ABSTRACT Millimetre wave (mmWave) multiple‐input multiple‐output (MIMO) has emerged as a promising physical layer solution to address traffic demands in beyond 5G wireless communication systems. However, modern signal processing techniques for mmWave hybrid MIMO systems fall short of addressing the performance degradations due to residual transceiver hardware impairments (HIs). This paper thus considers a mmWave hybrid MIMO system with residual HIs. Using Bussgang decomposition, the residual transceiver HIs are modelled as an additive non‐Gaussian noise that severely affects the received pilot and information signals, which makes channel state acquisition challenging. In this context of channel‐estimation over non‐Gaussian noise due to HIs, this work presents a hyperparameter‐free maximum Versoria criterion (MVC)‐based channel estimation technique. In details, the MVC‐based channel‐estimator is rendered hyperparameter‐free by proposing a sampling rule for its spread parameter and a gradient descent‐based optimisation for its shape parameter . Finally, simulations are presented to show the generalisation and scenario‐invariance of the proposed MVC‐based channel‐estimator. The analytical expression for steady‐state misadjustment is also derived and validated by simulations.

Generative artificial intelligence (AI) is rapidly transforming software development, especially in code generation. Large Language Models (LLMs) show strong potential for automating programming tasks, though their performance varies with task complexity. This study systematically evaluates state-of-the-art models, including OpenAI GPT-4.5 Preview, GPT-4o, GPT-4o Mini, GPT-4 Turbo, GPT-3.5 Turbo, GPT-o1, GPT-o3 Mini, Google’s Gemini 1.5 Pro, Gemini 1.5 Flash, Gemini 2.0 Flash, Gemini 2.0 Flash Lite, Anthropic’s Claude 3 Opus, Claude 3 Sonnet, Claude 3 Haiku, Claude 3.5 Sonnet, Claude 3.5 Haiku, Claude 3.7 Sonnet, and Meta’s LLaMA 3.0 8B Instruct, LLaMA 3.1 8B Instruct. These models were tested on ten Python programming tasks—five simple and five complex—and evaluated using an embedding-based semantic similarity approach. High-performing models such as GPT-4.5 Preview and GPT-4o Mini consistently produced accurate outputs, while LLaMA 3.1 8B Instruct performed weakest. Interestingly, complex tasks yielded higher similarity scores, likely due to their structured outputs. The results highlight the need for complementary metrics beyond semantic similarity, including execution correctness and efficiency. This study offers practical insights into AI-assisted coding and points toward future research directions for improving generative models in real-world applications.

The article presents a proposal of an algorithm for predicting the energy potential of a photovoltaic system. After a thorough study of real data on the amount of energy produced for the year 2022 from the photovoltaic system and meteorological data for the monitored period, a methodology for the design of a predictive solution algorithm was proposed, and subsequently the algorithm itself was developed to predict the amount of energy produced depending on meteorological data.

The selection between Deep Learning (DL) approaches is not easy for URL phishing due to the variety of attacks and scammers. There are various DL techniques to detect phishing URLs, and choosing the suitable algorithm and affecting the model formed is very important. Wrong-choosing DL techniques might lead to low maturity and produce bias. The trained model’s performance and accuracy would also be unsatisfactory if the wrong algorithms and methods were used. It often happens when the attackers change their phishing strategies frequently to target the system’s weaknesses and users’ naivety. The robust characteristics of DL algorithms have led the researchers to develop several URL phishing mitigation strategies. The techniques have been used to detect phishing attacks by using various URL features like URL length, URL domain, and other known features and further incorporating the new features. From the perspective of the Natural Language Processing (NLP) technique perspective, transformers are models designed to handle sequential text, such as summarizing and translating. One of the well-known transformers, called Keras Embedding, has a good application in detecting spam emails. As the transformers proved their usage in URL phishing detection, it is further hypothesized that the URLs can directly parse out the contextual meaning of the string and identify whether the website is benign or phishing. Therefore, this paper provides a URL phishing detection model with a combination of deep learning and natural language processing methods. As shown in the experiments, the result produces and improves with high performance and accuracy for URL phishing detection. We also examined and compared the findings of the proposed solution with deep learning only and NLP-only URL phishing detection approaches.

A large number of emerging applications, such as autonomous navigation, space exploration, surveillance, military target detection, and remote sensing, use outdoor images to monitor various activities of interest. However, images acquired under unfavorable weather conditions usually suffer from atmospheric scattering due to environmental pollution causing color-shift and low-contrast images. Dehazing is an emerging research area in the computer vision domain that intends to restore the visibility of images by eliminating the latter types of degradation. Single image dehazing, on the other hand, is more challenging since it necessitates a precise assessment of atmospheric light and transmission map. This study aims to design a dual-channel deep neural network (DCD-Net) for estimating the transmission map, further utilized to compute atmospheric light. Finally, a dehazed image is generated using the obtained atmospheric light and the transmission map. The experimental results are compared qualitatively and quantitatively with eight existing dehazing approaches based on ten metrics on six publicly available standard datasets: Foggy Road Image DAtabase, HazeRD, REalistic Single Image DEhazing, NYU-Depth, O-HAZE, I-HAZE, a few natural hazy images, and underwater images. The DCD-Net outperforms conventional techniques, according to extensive studies. Moreover, a range of relative improvements of the proposed method over other approaches is calculated for better analysis of the results. A visual internet of things (VIoT) framework employing a PYNQ-Z2 board is presented in addition to the DCD-Net. It can be applied in real-time applications, particularly in the transportation and surveillance industries. The DCD-Net is suitable for image dehazing by virtue of its multilayered structure. The VIoT uses the DCD-Net for dehazing, while the PYNQ-Z2 board serves as the central processing unit. Note to Practitioners—This paper is motivated by the problems occurring due to haze. Haze reduces the visibility of a scene, causing major concerns in transportation and surveillance. Existing approaches have attempted to address this issue, albeit the methods are limited. As a result, this study proposes a new dual-channel CNN model with two modules, where the first module calculates fine details of the image and the second module estimates the transmission map. Furthermore, both features are combined to produce a more reliable transmission map. The training process highly influences the resulting output of the network. Therefore, an algorithm explaining the training instructions for the practitioners is given in the appendix. The obtained transmission map is further used to estimate atmospheric light. The images are then dehazed using atmospheric light and transmission maps. In addition, we have designed a framework for image dehazing using VIoT with a PYNQ-Z2 board. Experimental results suggest that this approach gives expected results, yet, there is one limitation. This method requires a haze image and a corresponding transmission map, which is not always possible. Therefore, we will attempt to design a semi-supervised learning approach in the future.

Visibility issues in intelligent transportation systems are exacerbated by bad weather conditions such as fog and haze. It has been observed from recent studies that major road accidents have occurred in the world due to low visibility and inclement weather conditions. Single image dehazing attempts to restore a haze-free image from an unconstrained hazy image. We proposed a dehazing method by cascading two models utilizing a novel parameter-adaptive dual-channel modified simplified pulse coupled neural network (PA-DC-MSPCNN). The first model uses a new color channel for removing haze from images. The second model is the improved brightness preserving model (I-GIHE), which retains the brightness of the image while improving the gradient strength. To integrate the results from these two models and provide a pleasing haze-free image, a PA-DC-MSPCNN-based fusion is used. Furthermore, the proposed approach is deployed on a Xilinx Zynq SoC by exploiting the recently released PYNQ platform. The dehazing system runs on a PYNQ-Z2 all-programmable SoC platform, where it will input the camera feed through the FPGA unit and carry out the dehazing algorithm in the ARM core. This configuration has allowed reaching real-time processing speed for image dehazing. The results of dehazing are analyzed using both synthetic and real-world hazy images. Synthetic hazy images are acquired from the O-HAZE, I-HAZE, SOTS, and FRIDA datasets, while real-world hazy images are taken from the RailSem19, E-TUVD dataset, and the internet. For evaluation, twelve cutting-edge approaches are chosen. The proposed method is also analyzed on underwater and low-light images. Extensive experiments indicate that the proposed method outperforms state-of-the-art methods of qualitative and quantitative performances.

Polyps are very common abnormalities in human gastrointestinal regions. Their early diagnosis may help in reducing the risk of colorectal cancer. Vision-based computer-aided diagnostic systems automatically identify polyp regions to assist surgeons in their removal. Due to their varying shape, color, size, texture, and unclear boundaries, polyp segmentation in images is a challenging problem. Existing deep learning segmentation models mostly rely on convolutional neural networks that have certain limitations in learning the diversity in visual patterns at different spatial locations. Further, they fail to capture inter-feature dependencies. Vision transformer models have also been deployed for polyp segmentation due to their powerful global feature extraction capabilities. But they too are supplemented by convolution layers for learning contextual local information. In the present paper, a polyp segmentation model CoInNet is proposed with a novel feature extraction mechanism that leverages the strengths of convolution and involution operations and learns to highlight polyp regions in images by considering the relationship between different feature maps through a statistical feature attention unit. To further aid the network in learning polyp boundaries, an anomaly boundary approximation module is introduced that uses recursively fed feature fusion to refine segmentation results. It is indeed remarkable that even tiny-sized polyps with only 0.01% of an image area can be precisely segmented by CoInNet. It is crucial for clinical applications, as small polyps can be easily overlooked even in the manual examination due to the voluminous size of wireless capsule endoscopy videos. CoInNet outperforms thirteen state-of-the-art methods on five benchmark polyp segmentation datasets.

One of the challenging issues in detecting the malware is that modern stealthy malware prefers to stay hidden during their attacks on our devices and be obfuscated. They can evade antivirus scanners or other malware analysis tools and might attempt to thwart modern detection, including altering the file attributes or performing the action under the pretense of authorized services. Therefore, it’s crucial to understand and analyze how malware implements obfuscation techniques to curb these concerns. This paper is dedicated to presenting an analysis of anti-obfuscation techniques for malware detection. Furthermore, an empirical analysis of the performance evaluation of malware detection using machine learning algorithms and the obfuscation techniques was conducted to address the associated issues that might help researchers plan and generate an efficient algorithm for malware detection.

Worm-like ransomware strains spread quickly to critical systems such as IoMT without human interaction. Therefore, detecting different worm-like ransomware attacks during their spread is vital. Nevertheless, the low detection rate due to the imbalanced ransomware data and the detection systems’ disability for multiclass simultaneous detection are two apparent problems. In this work, we proposed a new approach for multi-classifying ransomware using preprocessing, resampling, and different classifiers. The proposed system uses network traffic NetFlow data, which is privacy-friendly and not heavy. In the first phase, preprocessing techniques were used on the collected and aggregated ransomware traffic, and then an optimized Synthetic Minority Oversampling Technique (SMOTE) was used for resampling the low-class samples. After that, four classifiers were applied, namely, Bayes Net, Hoeffding Tree, K-Nearest Neighbor, and a lightweight Multi-Layered Perceptron (MLP). The experimental results showed that the efficient preprocessing ensured accurate and simultaneous ransomware detection while the resampling technique improved the detection rate, F1, and PRC curve.

There exist numerous solutions to detect malicious URLs based on Natural Language Processing and machine learning technologies. However, there is a lack of comparative analysis among approaches using distributed representation and deep learning. To solve this problem, this paper performs a comparative study on phishing URL detection based on text embedding and deep learning algorithms. Specifically, character-level and word-level embedding were combined to learn the feature representations from the webpage URLs. In addition, three deep learning models, including Convolutional Neural Network (CNN), Bidirectional Gated Recurrent Unit (BiGRU), and Bidirectional Long Short-Term Memory (BiLSTM), were constructed for effective classification of phishing websites. Several experiments were conducted and various evaluation metrics were used to assess the performance of these deep learning models. The findings obtained from the experiments indicated that the combination of the character-level and word-level embedding approach produced better results than the individual text representation methods. Also, the CNN-based model outperformed the other two deep learning algorithms in terms of both detection accuracy and execution time.

With the advent of deep learning, the research on facial expression recognition (FER) has received a lot of interest. Different deep convolutional neural network (DCNN) architectures have been developed for real-time and efficient FER. One of the challenges in FER is obtaining trustworthy features that are strongly associated with facial expression changes. Furthermore, traditional DCNNs for FER problems have two significant issues: insufficient training data, which leads to overfitting, and intra-class facial appearance variations. FLEPNet, a texture-based feature-level ensemble parallel network for FER, is proposed in this study and proved to solve the aforementioned problems. Our parallel network FLEPNet uses multi-scale convolutional and multi-scale residual block-based DCNN as building blocks. First, we consider modified homomorphic filtering to normalize the illumination effectively, which minimizes the intra-class difference. The deep networks are then protected against having insufficient training data by using texture analysis on face expression images to identify multiple attributes. Four texture features are extracted and combined with the image's original characteristics. Finally, the integrated features retrieved by two networks are used to classify seven facial expressions. Experimental results reveal that the proposed technique achieves an average accuracy of 0.9914, 0.9894, 0.9796, 0.8756, and 0.8072 on Japanese Female Facial Expressions, Extended CohnKanade, Karolinska Directed Emotional Faces, Real-world Affective Face Database, and Facial Expression Recognition 2013 databases, respectively. Moreover, experimental outcomes depict significant reliability when compared to competing approaches.

Heterogeneous Face Recognition (HFR) is a challenging task due to the significant intra-class variation between the query and gallery images. The reason behind this vast intra-class variation is the varying image capturing sensors and the varying image representation techniques. Visual, Infrared, thermal images are the output of different sensors and viewed sketches, and composite sketches are the output of different image representation techniques. Conventional deep learning models are trying to solve the problem. Still, progress is impeded due to small HFR data samples, task-specific models (one model trained for face sketch-photo matching can’t perform well for NIR-VIS face matching), joint learning of two different HFR scenarios are not possible by one single deep network, and models are not interpretable. In this paper, to solve these major problems, we presented a novel interpretable Local Frequency Binary Pattern (LFrBP) based continual learning shallow network for HFR. The model is divided into two parts. A modality-invariant CNN model using the LFrBP feature, fine-tuned with CNN, is presented in the first part. The second part is based on continual learning to jointly learn the two HFR scenarios (face sketch-photo and NIR-VIS face matching) using a single network. Recognition results on different challenging HFR databases depict the superiority of the proposed model over other state-of-the-art deep learning-based methods.

This paper proposes a novel local feature descriptor coined as a local instant-and-center-symmetric neighbor-based pattern of the extrema-images (LINPE) to detect breast abnormalities in thermal breast images. It is a hybrid descriptor that combines two different feature descriptors: one is the inverse-probability difference extrema (IpDE), and another is the local instant and center-symmetric neighbor-based pattern (LICsNP). IpDE is developed to compute the intensity-inhomogeneity-invariant feature-based image of the breast thermogram. Besides, the LICsNP is intended to capture the local microstructure pattern information in the IpDE image. A new paradigm, named Broad Learning (BL) network, is introduced here as a classifier to differentiate the healthy and sick breast thermograms efficiently. The efficacy of the proposed system is quantitatively validated on the images of DMR-IR and DBT-TU-JU databases. Extensive experimentation on these databases with an average accuracy of 96.90% and 94%, respectively, justifies proposed system's superiority in the differentiation of healthy and sick breast thermograms over the other related existing state-of-the-art methods. The proposed system also performs consistently in the presence of noise and rotational changes.

Data-driven plays an important role in determining the quality of services in institutions of higher learning (HEIs). Increasingly data in education is encouraging institutions to find ways to improve student academic performance. By using machine learning with visual analytics, data can be predicted based on valuable information and presented with interactive visualizations for institutions to improve decision making. Therefore, predicting students’ academic performance is critical to identifying students at risk of failing a course. In this paper, we propose two approaches, such as (i) a prediction model for predicting students’ final grade based on machine learning that interacts with computational models; (ii) visual analytics to visualize predictive models and insightful data for educators. The data were tested using student achievement records collected from one of the Malaysian Polytechnic databases. The data set used in this study involved 489 first semester students in Computer System Architecture (CSA) course from 2016 to 2019. The decision tree algorithms (J48), Random Tree (RT), Random Forest (RF), and REPTree) was used on the student data set to produce the best predictions of the model. Experimental results show that J48 returns the highest accuracy with 99.8 %, among other algorithms. The findings of this study can help educators predict student success or failure for a particular course at the end of the semester and help educators make informed decisions to improve student academic performance at Polytechnic Malaysia.

This paper presents and discusses a comparative study of three major clustering categories namely Hierarchical-based, Iterative mode-based and Partition-based in analyzing and prioritizing Mobile Augmented reality (MAR) Learning (MAR-learning) usability data. This paper first discusses the related works in usability and clustering before moving on to the identification of gaps that can be addressed through experimentation. This paper will then propose a research methodology to measure four common clustering techniques on MAR-learning usability data. The paper will then discourse comparative results showing how Mini-batch K-means to be an ideal technique within the experimental setup. The paper will then present important research highlights, discussion, conclusion and future works.

Student performance is the most factor that can be beneficial for many parties, including students, parents, instructors, and administrators. Early prediction is needed to give the early monitor by the responsible person in charge of developing a better person for the nation. In this paper, the improvement of Bagged Tree to predict student performance based on four main classes, which are distinction, pass, fail, and withdrawn. The accuracy is used as an evaluation parameter for this prediction technique. The Bagged Tree with the addition of Bag, AdaBoost, RUSBoost learners helps to predict the student performance with the massive datasets. The use of the RUSBoost algorithm proved that it is very suitable for the imbalance datasets as the accuracy is 98.6% after implementing the feature selection and 99.1% without feature selection compared to other learner types even though the data is more than 30,000 datasets.

BACKGROUND: In biomedicine, infrared thermography is the most promising technique among other conventional methods for revealing the differences in skin temperature, resulting from the irregular temperature dispersion, which is the significant signaling of diseases and disorders in human body. Given the process of detecting emitted thermal radiation of human body temperature by infrared imaging, we, in this study, present the current utility of thermal camera models namely FLIR and SEEK in biomedical applications as an extension of our previous article. RESULTS: The most significant result is the differences between image qualities of the thermograms captured by thermal camera models. In other words, the image quality of the thermal images in FLIR One is higher than SEEK Compact PRO. However, the thermal images of FLIR One are noisier than SEEK Compact PRO since the thermal resolution of FLIR One is 160 × 120 while it is 320 × 240 in SEEK Compact PRO. CONCLUSION: Detecting and revealing the inhomogeneous temperature distribution on the injured toe of the subject, we, in this paper, analyzed the imaging results of two different smartphone-based thermal camera models by making comparison among various thermograms. Utilizing the feasibility of the proposed method for faster and comparative diagnosis in biomedical problems is the main contribution of this study.

Parameter selection or attribute selection is one of the crucial tasks in the data analysis process. Incorrect selection of the important attribute might generate imprecise or event for a wrong decision. It is an advantage if the decision-maker could select and apply the best model that helps in identifying the best-optimized attribute set — in the decision analysis process. Recently, many data scientists from various application areas are attracted to investigate and analyze the advantages and disadvantages of big data. One of the issues is, analyzing large volumes and variety of data in a big data environment is very challenging to the data scientists when there is a lack of a suitable model or no appropriate model to be implemented and used as a guideline. Hence, this paper proposes an alternative parameterization model that is able to generate the most optimized attribute set without requiring a high cost to learn, to use, and to maintain. The model is based on two integrated models that are combined with correlation-based feature selection, best-first search algorithm, soft set, and rough set theories which were compliments to each other as a parameter selection method. Experimental have shown that the proposed model has significantly shown as an alternative model in a big data analysis process.

This paper presents and discusses an empirical work of using machine learning K-means clustering algorithm in analyzing and processing Mobile Augmented Reality (MAR) learning usability data. This paper first discusses the issues within usability and machine learning spectrum, then explain in detail a proposed methodology approaching the experiments conducted in this research. This contributes in providing empirical evidence on the feasibility of K-means algorithm through the discreet display of preliminary outcomes and performance results. This paper also proposes a new usability prioritization technique that can be quantified objectively through the calculation of negative differences between cluster centroids. Towards the end, this paper will discourse important research insights, impartial discussions and future works.

This article presents the experimental work of comparing the performances of two machine learning approaches, namely Hierarchical Agglomerative clustering and K-means clustering on Mobile Augmented Reality Usability datasets. The datasets comprises of 2 separate categories of data, namely performance and self-reported, which are completely different in nature, techniques and affiliated biases. This research will first present the background and related literature before presenting initial findings of identified problems and objectives. This paper will the present in detail the proposed methodology before presenting the evidences and discussion of comparing this two widely used machine learning approach on usability data. This paper contributes in presenting evidences showing K-means as the better performing clustering algorithm when compared to Hierarchical Agglomerative when implemented on the usability datasets. The results shown has contradicted with some recent studies claiming otherwise, and the findings have created more research gaps pertaining the combined utilization of machine learning and usability analysis.

This paper presents fractions of research outcome from a bigger project involving machine learning, Hierarchical Agglomerative Clustering (HAC) Algorithms on usability data gathered through performance and self-reported data. This paper highlights the common problems in usability studies where the conventional analysis was frequently utilized while prioritizing usability issues. The utilization of clustering techniques is limited in the area of this study. A previous publication has shown how HAC was used in clustering usability problems in Mobile Augmented Reality (MAR) learning applications. However, there has not been a triangulation effort to confirm the first gathered results due to small datasets. This research presents a methodology adopted from previous studies in confirming earlier usability analysis results. The experiments found consistent evidence approving the feasibility of HAC in clustering and prioritizing Usability performance and self-reported data.

Handling the missing values play important step in the preprocessing phase of hydrological modeling analysis. One of the challenges in preprocessing phase is to deal with the problems of missing data with good consideration on the pattern and approaches of the missing data. Hence, this paper presents a study on Feedforward neural network algorithm (FFNN) and Elman neural network (ENN) imputation algorithm in estimating missing rainfall data at different percentages of missingness. Reliable rainfall data series from nearest neighbor gauging stations were used as inputs to predict the missing rainfall data for an output station. The selected study area is Sungai Merang, East Malaysia. The study revealed that ENN method demonstrated a superior prediction of the missing daily rainfall data than FFNN method. It is also observed that the ENN model-infilling method could be highly beneficial in reducing the data gaps for continuous hydrological modelling analysis.

Botnets are the most deadly threat in the network due to the capability of exploiting resources within a network as an army to launch huge attacks such as Denial-Distributed-of-Service (DDOS) or spam emails. Network Intrusion Detection System (NIDS) that designed based on the behavior of botnets in network traffic is seen as the promising technique in detecting botnets that are hiding by using encryption technique or any hiding techniques. This paper proposes on K-means clustering algorithm as the first phase of botnet's behaviour detection model that extracts data from network traffic. The criterion highlighted for our behaviour detection model is that it should be able to detect botnet in encrypted packets(hiding techniques), structure-independent (centralized and peer-to-peer), requiring minimal computing resources and minimal time processing. Other than that, by representing the real-time of network traffics, the detection model also must be resistant to noise and able to identify the anomaly of botnets behavior among a huge number of normal traffic. We are using the botnet benchmark dataset and normal traffic from Malware Capture Facility Project and comparing our proposed method using K-means algorithm with Expectation Maximization algorithm that proposed by the previous researcher in clustering the similar pattern of botnet behavior. The result shows that the K-means algorithm producing much higher accuracy, 94% and lower false negative rate, 0.1413. While, average accuracy for Expectation Maximization algorithm is 88% and False Negative Rate, 0.2245 with the insertion of uncertain data from real network traffic.

This paper highlights the current literatures in usability studies, performance metrics, self-reported metrics and hierarchical agglomerative clustering algorithms. A literature review is done in these three areas of studies to find a research gap that can be explored further. The paper will then propose a research methodology to study comparatively feature selection based on performance and self-reported usability data. This paper will highlight methods used to compare the feasibility and performance of hierarchical agglomerative clustering algorithms on both performance and self-reported data. The results of the experiment will then be presented and discussed before proceeding to the conclusion and future works of this study.

During disaster the communication behavior between society and crisis management authorities significantly changed due to technological evolutions and social media modernizations. The crowdsourcing and its platforms have gained importance for information exchange during and after disaster events. Social media and mobile apps are powerful and effective crowdsourcing platforms for collection of the data from various sources to collaborate and disseminate the processed information during emergency. However, data collection, integration and processing of unstructured data from diverse platforms, reliability and validity of data as well as privacy and security issues are the challenging part of crowdsourcing. The purpose of this study is to identify the various challenges through the systematic literature review during disaster management by using the crowdsourcing. Fifteen challenges have been highlighted and prioritized on the basis of their frequency of occurrence. Crowdsourcing has an important aspect of collecting and sharing of information during the disaster situation and aiming to reduce the disaster effects. It is recommended to address the identified crowdsourcing challenges for providing relief to affected community.

Fog Computing is a part of edge computing that define as intermediate layer between “Things” and the Cloud. Fog Health is an implementation of fog Computing's concept in health care and its related area. The purpose of this study is to extract and analyze the concept and application of Fog Health. The goal of this study are to identify the trend or patterns in Fog Health publications, to identify the application domain of Fog Health and to identify the research gap and future direction of implementing Fog Computing in health care related areas. Search term with relevant keywords were used to identify primary study related to the topic. About 53 of primary study were identified and selected. 46% (the largest portion) of the selected paper were journal articles. 47% of the publications were published by IEEE and 25 of the publications were published in 2017. We have found that the most three major issues that mostly discussed in fog Computing literatureas are related to the implementation of real-time system with minimum delay, the issues related to the performance on complex data processing that not affecting the system performance and the security & privacy issues related to the Fog Health implementations in the medical related facilities.

Providing sensory innervation, the ulnar nerve runs from the shoulder to the little finger; however, when entrapped in wrist, numbness and decreased sensation would occur. One of the reason is habitual misuse of computer keyboards in harmful wrist angles, causing constant pressure on the nerve. Therefore in this paper, we present a methodology for identification of the hands including localization of the wrists, supported by a fuzzy warning system. Initially, on the images taken by a camera mounted above a laptop monitor, hands and the left wrist are recognized. Subsequently, angle of the wrist is estimated and the fuzzy warning system is triggered by the angle and the duration, as inputs. While putting forward a new monitoring protocol, a novel application area of fuzzy warning system in preventive medicine is presented to improve the health of individuals and quality of life.

Identification methods based on biometrics are going through great expansion lately. Therefore in this article we propose a realization of experimental multibiometric system for laboratory verification of theoretical knowledge. The system identifies hand veins using biometric characteristics of hand contour and bloodstream on the dorsum of the hand. Moreover we alternatively put forward a fuzzy approach for segmentation of veins and estimation of the vein system by maximum curvature method.

This paper highlights the current literatures in usability studies, performance metrics and machine learning algorithm. A literature review is done in these three areas of studies to find a research gap that can be explored further. The paper will then propose a research methodology to attend to the issues of machine learning and usability. An experiment is proposed to compare the efficiency results in between data consistency, correlation between performance metrics and self-reported metrics of a Mobile Augmented Reality learning application. The methodology proposes hierarchical agglomerative clustering technique as a solution in differentiating usability issues according to priority in order to help with usability re-engineering decisions. This paper proposes two objectives through the proposed framework and present evidence on how to achieve them. Lastly, this paper will discuss the results, conclusion and future works of the proposed study.

With the benefits of context-aware and the smartphone's participatory sensing potential, individual activity recognition (IAR) has proven to be enormously importance for stampede prediction in a crowd using a geographical location and global positioning system (GPS) data. In the case of an unforeseen incident and in an emergency situations whether in a small or large gathering. The research effort used Kalman filter to remove uncertainty through sensor fusion to create room for a reliable measurement for abnormality prediction. This paper, addressed the following questions. (i) How to determine the flow direction and the velocity of peoples' movement in a crowd to know when stampede will occur? (ii) What is the role of sensor fusion in a crowd scenario? Two scenarios experimented on IAR with accelerometer, GPS, and digital compass sensors to determine the flow pattern of participants' movement in a crowd using the flow velocity Vsi and flow direction Dsi, in the proposed stampede prediction approach. The experimental results show the effect of Vsi and Dsi for different group locations and serve as a pointer to reduce risk towards mitigation of crowd disaster and enhanced the existing context-aware framework to save human lives in our society if used in crowd scenarios.

This paper deals with the issue of procrastination in e-learning. Suggested approach is based on compensating e-learning shortcomings and applying principles of forming a habit. Technical implementation is possible through use of mobile devices, incorporated in e-learning strategy. Respective habit loop would consist of immediate trigger (delivered by a reminder application), desired behavior (engagement in learning session) and immediate reward. Requirements on learning strategy, software and hardware are discussed, as well as a reminder mechanism and relevant system of rewards. Data processing in the reminder application is outlined for computing initial settings of the application.

This paper presents a framework for web-based application, which aspires to maintain learning resources, for purposes of learning courses or organizations. The goal is to create resource-rich hierarchical learning environment, which supports collaborative building of learning resources for specific domain. This paper presents basic model with essential user activities, proposal of information architecture and implementation. Presented version of the learning management system is based on principle of shared hierarchy, user contribution and moderated improvement of learning resources. This includes possibility of entry customization for active students as well as creating new entries and custom groups. Teachers can then evaluate new custom entries, which – if approved – they can add to shared hierarchy, accessible for other learners as well. This continuous process can successfully lead to collaborative building of learning resources and thus enrich communities of teachers and active students, interested in one specific domain.

Existence of software platform, which will allow us to monitor the patients' bio-parameters and provide us with services which help with full health care, is more than relevant these days. Increasing amount of information, a new trend in home health care or desire of individuals to increase their life quality are only some aspects which confirm this need. Project Guardian concerns with this problem. Its aim is to provide solution which can be used in different spheres of health care and which will be available through PDA (Personal Digital Assistant), web or desktop clients.