High Processing Power Research Articles

ARGarden: 3D outdoor landscape design using handheld augmented reality with multi-user interaction

Augmented Reality (AR) is a technology to display the virtual object by overlaying on the physical marker in the real environment. Nowadays, the handheld device becomes a trend in AR since it has high processing power and compatible with the AR system. However, most of the AR applications did not fully support the interaction from multiple users for the collaborative AR interface. Therefore, this research aim is to develop multi-user interaction for collaborative handheld AR. There are a few phases involved to design and develop the collaborative AR application for the handheld device called ARGarden. All the interaction that has been done for the virtual object transformation such as translation, rotation and scaling can be seen by the other users, and this data is synchronized by using a network protocol. Then, a queue system is implemented for the multi-user interaction to allow the virtual object can only be selected by one user at one time, and the other users must wait for their turn. The significance of this research is all the method that has been implemented for the ARGarden can be applied in different fields such as simulation, modelling, medical and education.

Using distributed raspberry PIs to enable low-cost energy-efficient machine learning algorithms for scientific articles recommendation

In recent decades, machine learning has become a crucial factor in terms of automating business operations and assisting in the decision-making process within an organization. With the huge volume of data generated at an unprecedented rate has motivated researchers and industry analysts to constantly develop effective and efficient analytical models machine learning techniques. This study adds to the data mining community by evaluating some of the most significant text mining techniques and presenting a predictive model that will supposedly ease the process of literature review for researchers. In addition, it compares the execution of the model in terms of cost, energy consumption, accuracy and scalability in three different environments, namely, google cloud instance, google cloud functions and distributed raspberry PIs. Results yielded in our study showed that distributed Raspberry PIs can have a highly positive impact in terms of lowering costs and being energy efficient. On one hand, we found out that machine learning algorithms can be adapted and run on distributed raspberry PIs with low cost and low energy consumption compared to cloud alternatives. On the other hand, this solution does not offer great scalability and it requires more time on management, deployment and configuration. The distributed Raspberry PIs also showed bad performance on execution time compared to the other alternatives when comes to high processing power.

Comparisons of Facial Recognition Algorithms Through a Case Study Application

<p class="affiliations"><strong>Abstract— </strong>Computer visions and their applications have become important in contemporary life. Hence, researches on facial and object recognition have become increasingly important both from academicians and practitioners. Smart gadgets such as smartphones are nowadays capable of high processing power, memory capacity, along with high resolutions camera. Furthermore, the connectivity bandwidth and the speed of the interaction have significantly impacted the popularity of mobile object recognition applications. These developments in addition to computer vision’s algorithms advancement have transferred object’s recognitions from desktop environments to the mobile world. The aim of this paper to reveal the efficiency and accuracy of the existing open-source facial recognition algorithms in real-life settings. We use the following popular open-source algorithms for efficiency evaluations: Eigenfaces, Fisherfaces, Local Binary Pattern Histogram, the deep convolutional neural network algorithm, and OpenFace. The evaluations of the test cases indicate that among the compared facial recognition algorithms the OpenFace algorithm has the highest accuracy to identify faces. The findings of this study help the practitioner on their decision of the algorithm selections and the academician on how to improve the accuracy of the current algorithms even further.</p>

ASAS-NANP SYMPOSIUM: RUMINANT/NONRUMINANT FEED COMPOSITION: Challenges and opportunities associated with creating large feed composition tables.

Traditional feed composition tables have been a useful tool in the field of animal nutrition throughout the last 70 yr. The objective of this paper is to discuss the challenges and opportunities associated with creating large feed ingredient composition tables. This manuscript will focus on three topics discussed during the National Animal Nutrition Program (NANP) Symposium in ruminant and nonruminant nutrition carried out at the American Society of Animal Science Annual Meeting in Austin, TX, on July 11, 2019, namely: 1) Using large datasets in feed composition tables and the importance of standard deviation in nutrient composition as well as different methods to obtain accurate standard deviation values, 2) Discussing the importance of fiber in animal nutrition and the evaluation of different methods to estimate fiber content of feeds, and 3) Description of novel feed sources, such as insects, algae, and single-cell protein, and challenges associated with the inclusion of such feeds in feed composition tables. Development of feed composition tables presents important challenges. For instance, large datasets provided by different sources tend to have errors and misclassifications. In addition, data are in different file formats, data structures, and feed classifications. Managing such large databases requires computers with high processing power and software that are also able to run automated procedures to consolidate files, to screen out outlying observations, and to detect misclassified records. Complex algorithms are necessary to identify misclassified samples and outliers aimed to obtain accurate nutrient composition values. Fiber is an important nutrient for both monogastrics and ruminants. Currently, there are several methods available to estimate the fiber content of feeds. However, many of them do not estimate fiber accurately. Total dietary fiber should be used as the standard method to estimate fiber concentrations in feeds. Finally, novel feed sources are a viable option to replace traditional feed sources from a nutritional perspective, but the large variation in nutrient composition among batches makes it difficult to provide reliable nutrient information to be tabulated. Further communication and cooperation among different stakeholders in the animal industry are required to produce reliable data on the nutrient composition to be published in feed composition tables.

Medical Image Analysis (MedIA) using Deep Learning

Medical Image analysis has gained momentum in the research since last ten years. Medical images of different modalities like X-rays, Computed Tomography (CT), Magnetic Resonance Imaging (MRI), Ultrasound etc. are generated with an increase of 15% to 20% every year. Medical image analysis requires high processing power and huge memory for storing the medical images, processing them, extracting features for useful information and segment the interested area for analysis. Thus, here comes the role of deep learning which proves to be promising for medical image analysis. The major focus of the paper is on exploring the literature on the broad areas of medical image analysis like Image Classification, Tumor/lesion classification and detection, Organ/Sub-structure Segmentation, Image Registration and Image Construction/ Enhancement using deep learning. Paper also highlights the physiological and medical challenges to be taken care, while analyzing medical images. It also discusses the technical challenges of using deep learning for medical image analysis and its solutions.

A Framework for Object Detection using Deep-Reinforcement Machine Learning

Machine learning area enable the utilization of Deep learning algorithm and neural networks (DNNs) with Reinforcement Learning. Reinforcement learning and DL both is region of AI, it’s an efficient tool towards structuring artificially intelligent systems and solving sequential deciding problems. Reinforcement learning (RL) deals with the history of moves; Reinforcement learning problems are often resolve by an agent often denoted as (A) it has privilege to make decisions during a situation to optimize a given problem by collective rewards. Ability to structure sizable amount of attributes make deep learning an efficient tool for unstructured data. Comparing multiple deep learning algorithms may be a major issue thanks to the character of the training process and therefore the narrow scope of datasets tested in algorithmic prisons. Our research proposed a framework which exposed that reinforcement learning techniques in combination with Deep learning techniques learn functional representations for sorting problems with high dimensional unprocessed data. The faster RCNN model typically founds objects in faster way saving resources like computation, processing, and storage. But still object detection technique typically require high computation power and large memory and processor building it hard to run on resource constrained devices (RCD) for detecting an object during real time without an efficient and high computing machine.

A platform for cloudification of network and applications in the Internet of Vehicles

AbstractVehicular networks have evolved from traditional vehicle‐to‐vehicle (V2V) and vehicle‐to‐infrastructure (V2I) communications to the so‐called Internet of Vehicles (IoV), which encompasses cloud computing and the Internet of Things. Thus, IoV creates novel scenarios and applications such as autonomous driving, as well as research challenges regarding the requirements for connected cars such as ultra‐reliable low‐latency, communications and the need for high processing power closer to vehicles. In order to address the above issues for IoV, this article proposes integrating state‐of‐the‐art technology pillars that are well accepted by both industry and academia as building blocks for future 5G networks. We propose a flexible, hierarchical, and scalable IoV architecture that integrates software‐defined networking, edge/fog computing, network function virtualization, and containers. We evaluate the proposal in both infrastructure‐based and peer‐to‐peer content distribution scenarios. The results are promising and point out that the architecture can cope with the stringent requirements that need to be met in terms of latency for 5G‐based IoV applications.

High-performance GPU computations in nonlinear dynamics: an efficient tool for new discoveries

The main aim of this paper is to demonstrate the benefit of the application of high-performance computing techniques in the field of non-linear science through two kinds of dynamical systems as test models.It is shown that high-resolution, multi-dimensional parameter scans (in the order of millions of parameter combinations) via an initial value problem solver are an efficient tool to discover new features of dynamical systems that are hard to find by other means. The employed initial value problem solver is an in-house code written in C++ and CUDA C software environments, which can exploit the high processing power of professional graphics cards (GPUs). The first test model is the Keller–Miksis equation, a non-linear oscillator describing the dynamics of a driven single spherical gas bubble placed in an infinite domain of liquid. This equation is important in the field of cavitation and sonochemistry. Here, the high-resolution parameter scans gave us the opportunity to lay down the basis of a non-feedback technique to control multi-stability in which direct selection of the desired attractor is possible. The second test model is related to a pressure relief valve that can exhibit a special kind of impact dynamics called grazing impact. A fine scan of the initial conditions revealed a second focal point of the grazing lines in the initial-condition space that was hidden in previous studies.

Machine learning algorithms for improving security on touch screen devices: a survey, challenges and new perspectives

Mobile phone touch screen devices are equipped with high processing power and high memory. This led to users not only storing photos or videos but stored sensitive application such as banking applications. As a result of that the security system of the mobile phone touch screen devices becomes sacrosanct. The application of machine learning algorithms in enhancing security on mobile phone touch screen devices is gaining a tremendous popularity in both academia and the industry. However, notwithstanding the growing popularity, up to date no comprehensive survey has been conducted on machine learning algorithms solutions to improve the security of mobile phone touch screen devices. This survey aims to connect this gap by conducting a comprehensive survey on the solutions of machine learning algorithms to improve the security of mobile phone touch screen devices including the analysis and synthesis of the algorithms and methodologies provided for those solutions. This article presents a comprehensive survey and a new taxonomy of the state-of-the-art literature on machine learning algorithms in improving the security of mobile phone touch screen devices. The limitation of the methodology in each article reviewed is pointed out. Challenges of the existing approaches and new perspective of future research directions for developing more accurate and robust solutions to mobile phone touch screen security are discussed. In particular, the survey found that exploring of different aspects of deep learning solutions to improve the security of mobile phone touch screen devices is under-explored.

Unleashing the hidden powers of low-cost IoT boards: GPU-based edutainment case study

The Ubiquitous interconnected smart devices enabled by the recent evolution of low-cost, generic, small-size, powerful computing platforms devised the term Internet of Things (IoT), which cross-cuts many areas of our modern day living. IoT applications go way beyond simple sensing and actuation to sophisticated localized processing and decision-making. The recent advances in embedded systems produced a long list of IoT boards equipped with powerful central processing units (CPUs), and graphics processing units (GPUs). Unfortunately, even with the limited energy consumption and high processing power of such GPUs, CPUs are usually the only computational element utilized by the hosted applications, thus hindering the capabilities of the entire board. This is mainly due to the complicated nature of GPU-based programming. In this paper, we are presenting a case study showing the effect of offloading the computationally intensive part of a latency-sensitive educational game to a low-cost Raspberry Pi’s GPU, thus enabling the board to seamlessly host the entire game operations. Relying mainly on the boards CPU shows very long interaction latency i.e., 4.82 s. By efficiently leveraging the powerful coprocessor, VideoCore GPU, we are able to significantly improve the interaction latency to a fraction of a second, making the game conveniently playable.

Relationship between the radial dynamics and the chemical production of a harmonically driven spherical bubble

The sonochemical activity and the radial dynamics of a harmonically excited spherical bubble are investigated numerically. A detailed model is employed capable to calculate the chemical production inside the bubble placed in water that is saturated with oxygen. Parameter studies are performed with the control parameters of the pressure amplitude, the forcing frequency and the bubble size. Three different definitions of collapse strengths (extracted from the radius vs.time curves) are examined and compared with the chemical output of various species. A mathematical formula is established to estimate the chemical output as a function of the collapse strength; thus, the chemical activity can be predicted without taking into account the chemical kinetics into the bubble model. The calculations are carried out by an in-house code exploiting the high processing power of professional graphics cards (GPUs). The results shown that chemical activity can be approximated qualitatively from the values of relative expansion. This could be helpful in order to optimise chemical output of sonochemical reactors either from measurement data or simulations as well.

Practical Implementation of Privacy Preserving Clustering Methods Using a Partially Homomorphic Encryption Algorithm

The protection and processing of sensitive data in big data systems are common problems as the increase in data size increases the need for high processing power. Protection of the sensitive data on a system that contains multiple connections with different privacy policies, also brings the need to use proper cryptographic key exchange methods for each party, as extra work. Homomorphic encryption methods can perform similar arithmetic operations on encrypted data in the same way as a plain format of the data. Thus, these methods provide data privacy, as data are processed in the encrypted domain, without the need for a plain form and this allows outsourcing of the computations to cloud systems. This also brings simplicity on key exchange sessions for all sides. In this paper, we propose novel privacy preserving clustering methods, alongside homomorphic encryption schemes that can run on a common high performance computation platform, such as a cloud system. As a result, the parties of this system will not need to possess high processing power because the most power demanding tasks would be done on any cloud system provider. Our system offers a privacy preserving distance matrix calculation for several clustering algorithms. Considering both encrypted and plain forms of the same data for different key and data lengths, our privacy preserving training method’s performance results are obtained for four different data clustering algorithms, while considering six different evaluation metrics.

Hierarchical algorithms on hierarchical architectures.

A traditional goal of algorithmic optimality, squeezing out flops, has been superseded by evolution in architecture. Flops no longer serve as a reasonable proxy for all aspects of complexity. Instead, algorithms must now squeeze memory, data transfers, and synchronizations, while extra flops on locally cached data represent only small costs in time and energy. Hierarchically low-rank matrices realize a rarely achieved combination of optimal storage complexity and high-computational intensity for a wide class of formally dense linear operators that arise in applications for which exascale computers are being constructed. They may be regarded as algebraic generalizations of the fast multipole method. Methods based on these hierarchical data structures and their simpler cousins, tile low-rank matrices, are well proportioned for early exascale computer architectures, which are provisioned for high processing power relative to memory capacity and memory bandwidth. They are ushering in a renaissance of computational linear algebra. A challenge is that emerging hardware architecture possesses hierarchies of its own that do not generally align with those of the algorithm. We describe modules of a software toolkit, hierarchical computations on manycore architectures, that illustrate these features and are intended as building blocks of applications, such as matrix-free higher-order methods in optimization and large-scale spatial statistics. Some modules of this open-source project have been adopted in the software libraries of major vendors.This article is part of a discussion meeting issue ‘Numerical algorithms for high-performance computational science’.

Multi-level trust-based intelligence schema for securing of internet of things (IoT) against security threats using cryptographic authentication

The internet of things (IoT) is able to provide a prediction of linked, universal, and smart nodes that have autonomous interaction when they present services. Because of wide openness, relatively high processing power, and wide distribution of IoT things, they are ideal for attacks of the gray hole. In the gray hole attack, the attacker fakes itself as the shortest path to the destination that is a thing here. This causes the routing packets not to reach the destination. The proposed method is based on the AODV routing protocol and is presented under the MTISS-IoT name which means for the reduction of gray hole attacks using check node information. In this paper, a hybrid approach is proposed based on cryptographic authentication. The proposed approach consists of four phases, such as the verifying node trust in the IoT, testing the routes, gray hole attack discovery, and the malicious attack elimination process in MTISS-IoT. The method is evaluated here via extensive simulations carried out in the NS-3 environment. The experimental results of four scenarios demonstrated that the MTISS-IoT method can achieve a false positive rate of 14.104%, a false negative rate of 17.49%, and a detection rate of 94.5% when gray hole attack was launched.

Soil erosion prediction using the Revised Universal Soil Loss Equation (RUSLE) in Google Earth Engine (GEE) cloud-based platform

High-quality soils are an important resource affecting the quality of life of human societies, as well as terrestrial ecosystems in general. Thus, soil erosion and soil loss are a serious issue that should be managed, in order to conserve both artificial and natural ecosystems. Predicting soil erosion has been a challenge for many years. Traditional field measurements are accurate, but they cannot be applied to large areas easily because of their high cost in time and resources. The last decade, satellite remote sensing and predictive models have been widely used by scientists to predict soil erosion in large areas with cost-efficient methods and techniques. One of those techniques is the Revised Universal Soil Loss Equation (RUSLE). RUSLE uses satellite imagery, as well as precipitation and soil data from other sources to predict the soil erosion per hectare in tons, in a given instant of time. Data acquisition for these data-demanding methods has always been a problem, especially for scientists working with large and diverse datasets. Newly emerged online technologies like Google Earth Engine (GEE) have given access to petabytes of data on demand, alongside high processing power to process them. In this paper we investigated seasonal spatiotemporal changes of soil erosion with the use of RUSLE implemented within GEE, for Pindos mountain range in Greece. In addition, we estimated the correlation between the seasonal components of RUSLE (precipitation and vegetation) and mean RUSLE values.

ELASTIC CLOUD COMPUTING APPROACH FOR BIG DATA ANALYTICS

The latest developments in the communication and the information technologies have turned out to be the foundation for the emergence of the industrial developments and progress in the business causing digital transformation in the industrial and the business operations. The clubbing of the internet and the information technology along with the tangible things that are the responsible for the industrial operations, generate a huge set of data that requires enormous network bandwidth, high processing power and accessible resolutions. So the paper presents the elastic cloud computing approach to bring down the complexities in deploying, cost of the groundwork and maintenance and provide an automated resource provisioning according to the demands. This method increases the amount of the computing power available, network properties and the storage. The experimental results for the novel elastic approach for the big data analytics is obtained by the simulation through MATLAB in terms response time, processing power and cost.

Design and Implementation of Facial Recognition System for Visually Impaired using Image Processing

One of the most difficult tasks faced by the visually impaired students is identification of people. The rise in the field of image processing and the development of algorithms such as the face detection algorithm, face recognition algorithm gives motivation to develop devices that can assist the visually impaired. In this research, we represent the design and implementation of a facial recognition system for the visually impaired by using image processing. The device developed consists of a programmed raspberry pi hardware. The data is fed into the device in the form of images. The images are pre-processed and then the input image captured is processed inside the raspberry pi module using KNN algorithm, The face is recognized and the name is fed into text to speech conversion module. The visually impaired student will easily recognize the person before him using the device. Experiment results show high face detection accuracy and promising face recognition accuracy in suitable conditions. The device is built in such a way to improve cognition, interaction and communication of visually impaired students in schools and colleges. This system eliminates the need of a bulk computer since it employs a handy device with high processing power and reduced costs.

Distributed video coding scheme of multimedia data compression algorithm for wireless sensor networks

The emergence of multimedia data has enriched people’s lives and work and has penetrated into education, finance, medical, military, communications, and other industries. The text data takes up a small space, and the network transmission speed is fast. However, due to its richness, the multimedia data makes it occupy an ample space. Some high-definition multimedia information even reaches the GB level, and the multimedia data network transmission is relatively slow. Compared with the traditional scalar data, the multimedia data better describes the characteristics of the transaction, but at the same time, the multimedia data itself has a large capacity and must be compressed. Nodes of wireless multimedia sensor networks have limited ability to process data. Traditional data compression schemes require high processing power of nodes and are not suitable for sensor networks. Therefore, distributed video codec scheme in recent years becomes one of the hot multimedia sensor network technologies, which is a simple coding scheme, coding complexity of decoding performance. In this paper, distributed video codec and its associated knowledge based on the study present a distributed video coding scheme and its improvements. Aiming at the problem that the traditional distributed video coding scheme cannot accurately decode the motion severe region and the boundary region, a distributed video coding algorithm based on gradient-domain ROI is proposed, which can enhance the coding efficiency of the severe motion region and improve the decoded image while reducing the code rate and quality, ultimately reducing sensor node energy consumption.

A Novel Technique to Improve the Performance of fog Computing using fog Terminal Nodes

Cloud computing is considered to be technological revolution in the past decade, due to its reliability and flexibility in enabling anything-as-a-service to the end users based on the key principle of utility computing. With the advent of IoT and Real-time data processing continuous usage of cloud services have incremented the dependency levels of Cloud Data Centres which in a while required high processing power as well as it will be hazardous to the environment. Addressing this problem several research studies have identified FoG Computing as a next generation computing platform that enhances the performance of the cloud servers by processing the data at the edge devices. This paper presents a novel fog computing framework that enhances the performance of the data migration reducing the effort on cloud servers.

The contribution of research platforms to the pathway from off-line to real-time high-frame-rate imaging

Ultrasound research platforms are characterized by a great flexibility of parameters that are usually fixed on clinical scanners. They allow the transmission of arbitrary waveforms in arbitrary scan sequences, and to test novel methods by processing off-line the raw radiofrequency channel data received after each transmission (TX) event. In particular, the availability of research platforms has boosted the development of high-frame-rate (HFR) imaging methods. These are typically based on the TX of multi-line focused beams, plane waves or diverging waves, which can be optimally produced by linear power amplifiers. The echoes received after each TX event should thus be beamformed along multiple lines in parallel, which is feasible, on line, only when high processing power is available. In this talk, the characteristics of “hardware-based” research platforms, like the ULA-OP 256, are discussed. The presence of powerful FPGA devices, on board such systems, allows implementing efficient parallel beamforming methods, which is the key-requirement for HFR imaging. Sample results obtained in real-time at rates of hundreds frames/s will be presented for both multi-line and plane wave compounding scan sequences. Finally, the main challenges to be faced to extend such modalities to volumetric (3-D) imaging will be discussed.