel fake news detection method based on Natural Language Inference (NLI) approach. Instead of using only statistical features of the content or context of the news, the proposed method exploits a human-like approach, which is based on inferring veracity using a set of reliable news. In this method, the related and similar news published in reputable news sources are used as auxiliary knowledge to infer the veracity of a given news item. We also collect and publish the first inference-based fake news detection dataset, called FNID, in two formats: the two-class version (FNID-FakeNewsNet) and the six-class version (FNID-LIAR). We use the NLI approach to boost several classical and deep machine learning models including Decision Tree, Naïve Bayes, Random Forest, Logistic Regression, k-Nearest Neighbors, Support Vector Machine, BiGRU, and BiLSTM along with different word embedding methods including Word2vec, GloVe, fastText, and BERT. The experiments show that the proposed method achieves 85.58% and 41.31% accuracies in the FNID-FakeNewsNet and FNID-LIAR datasets, respectively, which are 10.44% and 13.19% respective absolute improvements.

Machine Reading Comprehension (MRC) is a challenging task and hot topic in Natural Language Processing. The goal of this field is to develop systems for answering the questions regarding a given context. In this paper, we present a comprehensive survey on diverse aspects of MRC systems, including their approaches, structures, input/outputs, and research novelties. We illustrate the recent trends in this field based on a review of 241 papers published during 2016–2020. Our investigation demonstrated that the focus of research has changed in recent years from answer extraction to answer generation, from single- to multi-document reading comprehension, and from learning from scratch to using pre-trained word vectors. Moreover, we discuss the popular datasets and the evaluation metrics in this field. The paper ends with an investigation of the most-cited papers and their contributions.

This paper provides a review of the consensus problem as one of the most challenging issues in the distributed control of the multi-agent systems (MASs). In this survey, firstly, the consensus algorithms for the agents with the single-integrator, double-integrator and high-order dynamic models were collected from various research works, and the convergence condition for each of these algorithms was explained. Secondly, all the consensus-related problems such as those in the sampled-data consensus, quantized consensus, random-network consensus, leader–follower consensus, finite-time consensus, bipartite consensus, group consensus/cluster consensus, and the scaled consensus were analyzed and compared with each other. Thirdly, we focused on the common control techniques used for the consensus problems in the presence of disturbance and divided all these control methods into two categories: robust control and adaptive control. Finally, we reviewed the most prevalent consensus applications in the MASs, including the subjects of rendezvous, formation control, axial alignment and the wireless sensor networks.

Machine Reading Comprehension (MRC) is an active field in natural language processing with many successful developed models in recent years. Despite their high in-distribution accuracy, these models suffer from two issues: high training cost and low out-of-distribution accuracy. Even though some approaches have been presented to tackle the generalization problem, they have high, intolerable training costs. In this paper, we investigate the effect of ensemble learning as a light approach to improve out-of-distribution generalization of MRC systems by aggregating the outputs of some pre-trained base models without retraining a big model. After separately training the base models with different structures on different datasets, they are ensembled using weighting and stacking approaches in probabilistic and non-probabilistic settings. Three configurations are investigated including heterogeneous, homogeneous, and hybrid on eight datasets and six state-of-the-art models. We identify the important factors in the effectiveness of ensemble methods. Also, we compare the robustness of ensemble and fine-tuned models against data distribution shifts. The experimental results show the effectiveness and robustness of the ensemble approach in improving the out-of-distribution accuracy of MRC systems, especially when the base models are similar in accuracies.

The detection of traffic signs in clean and noise-free images has been investigated by numerous researchers; however, very few of these works have focused on noisy environments. While in the real world, for different reasons (e.g. the speed and acceleration of a vehicle and the roughness around it), the input images of the convolutional neural networks (CNNs) could be extremely noisy. Contrary to other research works, in this paper, we investigate the robustness of the deep learning models against the synthetically modeled noises in the detection of small objects. To this end, the state-of-the-art architectures of Faster-RCNN Resnet101, R-FCN Resnet101, and Faster-RCNN Inception Resnet V2 are trained by means of the Tsinghua-Tencent 100K database, and the performances of the trained models on noisy data are evaluated. After verifying the robustness of these models, different training scenarios (1 – Modeling various climatic conditions, 2 – Style randomization, and 3 – Augmix augmentation) are used to enhance the model robustness. The findings indicate that these scenarios result in up to 13.09%, 12%, and 13.61% gains in the mentioned three networks by means of the mPC metric. They also result in 11.74%, 8.89%, and 7.27% gains in the rPC metric, demonstrating that improvement in robustness does not lead to performance drop on the clean data.

Despite their many advantages and positive features, the deep neural networks are extremely vulnerable against adversarial attacks. This drawback has substantially reduced the adversarial accuracy of the visual object detectors. To make these object detectors robust to adversarial attacks, a new Gabor filter-based method has been proposed in this paper. This method has then been applied on the YOLOv3 with different backbones, the SSD with different input sizes and on the FRCNN; and thus, six robust object detector models have been presented. In order to evaluate the efficacy of the models, they have been subjected to adversarial training via three types of targeted attacks (TOG-fabrication, TOG-vanishing, and TOG-mislabeling) and three types of untargeted random attacks (DAG, RAP, and UEA). The best average accuracy (49.6%) was achieved by the YOLOv3-d model, and for the PASCAL VOC dataset. This is far superior to the best performance and accuracy and obtained in previous works (25.4%). Empirical results show that, while the presented approach improves the adversarial accuracy of the object detector models, it does not affect the performance of these models on clean data.

Recent studies have recently exploited knowledge distillation (KD) technique to address time-consuming annotation task in semantic segmentation, through which one teacher trained on a single dataset could be leveraged for annotating unlabeled data. However, in this context, knowledge capacity is restricted, and knowledge variety is rare in different conditions, such as cross-model KD, in which the single teacher KD prohibits the student model from distilling information using cross-domain context. In this study, we aim to train a robust, lightweight student under the supervision of several expert teachers, which provide better instructive guidance compared to a single student-teacher learning framework. To be more specific, we first train five distinct convolutional neural networks (CNNs) as teachers for semantic segmentation on several datasets. To this end, several state-of-the-art augmentation transformations have also been utilized in training phase of our teachers. The impacts of such training scenarios are then assessed in terms of student robustness and accuracy. As the main contribution of this paper, our proposed multi-teacher KD paradigm endows the student with the ability to amalgamate and capture a variety of knowledge illustrations from different sources. Results demonstrated that our method outperforms the existing studies on both clean and corrupted data in the semantic segmentation task while benefiting from our proposed score weight system. Experiments validate that our multi-teacher framework results in an improvement of 9% up to 32.18% compared to the single-teacher paradigm. Moreover, it is demonstrated that our paradigm surpasses previous supervised real-time studies in the semantic segmentation challenge.

In this paper, we propose a novel Interval Type-2 (IT2) Fuzzy Cognitive Map (FCM)-based flight control system to solve the altitude, attitude and position control problems of quadcopters. The proposed IT2-FCM encompasses all concepts related to drone for a satisfactory path-tracking and stabilizing control performance. The degree of mutual influences of the concepts is designed with opinions of three experts that take account the dynamics of drone and rules governing proportional integral derivative (PID) controllers. To model the inter-uncertainty of the experts’ opinions, IT2 fuzzy logic systems are utilized as they are powerful tools to model high level of uncertainties. Thus, the proposed IT2-FCM has a qualitative representation as it merges the advantages of IT2 fuzzy logic systems and FCMs. We present comparative simulations results in presence of uncertainties where the superiority of the proposed IT2-FCM-based flight control system is shown in comparison with its type-1 fuzzy counterpart.

Advances in digital equipment and organizations interest in having comprehensive and real-time information about products have increased the use ofmachine vision systems in organizations. In this pa

Good and regular inspections of transportation infrastructures such as bridges and overpasses are necessary to maintain the safety of the public who uses them and the integrity of the structures. Until recently, these inspections were done entirely manually by using mainly visual inspection to detect defects on the structure. In the last few years, inspection by drone is an emerging way of achieving inspection that allows more efficient access to the structure. This paper describes a human-in-the-loop system that combines AI and AR for bridge inspection by drone., VAMR 2022, Virtual, Augmented and Mixed Reality 14th International Conference, held as part of HCII 2022, the 24th HCI International Conference, June 26th - July 1, 2022, Virtual Event, Series: Lecture Notes in Computer Science

In this study, a new approach based on fuzzy cognitive map (FCM) and neuro-fuzzy inference system (NFIS), called the neuro-fuzzy cognitive map (NFCM), is proposed. Here, the NFCM is used for diagnosis of autoimmune hepatitis (AIH). AIH is a chronic inflammatory liver disease. AIH primarily affects women and typically responds to immunosuppressive therapy with clinical, biochemical, and histological remission. An untreated AIH can lead to scarring of the liver and ultimately to liver failure. If rapidly diagnosed, AIH can often be controlled by medication. NFCM is a new extension of FCM, which employs a NFIS to determine the causal relationships between concepts. In the proposed approach, weights are calculated using the knowledge and experience of experts as well as the advantages of NFIS. This makes the presented model more accurate. Having a high convergence speed, the proposed NFCM model performs well by achieving an AIH diagnosis accuracy of 89.81%. The superiority of the proposed NFCM model over the conventional FCM is that, it uses the NFIS to determine the link weights which train system parameters.

In this paper, we propose a nonparametric Bayesian learning framework for clustering problem based on multivariate Beta mixture model. Mixture models have been widely used as an unsupervised learning method in many machine learning, data mining and pattern recognition applications. One critical challenge in applying mixture model is the selection of the proper number of mixture components which best describes the data. Our approach can be viewed as an extension of the finite mixture model to infinite to tackle the model selection problem. In particular, our learning approach is Bayesian and relies on estimation of posterior distribution using Markov Chain Monte Carlo technique. The performance of our proposed method is evaluated through multiple challenging applications and we show that clustering via infinite multivariate Beta mixture models provides a more powerful performance comparing with various other approaches.

A dangerous infectious disease of the current century, the COVID-19 has apparently originated in a city in China and turned into a widespread pandemic within a short time. In this paper, a novel method has been presented for improving the screening and classification of COVID-19 patients based on their chest X-Ray (CXR) images. This method eliminates the severe dependence of the deep learning models on large datasets and the deep features extracted from them. In this approach, we have not only resolved the data limitation problem by combining the traditional data augmentation techniques with the generative adversarial networks (GANs), but also have enabled a deeper extraction of features by applying different filter banks such as the Sobel, Laplacian of Gaussian (LoG) and the Gabor filters. To verify the satisfactory performance of the proposed approach, it was applied on several deep transfer models and the results in each step were compared with each other. For training the entire models, we used 4560 CXR images of various patients with the viral, bacterial, fungal, and other diseases; 360 of these images are in the COVID-19 category and the rest belong to the non-COVID-19 diseases. According to the results, the Gabor filter bank achieves the highest growth in the values of the defined evaluation criteria and in just 45 epochs, it is able to elevate the accuracy by up to 32%. We then applied the proposed model on the DenseNet-201 model and compared its performance in terms of the detection accuracy with the performances of 10 existing COVID-19 detection techniques. Our approach was able to achieve an accuracy of 98.5% in the two-class classification procedure; which makes it a state-of-the-art method for detecting the COVID-19.

In this paper, a novel approach is presented to identify the smart home residents. The different behavioral patterns of smart home’s inhabitants are exploited to distinguish the residents. The variation of a specific individual behavior in smart homes is a significant challenge. We introduce different features that are useful to handle this problem. Moreover, we introduce an innovative strategy which considers the Bag of Sensor Events and Bayesian networks. In the Bag of Sensor Events approach, the frequency of each sensor event occurrence is considered, regardless of the order of sensor events. The efficiency of the Bag of sensor Events approach is compared to the Sequence of Sensor Events. Our experiments confirm that the Bag of Sensor Events approach outperformed the previous approaches. When the smart homes residents are people who repeat their daily activities frequently, applying the Bag of Sensor Events on Activity Based Window Frame features, which considers the performed daily activities, would identify them more accurately. In contrast, in cases where residents perform their activities in different ways, considering the Time Based Window Frame leads to higher accuracy in distinguishing residents. In this approach, the features are created by considering the constant time intervals. The F-measure of our proposed approach on the Twor2009, Tulum2009, and Tulum2010 datasets is 96%, 100%, and 99%, respectively, which improves the results of the previous researches which consider behavioral patterns to identify smart home residents.

Fuzzy cognitive maps (FCMs) represent a graphical modeling technique based on the decision-making and reasoning rules and algorithms similar to those used by humans. The graph-like structure and the execution model of FCMs respectively allow static and dynamic analyses to be carried out. The learning algorithms of FCMs that are based on expert opinion are weak in dynamic analysis, and fully automatic algorithms are weak in static analysis. In this paper, for providing the facility for simultaneous static and dynamic analyses, a new training algorithm called the quantum FCM (QFCM) is presented. In our proposed algorithm, the quantum inspired evolutionary algorithm (QEA) and the particle swarm optimization algorithm are employed for generating static and dynamic analyses properties respectively. In the QFCM, instead of coding the presence and absence of links between concepts with 1 and 0, respectively, the probability of their existence or inexistence is modeled with a Q-bit (the smallest information unit in the QEA) and, depending on the outcome of dynamic analysis, the quantum state of this Q-bit is updated. Using a probabilistic representation instead of 0 and 1, in addition to creating diversity in the solution space, can lead to escapes from many local optima; which is an issue of concern in the optimization of FCM structure. Experiments on synthetic, real-life, and gene regulatory network reconstruction problems demonstrated that not only does QFCM find potentially good structures, providing static analysis, but also it brings about low data error, showing good dynamic property. Furthermore, QFCM successfully outshined most of the state-of-the-art FCM’s learning algorithms, without any need to human knowledge, illustrating its power in this regard.

According to previous researches, Persian consonants have been divided into seven categories based on viseme. It led to several consonants being placed in one category. Detecting between consonants in one category is so hard because the spots for the production of these consonants are the same. The forms of lips do not change at the time of production; these consonants are hardly distinguishable. The major challenge is to recognize the differences between lip shapes in one category. The purpose of this study is to recognize differences between bilabial consonants such as /p/, /b/, and /m/ in a word that composed of consonant/vowel called CV by computer vision. For the first time, this study attempts to distinguish these consonants. Proper pronunciations of words are required to identify consonants. Therefore, a database has formed based on the videos of the speech therapists. Generally, this kind of process is including 1-lip detection, 2- lip feature extraction, and 3- classification systems for the diagnosis of consonants. In this paper, consonants recognition in a category based on lip shape using the CLM algorithm for lip detection is presented. Geometric algorithms for feature extraction and DTW and equalizer as a classification system are proposed. Although this study is open because we could identify differences among consonants in just one class, we could reach remarkable CV video results for the first time. We could aim for acceptable results with reasonable accuracy for bilabial consonants detection. The principle purpose of this study is to improve lip-reading systems in security issues and help hearing-impaired people in interaction with their surroundings. The results of this paper can have a positive effect on speech systems.

Automatically annotating the game of tennis using video playback is a high potential but has many challenges. In this research, deep learning in annotating tennis games with the integration of computer vision and machine learning is discussed. The experiments of this research are performed using a set of video images and the implementation of the CNN algorithm. The proposed method was compared with NAIVE BAYES, SVM, HMM, and S-SVM methods. The results show that well-tuned channel neural networks have the best performance among the strategies. U sing deep neural network convolution in Comparisons and evaluations showed that annotation is performed with great accuracy. The accuracy obtained in this study is 0.92. CNN's proposed algorithm showed that with the necessary changes in network parameters, and this algorithm's techniques, the desired result achieved, and accuracy greatly increased.

Objective evaluation of images is one of the most essential and practical aspects of image inpainting. The existing objective evaluation methods of image inpainting are functional only on an individual basis and do not provide an accurate and useful objective evaluation of inpainted images. Currently, there is no objective measure for evaluating inpainted images. In this study, an objective evaluation method was developed for image inpainting. In the proposed method, first, 100 images were inpainted using an exemplar-based algorithm. Then, the saliency map and its complementary region in the original image were obtained and a new objective measure was proposed for the evaluation of inpainted images based on the saliency map features. To make the assessment more realistic and comparable to human judgments, two terms, namely penalty and compensation, were taken into account. To assess the performance of our proposed objective measure, the inpainted images were also evaluated using a subjective test. The experiments demonstrates that the proposed objective measure correlated with the qualitative opinion in a human observer study. Finally, the objective measure was compared against three other measures, and the results showed that our proposed objective measure performed better than the other evaluation measures.

Autonomous cars use images of the road to detect drivable areas, identify lanes, objects near the car, and necessary information. This information achieved from the road images are used to make suitable driving decisions for self-driving cars. Drivable area detection is a technique that segments the drivable parts of roads in the image. Modern methods often consider road detection as a pixel by pixel classification task, which is struggling to solve the problem of computational cost and speed. So to increase the speed of performance, we consider the process of drivable area recognition as a row-selection task. In this paper, special rows in the image are selected. Then, the boundaries of the drivable area are detected in these rows. Therefore computational costs reduce significantly. This model is evaluated on the Berkley Deep Drive dataset, and the speed of the process arrives at +300 frames per second, which is faster than previous methods.

Autonomous vehicles need to make intelligent decisions in complex environments such as highways or even less organized environments within cities. Driver Assistance Systems are intelligent systems in autonomous vehicles that assist drivers in such environments. In this article, we aim to employ an intelligent supervisor by means of reinforcement learning algorithms to adopt appropriate driving policies in relation to vehicle transmission on highways. To make the real situation similar to our problem and take into account the probabilistic situation of the vehicle movements, we utilize the Implicit Quantile Network, which is suitable for stochastic environments. Implicit Quantile Network is adopted so that we can estimate the full return distribution rather than a single expected return value. The quantile function with differentiable functions can also be calculated by the Implicit Quantile Network. Different training and validation experiments have been performed in a highway driving simulator developed in the unity environment to demonstrate the performance of the proposed approach. The results verify that the proposed method can select the best driving policies for driving scenarios compared to the state-of-the-art studies and can deliver better performance in terms of both speed and lane changing metrics.

Powerful yet simple augmentation techniques have significantly helped modern deep learning-based text classifiers to become more robust in recent years. Although these augmentation methods have proven to be effective, they often utilize random or non-contextualized operations to generate new data. In this work, we modify a specific augmentation method called Easy Data Augmentation or EDA with more sophisticated text editing operations powered by masked language models such as BERT and RoBERTa to analyze the benefits or setbacks of creating more linguistically meaningful and hopefully higher quality augmentations. Our analysis demonstrates that using a masked language model for word insertion almost always achieves better results than the initial method but it comes at a cost of more time and resources which can be comparatively remedied by deploying a lighter and smaller language model like DistilBERT.

Lately, data mining tools have received significant attention because of their capability in modeling and analyzing collected data in various fields including medical research. With the growing availability of medical data, it is crucial to develop models that can discover hidden patterns in data and analyze them. Among various techniques, mixture models have been widely used for categorization problems in statistical modeling. In this paper, a Bayesian learning framework is proposed for multivariate Beta mixture model. Previous works have shown that multivariate Beta distribution can be considered as an alternative to Gaussian due to the flexibility of its shape and convincing performance. In particular, we use the Birth and Death Markov Chain Monte Carlo (MCMC) algorithm, which allows simultaneous parameters estimation and model selection. Experimental results on medical applications demonstrate the effectiveness of the proposed algorithm.

For many researchers, human gait has become a subject of interest for its unique biometric characteristic. However, designing an efficient biometric system based on standard parameters is a challenging task as most of the existing systems run into problem while extracting the features that are robust to different gait challenging conditions. The three major problems with the common gait biometric systems are as follows: (1) eliminating temporal ordering of gait in the final template, (2) describing the gait without utilizing an efficient human’s motion model, and (3) aggregating noises and gait defects in the template feature. Numerous efforts have been made to solve these problems as well as to extract the effective spatio-temporal features. However, none of these approaches has been able to simultaneously solve the three mentioned problems and each has its own limitations in identifying human gait. In this chapter, while reviewing the recent approaches, a method for properly describing human’s gait is presented and its performance on three well-known datasets is evaluated. The three criteria of FAR and FRR errors, time and memory used, and scalability are considered as a benchmark of system efficacy. Using the USF dataset, the Rank1 and Rank5 accuracies of the proposed algorithm are 76.01% and 86.59%, respectively, which shows an improvement of about 5% compared to the recent methods. In addition, the proposed system achieves an FAR error of 38/1000, FRR error of 23%, computational speed of 5.5 frames/s, and requires 3.6 Gbytes of memory. The evaluation results indicate the superiority of the proposed system in terms of error minimization, and so it can be used to solve human gait problems under real conditions.

In this paper, we propose a novel framework for transferred deep learning-based anomaly detection in hyperspectral images. The proposed framework includes four main steps. Firstly, the image2019s spectral dimension is reduced by applying the principal component analysis (PCA) to decrease computational time. Secondly, a deep convolutional neural network (CNN) is trained using only one image to learn the pixels2019 similarities in a picture. Consequently, a novel and well-designed algorithm entitled object area filtering (OAF) is employed to benefit from this learned similarity for extracting objects in the image. The OAF removes irrelevant objects by comparing their area to an acceptable anomaly area range. Lastly, the final result is obtained by multiplying the network output and binary map of anomalies. The receiver operating characteristic (ROC) is employed to evaluate the proposed framework. Extensive experimental evaluations demonstrate that the proposed framework substantially outperforms a significant number of comparable state-of-the-art methods. Finally, we empirically verify that the deep network exhibits excellent domain adaptability.

In this paper, the type and location of unbalanced and bearing faults in rotary machines are investigated. The data used in this work is provided by the rotary machine fault simulator. These data include 14 different states, including 13 modes with unbalanced and bearings faults at different locations and a healthy mode, recorded by six acceleration sensors. The empirical mode decomposition (EMD) method is used for signal decomposition and feature extraction. Nine different features of the signals are extracted and the PCA method is used to reduce the dimensions of the data. Also, in the classification section, the performance of three types of classifiers RBF, MLP and FCM are evaluated. The results show that the best performance pertains to the RBF method using two Kurtosis and RMS features with the number of intrinsic mode functions (IMF) equals 9 and dimensions reduced to 25. In this case, accuracy in the test phase was obtained at 96.42%.

Mixture models have been widely used as statistical learning paradigms in various unsupervised machine learning applications, where labeling a vast amount of data is impractical and costly. They have shown a significant success and convincing performance in many real-world problems such as medical applications, image clustering and anomaly detection. In this paper, we explore a fully Bayesian analysis of multivariate Beta mixture model and propose a solution for the problem of estimating parameters using Markov Chain Monte Carlo technique. We exploit Gibbs sampling within Metropolis-Hastings for Monte Carlo simulation. We also obtained prior distribution which is a conjugate for multivariate Beta. The performance of our proposed method is evaluated and compared with Bayesian Gaussian mixture model via challenging applications, including cell image categorization and network intrusion detection. Experimental results confirm that the proposed technique can provide an effective solution comparing to similar alternatives.

In this paper, an active Hebbian learning (AHL) for intuitionistic fuzzy cognitive map (iFCM) is proposed for grading the celiac. This method performs the diagnosis procedure automatically, and it is more suitable for specialists in better understanding and assessment of the disease. Our approach shows potential in confronting hesitancy through considering experts’ uncertainty in modeling. In this study, we propose an automatic computer-aided diagnosis system based on iFCMs to determine the grade of celiac disease. By relying on the knowledge of experts, the key features of disease are extracted as the main concepts, and the iFCM model for the complex grading system is designed as a graph with eight concepts. The results obtained by applying our proposed method (iFCM-AHL) on the dataset verify the ability and effectiveness of this model. The proposed iFCM by considering hesitation of experts in modeling process and property of less sensitive to missing input data, not only increase accuracy in detecting the type of disease, but also obtain a higher robustness, in dealing with incomplete data. The obtained results have been compared with the findings of the FCM, interval type-2 fuzzy logic system, untrained iFCM and five extensions of the FCM. Comparative results show that our approach offers a robust classification method that produces better performance than other models.

Soccer is the most popular sport in the world and the information extracted from this match has many uses. It can be used to extract players ‘paths, recognize the performance, and evaluate the players‘ statistics, evaluate the referee‘s decision, and so on. One of the main steps in analyzing soccer video is tracking players that seeks to locate players when playing video. Player tracking involves various processes, such as playfield detection, player detection, tracking of players, apparent modeling of players, and identification of players overlapping. One of the challenges in this field is the tracing of players in the shaded play field, which challenges the tracking of players due to light changes in the field. In this paper, using the proposed algorithm to identify and Tracking players in television shows with two shaded playfield and sun shades. The proposed method is identified the playfield by using the saliency map algorithm and shadow elimination, which will minimize the noise from the stadium area. Then by using the features of the color, brightness and edge we will recognize the players. Using the combination of Top-hat transformation and the morphological operation the lines of the playfield is detected. Finally, by using the results of the detection step we will track players. The tracker used in this study is an improved particle filter that uses a combination of color and edge features. The results of the proposed method demonstrate that the detection of play field has 93% accuracy. Also the proposed method tracks the detected players with 90% precision. Therefore tracking accuracy shows that light variations have very little effect on it.

Digital image inpainting is one of the most important areas in image processing science. Digital image inpainting is a set of methods to inpaint or refill the damaged areas of the images. Given the increasing use of image inpainting and the lack of a good metric for evaluating image inpainting, there is a challenge in this field. In this study an objective evaluation method for image inpainting is developed. In the proposed method, first, 100 images were inpainted using exemplar-based algorithm, then, the saliency map and its complementary region in the original image are obtained and based on saliency map features, a new objective measure for evaluation of inpainted images is proposed. A term called compensation have been taken into account. To assess the performance of the proposed objective measure, inpainted images are also evaluated using a subjective test. The experiments demonstrate that the proposed objective measure correlates with qualitative opinion in a human observer study. Finally, the objective measure is compared against three other measures and the results show that our proposed objective measure is better than the others.

Stable control of haptic interfaces is one of the most important challenges in haptic simulations, because any instability of a haptic interface can cause it to get far from the realistic sense. In this paper, the control strategies employed for a stable haptic rendering in an interactive virtual control laboratory are presented. In this interactive virtual laboratory, there are different scenarios to teach the control concepts, in which a haptic interface is used in the two cases of force control and position control. In this regard, two control strategies are employed to avoid instability. An energy-compensating controller is utilized to remove energy leakage. Besides, a fuzzy impedance control is used along with the energy-compensating controller for the position control scenarios. The results obtained indicate the proposed approaches practically guarantee the stability of the haptic interface for an educational application in practice.

In scene understanding for autonomous vehicles (AVs), models trained on the available datasets fail to generalize well to the complex, real-world scenarios with higher dynamics. In this work, we attempt to handle the distribution mismatch by employing the generative adversarial network (GAN) and weather modeling to strengthen the intra-domain data. We also alleviate the fragility of our trained models against natural distortions with state-of-the-art augmentation approaches. Finally, we assess our method for cross-domain object detection through CARLA simulation. Our experiments demonstrate that: (1) Augmenting training class with even limited intra-domain data captured from the adverse weather conditions boosts the generalization of the two kinds of object detectors; (2) Exploiting GANs and weather modeling to elaborately simulate the adverse, intra-domain weather conditions manages to surmount the adverse data scarcity issue for intra-domain object detection; (3) A combination of Augmix and style augmentations not only can promote the robustness of our trained models against different natural distortions but also can boost their performance in the cross-domain object detection; (4) Training GANs for unsupervised image-to-image translation by means of the existing, large-scale datasets outside of our training domain is found beneficial to alleviate image-based and instance-based domain shifts.

This paper presents a new method based on fuzzy cognitive map (FCM) and possibilistic fuzzy c-means (PFCM) clustering algorithm for categorizing celiac disease (CD). CD is a complex disorder whose development is affected by genetics (HLA alleles) and gluten ingestion. The celiac patients who are not treated are at a high risk of cancer, malignant lymphoma, and small bowel neoplasia. Therefore, CD diagnosis and grading are of paramount importance. The proposed FCM models human thinking for the purpose of classifying patients suffering from CD. We used the latest grading method where three grades A, B1, and B2 are used. To improve FCM efficiency and classification capability, a nonlinear Hebbian learning algorithm is applied for adjusting the FCM weights. To this end, 89 cases are studied. Three experts extracted seven main determinant characteristics of CD which were considered as FCM concepts. The mutual effects of these concepts on one another and on the final concept were expressed in the form of fuzzy rules and linguistic variables. Using the center of gravity defuzzifier, we obtained the numerical values of these weights and obtained the total weight matrix. Ultimately, combining the FCM model with PFCM algorithm, we obtained the grades A, B1, and B2 accuracies as 88, 90, and 91%, respectively. The main advantage of the proposed FCM is the good transparency and interpretability in the decision-making procedure, which make it a suitable tool for daily usage in the clinical practice.

Acute lymphoblastic leukemia is one of the most common hematologic malignancies among children, caused by uncontrolled growth of leukocytes. Since the main hallmarks of the disease is not specific, a considerable number of patients have been being misdiagnosed. Early diagnosis of the disease is usually made by morphological investigation of leukocytes under microscope. In light of the facts that decrease in cytoplasm-to-nucleus ratio is one of the main indicators of cancerous cells, and an accurate segmentation phase will lead to extraction of representative features, segmentation step is acknowledged as being crucial in design of a computer aided diagnosis (CAD). Previous researches have utilized standard, pre-defined color spaces, such as CMYK, for the segmentation of leukocyte nucleus. However, since these color spaces were not designed specifically for the segmentation task, they cannot extract nuclei efficiently. Thus, in this paper, by using feed forward neural networks, we propose a color transformation method, which maps RGB to a special 2D space. Design parameters of the neural network are tuned by using genetic algorithm. In the new space, nuclei of the leukocytes have the highest discrimination against background, so by using Otsu thresholding, one can extract the nuclei easily. Efficacy of the proposed method is evaluated by using ALL-IDB2 dataset, which is publicly available. Our obtained Dice similarity coefficient is higher than that of the other newly devised algorithms, showing its great performance.

Viewing heavenly objects in the sky helps astronomers understand how the world is shaped. Regarding the large number of objects observed by modern telescopes, it is very difficult to manually analyze it manually. An important part of galactic research is classification based on Hubble's design. The purpose of this research is to classify images of the stars using machine learning and neural networks. Particularly in this study, the galaxy's image is employed. The galaxies are divided into regular two-dimensional Hubble designs and an irregular bunch. The regular bands that are presented in the shape of the Hubble design are divided into two distinct spiral and elliptical galaxies. Spiral galaxies can be considered as elliptical or circular galaxies depending on the shape of the spiral, so the identification or classification of the spiral galaxy is considered important from other galaxies. In the proposed algorithm, the Sloan Digital Sky is used for testing, including 570 images. In the first step, its preprocessing operation is performed to remove image noise. In the next step, extracting the attribute from the galactic images takes place in a total of 827 properties using the sub-windows, the moments of different color spaces and the properties of the local configuration patterns. Then the classification is performed after extracting the property using a Support vector machine. And then compared with other methods, which indicate that our approach has worked better. In this study, the experiments were carried out in two spiral and elliptic classes and three spiral, elliptic and zinc-edged classes with a precision of 96 and 94 respectively.

In this article, a fast terminal sliding mode control technique is used for robust tracking control of a nonlinear uncertain mass–spring system in the existence of external perturbation. This system is considered as a benchmark problem in the flexible joint mechanisms. The joints flexibility in the robotic systems creates one of the most significant sources of parametric uncertainties. The theory of Lyapunov stability is used for the formulation of the proposed control method, and the presence of the sliding around the switching surface is satisfied in the finite time. Simulation results as well as the experimental verifications prove the efficiency and applicability of the suggested approach in the presence of parametric uncertainty, noise, and exterior disturbance.

In order to classify the dark objects observed in SAR images into oil spills or lookalikes, many features need to be extracted from these images. In this paper, an algorithm is presented for selecting an optimum set of features from SAR images; which maximizes the discrimination between oil spills and their lookalikes in such images. The proposed algorithm consists of the following sections: detection of dark spots in SAR images, extraction of features, selection of features, and the classification of dark spots into oil spills or lookalikes. It is observed that the proposed algorithm can accurately detect and classify the dark spots in SAR images. In extracting the features, 74 different kinds of features consisting of 32 textural features, 19 geometrical features, 19 physical features and 4 contextual features are extracted. In the feature selection step, eight different evolutionary algorithms are employed to yield the desired feature subsets. The obtained subsets are then evaluated based on the classification error rate criterion; while Bayesian network is used to classify the dark spots into oil spills or lookalikes. The proposed algorithm is applied to a data set of 134 oil spills and 118 lookalikes. The classification rate obtained by using the optimum set of features is 93.19 %.

By applying an image-based visual servoing (IBVS) method, the intelligent image-based controlling of a quadrotor type unmanned aerial vehicle (UAV) tracking a moving target is studied in this paper. A fuzzy cognitive map (FCM) is a soft computing method which is classified as a fuzzy neural system and exploits the main aspects of fuzzy logic and neural network systems; so it seems to be a suitable choice for implementing a vision-based intelligent technique. An FCM has been employed in implementing an IBVS scheme on a quadrotor UAV, so that the UAV can track a moving target on the ground. For this purpose, by properly combining the perspective image moments, some features with the desired characteristics for controlling the translational and yaw motions of a UAV have been presented. In designing a vision-based control method for a UAV quadrotor, there are some challenges, including the target mobility and not knowing the height of UAV above the target. Also, no sensor has been installed on the moving object and the changes of its yaw angle are not available. Despite all the stated challenges, the proposed method, which uses an FCM in controlling the translational motion and the yaw rotation of a UAV, adequately enables the quadrotor to follow the moving target. The simulation results for different paths show the satisfactory performance of the designed controller.

In this paper, we provide a novel technique based on a high-order fuzzy cognitive map (HFCM) to predict autoimmune hepatitis (AIH). The basic features that are extracted by specialists are used as the input concepts of the HFCM model. Particle swarm optimization (PSO) algorithm is used to enhance the capability and increase the efficiency of HFCM classification. In order to evaluate the performance, our method is applied to 216 patients. In this paper, we have also used the chaotic PSO (CPSO) algorithm; which, as extensions of PSO algorithm, improve the performance of PSO in terms of global optimality, reliability, convergence speed and solution accuracy. The results of applying different CPSOs are compared with classical PSO. The best results in this case, which are achieved by applying the CPSO, are 85.71%, 86.21% and 87.88% for the definite, probable and improbable classes, respectively. Therefore, the highest grading accuracies are achieved by using the combination of fourth order learned HFCM by CPSO.