Search

Nowadays, simulators are being used more and more during the development of robotic systems due to the efficiency of the development and testing processes of such applications. Undoubtedly, these simulators save time, resources and costs, as well as enable ease of demonstrations of the system. Specifically, tools like the open source Robotic Operating System (ROS) and Gazebo have gained popularity in building models of robotic systems. ROS is extensively used in robotics due to the pros of hardware abstraction and code reuse. The Gazebo platform is used for visualisation because of its high compatibility with ROS. In this paper, ROS and Gazebo have been integrated to build an interface for the visualisation of the Katana Arm manipulator.

Biometric presentation attack detection is gaining increasing attention. Users of mobile devices find it more convenient to unlock their smart applications with finger, face or iris recognition instead of passwords. In this paper, we survey the approaches presented in the recent literature to detect face and iris presentation attacks. Specifically, we investigate the effectiveness of fine tuning very deep convolutional neural networks to the task of face and iris antispoofing. We compare two different fine tuning approaches on six publicly available benchmark datasets. Results show the effectiveness of these deep models in learning discriminative features that can tell apart real from fake biometric images with very low error rate. Cross-dataset evaluation on face PAD showed better generalization than state of the art. We also performed cross-dataset testing on iris PAD datasets in terms of equal error rate which was not reported in literature before. Additionally, we propose the use of a single deep network trained to detect both face and iris attacks. We have not noticed accuracy degradation compared to networks trained for only one biometric separately. Finally, we analyzed the learned features by the network, in correlation with the image frequency components, to justify its prediction decision., A preprint of a paper accepted by IET Biometrics journal and is subject to Institution of Engineering and Technology Copyright

In 2017, the Ministry of Health in Gaza introduced Early Essential Newborn Care (EENC) as its primary maternal and neonatal care strategy. EENC comprises a package of simple evidence-based interventions that are delivered during labour and delivery, and in the early post-partum period, to prevent or treat the most important causes of morbidity and mortality in newborn babies. Four public maternity hospitals in Gaza, responsible for approximately 75% of all deliveries, began implementation of EENC in June, 2017. Clinical coaching was delivered by national facilitators over 2 days, and targeted all clinical staff in maternity and neonatal units. Subsequently, EENC quality improvement teams were formed to address contextual factors that influence practice. This study aimed to determine whether introduction of EENC resulted in changes in clinical practices for vaginal births.A pre-intervention and post-intervention design was used to review key clinical practices before and after EENC introduction in the four hospitals. Trained data collection staff visited each hospital for 1 day in each of the months of March and June, 2017 (before EENC implementation), January and April, 2018 (in the early stages of EENC implementation), September, 2018, and June, 2019 (after full EENC implementation). Standard WHO data collection methods and tools were used to gather practice data using exit interviews and chart reviews of 10-15 randomly selected post-partum mothers who had delivered vaginally in the previous 2-24 h and had not experienced a newborn death or stillbirth. Delivery observations were conducted for five to ten randomly selected vaginal deliveries using a standard clinical skills observation checklist, beginning at the second stage of labour. The Ministry of Health in Gaza approved EENC assessments for programme use, and informed verbal consent was obtained before maternal interviews. No personal identifiers were used in assessments.259 maternal post-partum interviews and 139 observations of birth practices were done across the four maternity hospitals, representing 8·8% (259 of 2940) and 4·7% (139) of expected vaginal births during the observation periods, respectively. Comparing practices at baseline, early implementation, and after full implementation, significant trend improvements were noted for proportion of babies receiving thorough drying (0% [0 of 12], 49% [32 of 66], 72% [43 of 60], respectively, p0·0004), immediate skin-to-skin contact (SSC) for less than 1 min (0% [0 of 14], 33% [43 of 127], 66% [72 of 110], p0·0001), uninterrupted SSC for at least 60 min (0% [0 of 14], 21% [27 of 129], 48% [53 of 111], p0·0001), uninterrupted SSC for at least 90 min (0% [0 of 14], 10% [13 of 129], 36% [39 of 110], p0·0001), early breastfeeding (15-90 min after birth) (0% [0 of 15], 39% [50 of 130], 61% [65 of 107], p0·0001), breastfeeding before separation (0% [0 of 15], 28% [36 of 131], 52% [56 of 108], p0·0001), and exclusive breastfeeding before discharge (33% [5 of 15], 68% [89 of 131], 81% [87 of 107], p=0·0010). Average clinical practice scores rose from five of 42 (12%) to 16 of 42 (38%) and 24 of 42 (57%). Practice improvements were supported by updated clinical guidelines, hospital policies, and routines, by reorganisation of work, and by the provision of simple supplies, including gowns for mothers and caps for newborn babies.The EENC clinical coaching approach coupled with regular self-assessments and action by hospital teams has significantly improved care practices during delivery and in the early post-partum period. It is possible that periodic cross-sectional practice reviews were not representative of routine practices, which may have varied with time of day, case load, and case complexity. Limitations were mitigated by assessing a systematic random sample of post-partum women delivering throughout the previous 24 h, and by measuring practices in two different time periods in each phase of implementation. Post-partum interviews were used to limit the Hawthorne effect. No other maternal or newborn initiatives were introduced during the study period, and no additional staff training was available, therefore the EENC approach was the primary influence on health worker practices.Support for this work was provided by WHO, occupied Palestinian territory.

The Lateral Geniculate Nucleus (LGN) represents one of the major processing sites along the visual pathway. Despite its crucial role in processing visual information and its utility as one target for recently developed visual prostheses, it is much less studied compared to the retina and the visual cortex. In this paper, we introduce a deep learning encoder to predict LGN neuronal firing in response to different visual stimulation patterns. The encoder comprises a deep Convolutional Neural Network (CNN) that incorporates visual stimulus spatiotemporal representation in addition to LGN neuronal firing history to predict the response of LGN neurons. Extracellular activity was recorded in vivo using multi-electrode arrays from single units in the LGN in 12 anesthetized rats with a total neuronal population of 150 units. Neural activity was recorded in response to single-pixel, checkerboard and geometrical shapes visual stimulation patterns. Extracted firing rates and the corresponding stimulation patterns were used to train the model. The performance of the model was assessed using different testing data sets and different firing rate windows. An overall mean correlation coefficient between the actual and the predicted firing rates of 0.57 and 0.7 was achieved for the 10 ms and the 50 ms firing rate windows, respectively. Results demonstrate that the model is robust to variability in the spatiotemporal properties of the recorded neurons outperforming other examined models including the state-of-the-art Generalized Linear Model (GLM). The results indicate the potential of deep convolutional neural networks as viable models of LGN firing. Supplementary Information The online version contains supplementary material available at 10.1186/s40708-021-00132-6.

Over the past few years, Deep learning-based methods have shown encouraging and inspiring results for one of the most complex tasks of computer vision and image processing; Image Inpainting. The difficulty of image inpainting is derived from its’ need to fully and deeply understand of the structure and texture of images for producing accurate and visibly plausible results especially for the cases of inpainting a relatively larger region. Deep learning methods usually employ convolution neural network (CNN) for processing and analyzing images using filters that consider all image pixels as valid ones and usually use the mean value to substitute the missing pixels. This result in artifacts and blurry inpainted regions inconsistent with the rest of the image. In this paper, a new novel-based method is proposed for image inpainting of random-shaped missing regions with variable size and arbitrary locations across the image. We employed the use of dilated convolutions for composing multiscale context information without any loss in resolution as well as including a modification mask step after each convolution operation. The proposed method also includes a global discriminator that also considers the scale of patches as well as the whole image. The global discriminator is responsible for capturing local continuity of images texture as well as the overall global images’ features. The performance of the proposed method is evaluated using two datasets (Places2 and Paris Street View). Also, a comparison with the recent state-of-the-art is preformed to demonstrate and prove the effectiveness of our model in both qualitative and quantitative evaluations.

Cloud computing is increasing rapidly as a successful paradigm presenting on-demand infrastructure, platform, and software services to clients. Load balancing is one of the important issues in cloud computing to distribute the dynamic workload equally among all the nodes to avoid the status that some nodes are overloaded while others are underloaded. Many algorithms have been suggested to perform this task. Recently, worldview is turning into a new paradigm for optimization search by applying the osmosis theory from chemistry science to form osmotic computing. Osmotic computing is aimed to achieve balance in highly distributed environments. The main goal of this paper is to propose a hybrid metaheuristics technique which combines the osmotic behavior with bio-inspired load balancing algorithms. The osmotic behavior enables the automatic deployment of virtual machines (VMs) that are migrated through cloud infrastructures. Since the hybrid artificial bee colony and ant colony optimization proved its efficiency in the dynamic environment in cloud computing, the paper then exploits the advantages of these bio-inspired algorithms to form an osmotic hybrid artificial bee and ant colony (OH_BAC) optimization load balancing algorithm. It overcomes the drawbacks of the existing bio-inspired algorithms in achieving load balancing between physical machines. The simulation results show that OH_BAC decreases energy consumption, the number of VMs migrations and the number of shutdown hosts compared to existing algorithms. In addition, it enhances the quality of services (QoSs) which is measured by service level agreement violation (SLAV) and performance degradation due to migrations (PDMs).

In the last few years, deep learning has shown significant improvement for many computer vision open problems, especially Image inpainting. Image inpainting is the process of filling missing regions across images. One of the most challenging problems in image inpainting is face inpainting. In this work, a new novel approach for face inpainting is proposed which can capture and preserve the identity of each human face in images while reproducing the missing irregular region in images. A two-stage cascaded model is proposed. It is composed of a shape-predictor of the key-points of the face followed by an inpainting network. The shape-predictor identifies the human face's structure-preserving its local points, i.e. eyes, mouth, nose, and then the inpainting network fills any random-irregular missing regions guided by the obtained knowledge as a priori. The effectiveness of the proposed model was evaluated using the CelebA dataset. The obtained results from the trained model outperform the recently proposed technique with contextual attention.

The growing usage of smart mobile devices have made authentication with biometric data more convenient. On the other side, videos and photos of users are becoming more available online. This makes it easier for attackers to spoof the authentication systems which rely on face and eye-region data for instance. One major problem with current Presentation Attack Detection (PAD) systems is their lack of generalization to data captured by different sensors or in different environments. In this paper, we propose the use of unsupervised domain adaptation to solve this PAD problem, specifically the iris PAD. Our model is composed of symmetric classifiers and two per-class domain discriminators. Interaction between class probabilities and domain classification is utilized to jointly adversarialy train a mobile-oriented feature extraction network, capable of generating domain-invariant features. The approach is evaluated on three benchmark iris PAD datasets. Results show up to 40% improvement in cross-dataset Average Classification Error Rate (ACER) proving the effectiveness of the approach in increasing the robustness and generalization of biometric PAD systems.

Data preprocessing represents one of the most crucial stages in data analysis. While developing novel machine learning techniques has been the main focus of research in the data science field, less attention has been given to data preprocessing. This paper examines the importance of discretization as a preprocessing step and helps achieve better classification performance compared to using continuous attributes. We examine the performance of multiple parametric and non-parametric discretization methods in conjunction with several machine learning classifiers for the problem of predicting the Intensive Care Unit (ICU) mortality. Our results demonstrate the significance of discretizing the input attributes in this problem. Using discretized data achieved a classification accuracy and F1 score of 89.19% and 0.38, respectively, while using continuous attributes achieved a classification accuracy F1 score of 86.19% and 0.08, respectively. These results demonstrate that discretizing continuous attributes prior to applying machine learning models could significantly enhance performance enhancement.

Face Presentation Attack Detection (PAD) has drawn increasing attentions to secure the face recognition systems that are widely used in many applications. Conventional face anti-spoofing methods have been proposed, assuming that testing is from the same domain used for training, and so cannot generalize well on unseen attack scenarios. The trained models tend to overfit to the acquisition sensors and attack types available in the training data. In light of this, we propose an end-to-end learning framework based on Domain Adaptation (DA) to improve PAD generalization capability. Labeled source-domain samples are used to train the feature extractor and classifier via cross-entropy loss, while unsupervised data from the target domain are utilized in adversarial DA approach causing the model to learn domain-invariant features. Using DA alone in face PAD fails to adapt well to target domain that is acquired in different conditions with different devices and attack types than the source domain. And so, in order to keep the intrinsic properties of the target domain, deep clustering of target samples is performed. Training and deep clustering are performed end-to-end, and experiments performed on several public benchmark datasets validate that our proposed Deep Clustering guided Unsupervised Domain Adaptation (DCDA) can learn more generalized information compared with the state-of-the-art classification error on the target domain., 10 pages, 2 figures, to be published in IMPROVE 2021

With the huge amount of published research papers, retrieving relevant information is a difficult task for any researcher. Effective clustering algorithms can help improve and simplify the retrieval process. Here, we propose an approach for automatic clustering for text document using a Self-Organizing Map (SOM). It is one of unsupervised artificial neural network that widely used for data analysis, data compression, clustering, and data mining. The quality and accuracy of a SOM algorithm depends on the selection of values for some of its parameters which are its initial learning rate, SOM matrix dimensions, and the number of iterations. Best values are typically selected using trial and error; however, in the current paper we suggest a more systematic approach to parameters optimization using the genetic algorithm. The proposed method is applied to cluster 3 scientific papers datasets using their keywords. Similar research papers were mapped closer to each other. Clustering results were validated using the Dunn index.

With the increase of global accessibility of web applications, maintaining a reasonable security level for both user data and server resources has become an extremely challenging issue. Therefore, static code analysis systems can help web developers to reduce time and cost. In this paper, a new static analysis model is proposed. This model is designed to discover the security problems in scripting languages. The proposed model is implemented in a prototype SCAT, which is a static code analysis Tool. SCAT applies the phases of the proposed model to catch security vulnerabilities in PHP 5.3. Empirical results attest that the proposed prototype is feasible and is able to contribute to the security of real-world web applications. SCAT managed to detect 94% of security vulnerabilities found in the testing benchmarks; this clearly indicates that the proposed model is able to provide an effective solution to complicated web systems by offering benefits of securing private data for users and maintaining web application stability for web applications providers.

Goal models have been used for the last two decades in various disciplines to represent business, organizations, and individuals’ objectives. Several methodologies and standards have emerged, and various goal analysis and evaluation algorithms have been introduced serving different sectors, including decision support. In contrast to most researches which are based mainly on simulation to predict the satisfaction levels of final goals, this research proposes a new framework for evaluating goal models based on the state-space representation that is used extensively in control systems. This new approach brings the theories and literature of state-space representation of systems to goal models, opening a new direction for using its available mature techniques and tools, for goal model analysis and evaluation. A hypothetical goal model, which can be used for policymaking after the COVID-19 pandemic, is presented as an example of how the proposed framework can be used, and the results that can be obtained.

Mobile crowdsensing is a recent model in which a group of mobile users uses their smart devices such as smartphones or wearable devices to cooperatively perform a large-scale sensing task. In this paper, a novel model will be introduced for recognizing/classifying human activities that were collected from sensor units on the chest, legs, and arms. The proposed model employed the k-Nearest Neighbor (k-NN) classifier which is one of the most common classifiers. k-NN has only one parameter, k, to determine the number of selected nearest neighbors to the test or unknown samples for predicting the class labels of the unknown samples. Searching for the value of k which has a great impact on the classification performance is difficult especially with high dimensional data. This paper employs the Particle Swarm Optimization (PSO) algorithm to search for the optimal value of the k parameter in the k-NN Classifier. This paper shows first experimentally how the PSO in the proposed algorithm searches for the optimal value of k parameter to reduce the misclassification rate of the k-NN classifier. Then, in the second experiment, ten standard datasets are utilized to benchmark the performance of the proposed algorithm. For verification, the results of the PSO-kNN algorithm are compared with two well-known algorithms: Genetic Algorithm (GA) and Ant Bee Colony Optimization (ABCO). In the third experiment, the proposed PSO-kNN algorithm was employed for recognizing human activities. The experimental results proved that the PSO-kNN algorithm is able to find the optimal or near optimal value(s) of the k parameter which enhances the accuracy of k-NN classifier. The results also demonstrated lower error rates compared when GA and ABCO algorithms.

Solving the traffic congestion problem has many benefits financially and environmentally. The application of Artificial Intelligence to solving the traffic congestion problem has been going on for a while. However, most of the current research in this area depends on knowing lots of information about all vehicles in the network. While it produces promising results, applying these techniques in the current world is not easy. In this paper, we apply reinforcement learning to the field of traffic control under the assumption that only minimal information is available. Our approach produces results that are better than currently deployed fixed-time traffic lights without having heavy requirements. In our first test configuration, our agent's waiting time is 82.3% of the best fixed-time traffic lights' waiting time and the average CO 2 emissions produced by our agent is 97.5% of the emissions produced by the best fixed-time traffic lights. This shows the potential of applying reinforcement learning to the traffic control problem with limited state.

Discretizing continuous attributes is one essential and important data preprocessing step in data mining. Various data mining techniques are designed to be applied to discrete attributes. There have been tremendous efforts to propose discretization techniques with different characteristics. However, a clear pathway that can guide the choice of the needed discretization technique for different types of datasets is lacking. This paper proposes a taxonomy based on the existence of class information and relationship between attributes in the analyzed dataset. We review different discretization techniques classified according to the proposed taxonomy. The proposed taxonomy emphasizes the advantages and disadvantages of each discretization technique to be able theoretically to find a suitable discretization technique for a particular dataset.

A Self-Organizing Map (SOM) is a powerful tool for data analysis, clustering, and dimensionality reduction. It is an unsupervised artificial neural network that maps a set of n-dimensional vectors to a two-dimensional topographic map. Being unsupervised, SOMs need little input to be successfully deployed. The only inputs needed by a SOM are its own parameters such as its size, number of iterations, and its initial learning rate. The quality and accuracy of the solution offered by a SOM depend on choosing the right values for such parameters. Different attempts have been made to use the genetic algorithm to optimize these parameters for random inputs or for specific applications such as the traveling salesman problem. To the best knowledge of the authors, no roadmaps for selecting these parameters were presented in the literature. In this paper, we present the first results of a proposed roadmap for optimizing these parameters using the genetic algorithm and we show its effectiveness by applying it on the classical color clustering problem as a case study.

This work investigates which algorithm of scheduling algorithm will provide the best QoS for which type of application in addition to how stationary nodes and mobile nodes affect the network in terms of mobility and time. We used variable scheduling algorithms Weighted Fair Queuing (WFQ), Priority Queuing (PQ),First In First Out (FIFO) for different applications including real-time and non-real-time, in the context of mobility and stationery to investigate which algorithm of scheduling will provide the best quality of service with presenting a HetNet based LTE-A architecture including macrocell, picocells and femtocells.

Cloud computing meets numerous challenges at increasing number of users because the demand of resources sharing and usage are increased rapidly. Therefore, load balancing between resources is an important field for scheduling tasks to achieve better performance. In this chapter, a Hybrid Artificial Bee and Ant Colony optimization (H_BAC) load balancing algorithm is proposed. It depends on joining the important behavior of Ant Colony Optimization (ACO) such as discovering good solutions rapidly and Artificial Bee Colony (ABC) algorithm such as collective interaction of bees and sharing information by waggle dancing. The performance of the proposed algorithm is compared with ACO, ABC, and an existing hybrid algorithm. The simulation results show that H_BAC improves execution time, response time, makespan, resource utilization and standard deviation. This improvement reaches about 40% in the execution time and response time and 30% in the makespan over the other algorithms.

Images are exposed to deterioration over years due to many factors. These factors may include, but not limited to, environmental factors, chemical processing, improper storage, etc. Image inpainting has gained significant attention from researchers to recover the deteriorated parts in images. In this paper, two new techniques for image inpainting techniques using Deep Convolution Neural Networks (CNN) are proposed. In the first technique, a self-tuned Encoder-Decoder architecture based on a Fully Convolution Network (FCN) is used to generate different sized blocks from non-deteriorated image dataset with L2 being used as a loss measure. On the other hand, the second technique is a two-step technique inspired from Context Encoders. In the first step, Context Encoders are trained on non-deteriorated image dataset to select blocks from training images with minimum L2 loss. In the second step, the selected block is applied to Generative Adversarial Networks (GAN) in order to improve the quality of the recovered image. Several simulation examples were made to proof that the performance self-tuned Encoder-Decoder and GAN is the same. Simulations have also shown that the proposed methods have superior performance in recovering missing regions in deteriorated images over other state-of-art techniques. Paris Street View dataset was used for training and validation to validate our results.