Search

With the fast-moving deployment of IoT (Internet of Things) applications, a large amount of data is generated and the data is in need for processing to be able contribute to the user's requirements. All of the IoT applications require low latency and certain bandwidth to provide us with new features which are Time-Dependent. Thus, it would be advantageous to redesign the current IoT architecture by understanding the unique properties of communications and computational power to support intelligent processing of IoT data and its analysis. The paper is focused on topics such as characteristics of IoT data, trends in IoT network architecture, problems in IoT data analysis and their solutions. Thus, in particular, the paper is a discussion that Edge software-defined computing and CEP (Complex Event Processing) based fog computing as a architecture that promises to support the different requirements of IoT data analytics. And also, the paper is a discussion about how machine learning, deep learning are useful in removing data processing anomalies. The leader node election for data processing at cloud is also a time consuming task, the proposed genetic based Genetic based Leader Election Algorithm (GLEA) algorithm in the paper show promising results for the solution. [ABSTRACT FROM AUTHOR]

The development of stereo matching algorithm is still one of the challenging problems, especially in ill-posed regions. Hence, this article presents a survey on the algorithm frameworks related to the stereo matching algorithm. Based on the early survey that had been conducted, two major frameworks available in current stereo matching algorithm development, they are traditional and artificial intelligence (AI) frameworks. Most of the traditional methods are very low accuracy compared to the AI-based approach. This can be observed in the standard benchmarking dataset, such as from the KITTI and the Middlebury, where AI methods rank at the top of the accuracy list. Additionally, the trend for solving computer vision problems uses AI or machine learning tools that become more apparent in recent years. Thus, this paper is focusing on the survey between the deep learning frameworks, which is one of the machine learning tools related to the convolutional neural network (CNN). Several mixed approaches between CNN based method and traditional handcraft method, as well as the end to end CNN method also discussed in this paper. [ABSTRACT FROM AUTHOR]

Intelligent machines are an important carrier of artificial intelligence, which are widely used not only in the industrial field, but also in various fields of daily life and work. Path planning is an important research direction for studying machine intelligence. The main purpose is to search for a feasible path so that the robot can walk along the searched path in a complex environment. This paper aims to study the evolutionary algorithm of intelligent machines based on artificial intelligence. On the basis of introducing the characteristics of artificial potential field algorithm, the concept and characteristics of genetic algorithm, using genetic algorithm as the framework, the artificial potential field algorithm is combined with the crossover of genetic algorithm. In the operator, the optimization performance of the crossover operator is improved. The research results show that the length of the path obtained by the improved algorithm is shorter and better than that of the genetic algorithm of the single-point crossover operator. [ABSTRACT FROM AUTHOR]

This paper presents a Text similarity detection method based on artificial intelligence and natural language processing. The method combines statistical machine learning and deep learning techniques and designs six models from three perspectives: character-level, word-level, and semantic-level. These models include the diff model based on machine learning, cosine similarity model, Jaccard model, TF-IDF model, as well as the SimCSE and SBERT models based on deep learning methods. To fully leverage the characteristics of these models, three scenarios are designed to calculate the similarity scores based on experience and multiple experimental results. The results show that calculating the similarity scores using these three scenarios not only achieves high accuracy but also requires fewer computational resources. As deep learning and natural language processing technologies continue to advance, Text similarity detection methods based on artificial intelligence and natural language processing will continue to be improved and play a more significant role in practice. Future research can explore more models and algorithms to enhance the accuracy and robustness of plagiarism detection. [ABSTRACT FROM AUTHOR]

The amount of network traffic generated from the Internet of Things (IoT) devices is massive which leads to Big data. In present scenario, IoT devices has taken a leap ahead in terms of growing technology and insecure network traffic flow. The sensitivity of the data in the IoT network is becoming very high, which concern the security in legal and privacy issues. Traditional Intrusion Detection System (IDS) is used as a primary line security to differentiate between the attack and benign network traffic flow. As the size of the network traffic capture increases, it becomes more important to handle such big data. Similarly, when the number of classes and instances increases, the complexity of the data also increases. In this paper, we propose a big data platform based intrusion detection system which can differentiate between the types of network traffic flow generated from the IoT devices. We compared the performance of deep learning algorithms with machine learning algorithms on Apache Spark platform and found that machine learning algorithms outperform deep learning algorithms with greater accuracy and less training time for the model. We evaluated our research based on real world network traffic dataset BoT-IoT [1] for an improved intrusion detection system in IoT network traffic. [ABSTRACT FROM AUTHOR]

Hybrid quantum machine learning (QML) algorithms have potentials for current quantum computing technologies since only part of the model is computed by a quantum device. This paper proposes a novel hybrid quantum-classical deep learning model for cybersecurity application: domain generation algorithms (DGA)-based botnet detection. We analyzed our hybrid model's performance compared with the classical model counterpart to investigate the quantum circuit's effectivity as a layer in a deep learning model. We used four features of the Botnet DGA dataset: MinREBotnets, CharLength, TreeNewFeature, and nGramReputation_Alexa. The hybrid model's quantum circuit is a combination of Pennylane's embedding (Angle Embedding and IQP Embedding) and layers circuit (Basic Entangler Layers, Random Layers, and Strongly Entangling Layers). Also, we implemented noise models for assessing the applicability of the model for current Noisy Intermediate Scale Quantum (NISQ) technology. We found that in some cases, the hybrid model reached high performance (maximum accuracy up to 94,7% using n= 100; 93,9% using n= 1,000). We discovered that the combination of Angle Embedding and Strongly Entangled delivers high accuracy, superior to the classical deep learning model in n= 100 experiments). However, the overall performance is still inferior for the rest cases compared to the classical deep learning model counterpart. [ABSTRACT FROM AUTHOR]

An approach is presented that generalize OCR task including polish letters using deep learning technique. The paper extends EMNIST dataset so that two new classes of polish diacritics "Ą" and "Ć" are attached. Using this new dataset and deep learning technique one can analyze sensitives of standard as well as presented extended approach with different parameters of convolutional neural network. The analysis of the results shows that the shadows and noises that Polish characters, leave with their hooks leads, can be properly recognized and two similar letters "A" and "Ą" can be distinguished by convolutional neural network. On the other hand a neural network trained on dataset without Polish characters do not treat letters "Ą" and "Ć" properly. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Medical Imaging & Radiation Sciences is the property of Elsevier B.V. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

• This paper proposes an unsupervised deep learning model. • This paper proposes an optimized sparse representation method. • This paper proposes a new image fusion algorithm based on unsupervised deep learning-optimized sparse representation. The image fusion method based on deep learning has problems such as the supervised learning of the model, the edge and noise of the fused image, and the setting of the image fusion weight map. To solve these problems, this paper proposes an end-to-end unsupervised deep learning model that performs one-to-many focus image fusion. It solves the training problem encountered by supervised deep learning models while avoiding the unreasonable image fusion weight maps. In addition, this paper proposes an optimized sparse representation method that divides an image into a target area and a background area. Then, it uses super complete dictionary learning to obtain a sparse representation of the image background area. This approach makes the proposed unsupervised deep learning image fusion method robust to noise. Finally, using this method to carry out image fusion experiments, the results show that the quality evaluation indicators of the fused image obtained by this method substantially outperform those of both mainstream machine learning and deep learning image fusion methods. [ABSTRACT FROM AUTHOR]

Abstract The emergence of machine learning algorithms enhanced the effectiveness of satellite image applications. On account of the high variability in-born in satellite data, a large portion of the present classification proposals are not sensible for handling satellite datasets. In this paper, we assessed profound learning systems in view of Convolutional Neural Networks for the precise order of multispectral remote sensing information. The Normalized Difference Vegetation Index (NDVI) is a solitary parameter for detecting landcover by utilizing the red and near-infrared band (NIR) information of the electromagnetic spectrum. NDVI is used to break down remote detecting images and survey the presence or absence of live green vegetation. The experiment is conducted on new publicaly available SAT-4 dataset, where the classes involved are types of vegetation. As NDVI computation require just two band data, it takes the benefit of both RED and NIR band information to classify diverse land covers. In the present work, the AlexNet architecture with two band information along with the reduced number of filters were trained and high-level features obtained from tested model managed to classify different land cover classes in the dataset. The proposed architecture is compared against the benchmark and results are estimated in terms of accuracy, precision and total number of trainable parameters. The proposed design brings about the aggregate diminishment of trainable parameters, while retaining high accuracy and precision, when compared against the existing architecture, which utilizes four bands. [ABSTRACT FROM AUTHOR]

In massive multiple-input multiple-output (MIMO), it is much challenging to obtain accurate channel state information (CSI) after radio frequency (RF) chain reduction due to the high dimensions. With the fast development of machine learning(ML), it is widely acknowledged that ML is an effective method to deal with channel models which are typically unknown and hard to approximate. In this paper, we use the low complexity vector approximate messaging passing (VAMP) algorithm for channel estimation, combined with a deep learning framework for soft threshold shrinkage function training. Furthermore, in order to improve the estimation accuracy of the algorithm for massive MIMO channels, an optimized threshold function is proposed. This function is based on Gaussian mixture (GM) distribution modeling, and the expectation maximum Algorithm (EM Algorithm) is used to recover the channel information in beamspace. This contraction function and deep neural network are improved on the vector approximate messaging algorithm to form a high-precision channel estimation algorithm. Simulation results validate the effectiveness of the proposed network. [ABSTRACT FROM AUTHOR]

SMS spams are one of the concerns and many people do not like to receive them since they are annoying. Many SMS spam detection methods already exist and different classifiers were used, such classifiers depended on Support Vector machine, Naïve Bays and many other machine learning algorithms. In this paper, new classifier is proposed which depends mainly on using H2O as platform to make comparisons between different machine learning algorithms. Moreover, Machine learning algorithms that are used for comparisons are random forest, deep learning and naïve bays. In addition to using deep learning and random forest as classifiers, they are also used to determine the most important features that can be used as input to random forest, deep learning and naïve bays classifiers. Experimental results show that the most significant features that can affect the detection of SMS spam are the number of digits and existing of URL in SMS text. The dataset that is used in experiment is the one proposed by UCI Machine Learning Repositories. Therefore, experiments show that the faster algorithm that achieves high performance is naïve bays with runtime 0.6 seconds, however after comparing it with deep learning and random forest it has the lowest precision, recall, f-measure and accuracy. On the other hand, random forest is the best in term of accuracy with 50 trees and 20 maximum depths, where precision, recall, f-measure and accuracy are 96%, 86%, 91% and 0.977% respectively; nevertheless the runtime is high 30.28 seconds. [ABSTRACT FROM AUTHOR]

• This paper proposes a deep learning algorithm for rapid detection of aortic dissection using contrast-enhanced CT images. • We built a segmentation network that is more suitable for segmenting CT images of the aorta. • This research has laid the foundation for the development of an effective aortic dissection detection algorithm. Aortic dissection (AD) is one of the most common aortic diseases, where blood enters the aortic wall through the aortic intimal rift and causes separation of the arterial wall. Any delay or misdiagnosis can have severe consequences for patients with aortic dissection and even lead to higher mortality rates. Therefore, rapid and accurate detection of aortic dissection saves patients valuable time and provides assistance for the selection of clinical treatment options. This paper describes a deep learning algorithm that uses contrast-enhanced CT images for segmentation and automatic detection of aortic dissection. First, we construct a U-Net based semantic segmentation architecture and apply it to contrast-enhanced CT images to segment the aortic true lumen. Then, we use the segmentation results for aortic circularity analysis to obtain slice-level detection results. Finally, we aggregated the slice-level results to present patient-level detection results. We tested our algorithm on 20 contrast-enhanced CT datasets, of which 10 were aortic dissections. In terms of temporal performance, we have achieved millisecond prediction on sliced images. At the same time, we achieved 85.00% accuracy, 90.00% sensitivity and 80.00% specificity in patient-level testing. [ABSTRACT FROM AUTHOR]

• Analytic comparison between traditional deep learning and shallow machine learning methods for urban remote sensing image segmentation. • Use of machine learning classification methods and SLIC for urban image segmentation. • The combined SLIC algorithm and deep learning methods resulted in an accuracy of 93.20\%, IoU of 0.700, and a variation of 1\% for the difference in the area of tree canopies. • Methodology to automate data extraction ana analysis of a given object using remote sensing data and machine learning. This paper presents a novel approach combining the Simple Linear Iterative Clustering (SLIC) superpixel algorithm with a Convolutional Neural Network (CNN) over high-resolution imagery to detect trees in a typical urban environment of the Brazilian Cerrado biome. Our analysis approach for better results uses the deep learning classifier ResNet-50, with a variation in the batch size and five traditional shallow learning methods. The results were processed and avaliated using mainly accuracy as a metric, but we show that the accuracy poorly represent the overlap between the manual annotation and the resulting map, so we bring the IoU metric results to better show the Network learning classification maps results. The combined SLIC algorithm and the best CNN resulted in an accuracy of 93.20%, IoU of 0.700 and a variation of 1% for difference in the area of tree canopies if compared to our labels, while the best shallow presented an accuracy of 91.70%, IoU of 0.200 and a variation in area of 12.52%. Demonstrating that the proposed CNN method is suitable for segmenting trees from high-resolution images acquired over urban environments. The segmentation with SLIC and CNN can provide very useful results for urban management using low cost RGB images. Such outcomes are of great interest for local managers since reliable maps showing the spatial distribution of trees in urban areas are often required for many applications. [ABSTRACT FROM AUTHOR]

With the developments of smart city, wireless passive localization (WPL) technique that detects targets without carrying any devices draws a lot of research attention. Some machine learning methods, such as sparse coding and deep learning, have been developed to formulate the WPL as a classification problem. Despite these methods have been proved to be effective in WPL, it remains hot topics to design a precise objective function that could detect the position of the targets with high accuracy and robustness. In this paper, we exploited a sparsity regularizer, named l o g -regularizer, in the classification objective function. By virtue of the distinguished ability in measuring sparsity, our proposed improved sparse coding algorithm (ISCA) with l o g -regularizer could position targets accurately with robust performance even in the challenging environments Experimental results show that our proposal achieves better results compared with five other machine learning algorithms. Even the input data is severely polluted by noise (SNR = −10 dB), the proposed method could still obtain high localization accuracy of 99.4%. [ABSTRACT FROM AUTHOR]

• A supervised multinomial Bayesian learning algorithm for breast cancer detection with terahertz (THz) images is proposed. • The model-based learning approach requires much less training data when compared to alternative deep learning methods. • The algorithm represents a promising technique for region detection within THz images of freshly excised cancer samples. This paper proposes a supervised multinomial Bayesian learning algorithm for breast cancer detection using terahertz (THz) imaging of freshly excised murine tumors. The proposed algorithm utilizes a multinomial Bayesian probit regression approach, which establishes the link between THz data and classification results by using two different models, a polynomial regression model and a kernel regression model. Such a model-based learning approach employs only a small number of model parameters, thus it requires much less training data when compared with alternative deep learning methods. The training phase of the algorithm is performed by using the histopathology results of formalin-fixed, paraffin embedded (FFPE) samples as ground truth. There is usually a considerable shape mismatch between the freshly excised sample and its FFPE counterpart due to sample dehydration, and such mismatch negatively impacts the quality of the training data. We propose to address this challenge by using an innovative reliability-based training data selection method, where the reliability of the training data is quantified and estimated by using an unsupervised expectation maximization (EM) classification algorithm with soft probabilistic output. Experiment results demonstrate that the proposed multinomial Bayesian probit regression models with reliability-based training data selection achieve better performance than existing methods. Overall, these results demonstrate that the proposed supervised segmentation models represent a promising technique for the region detection with THz imaging of freshly excised breast cancer samples. [ABSTRACT FROM AUTHOR]

• Performs the pulmonary emphysema diagnosis with binary thresholding and hybrid classification. • Produces a novel hybrid meta-heuristic algorithm called BM-BOA that improves the segmentation. • WLBP is the input for the NN and the WLBP pattern of segmented images are the input. • Experimentally proves the ability of proposed pulmonary emphysema diagnosis. • Proves that the proposed BM-BOA-HC performs better than all other existing methods. This paper is to estimate the potential of a deep learning method for automatic diagnosis of pulmonary emphysema. In the initial step, the dataset acquisition is performed by gathering a set of real-time dataset and the publicly available benchmark datasets known as the Computed Tomography Emphysema Database. After pre-processing of images, the lung segmentation is performed by the optimized binary thresholding. Here, the improvement of the segmentation is accomplished by the adoption of a hybrid meta-heuristic algorithm with Barnacles Mating Optimization (BMO), and Butterfly Optimization Algorithm (BOA) called Barnacles Mating-based Butterfly Optimization Algorithm (BM-BOA), in such a way to attain the multi-objective function concerning the variance and entropy of the image. Further, the feature descriptor called Weber Local Binary Pattern (WLBP) is used for generating the pattern image and the feature vectors. Two types of machine learning algorithms are used for the classification, in which Neural Network (NN) considers the feature vector from WLBP as input, and the deep learning model called Convolutional Neural Network (CNN) considers the WLBP pattern of the segmented image as input. In the hybrid classification model, the activation function is optimized by the same BM-BOA, which results in classifying the normal lung, mild emphysema, moderate (medium) emphysema, and severe emphysema. According to the experimental results with the comparison over the state-of-art-techniques, the proposed system permits inexpensive and reliable identification of emphysema on digital chest radiography. [ABSTRACT FROM AUTHOR]

Industry 4.0 is the new industrial revolution. By connecting every machine and activity through network sensors to the Internet, a huge amount of data is generated. Machine Learning (ML) and Deep Learning (DL) are two subsets of Artificial Intelligence (AI), which are used to evaluate the generated data and produce valuable information about the manufacturing enterprise, while introducing in parallel the Industrial AI (IAI). In this paper, the principles of the Industry 4.0 are highlighted, by giving emphasis to the features, requirements, and challenges behind Industry 4.0. In addition, a new architecture for AIA is presented. Furthermore, the most important ML and DL algorithms used in Industry 4.0 are presented and compiled in detail. Each algorithm is discussed and evaluated in terms of its features, its applications, and its efficiency. Then, we focus on one of the most important Industry 4.0 fields, namely the smart grid, where ML and DL models are presented and analyzed in terms of efficiency and effectiveness in smart grid applications. Lastly, trends and challenges in the field of data analysis in the context of the new Industrial era are highlighted and discussed such as scalability, cybersecurity, and big data. [ABSTRACT FROM AUTHOR]

• Propose a deep learning-based location-to-segmentation method for NPC segmentation. • Location: define recall preserved loss for ResNeXt-50 U-Net to ensure over-segmentation. • Segmentation: focus adaptively on NPC-related 3D volumes in each CT scan. • Achieve Dice of 62.88 ± 8.12% on StructSeg-NPC 2019 training set and mean DSC of 61.81% in online testing. Automated segmentation of Nasopharyngeal carcinoma (NPC) plays a critical role in the radiotherapy or chemo-radiotherapy for this cancer. Despite their improved performance, most deep learning models designed for this segmentation task use either magnetic resonance imaging (MRI) or multimodality data as input. In this paper, we propose a deep learning based algorithm called NPC-Seg for the segmentation of NPC using computed tomography (CT), which is less expensive and more available than MRI. This algorithm uses the location-to-segmentation framework. In the location step, it fine-tunes the pre-trained ResNeXt-50 U-Net with a newly proposed recall preserved loss to roughly segment the gross tumor volume (GTV) of each NPC. In the segmentation step, it fine-tunes the ResNeXt-50 U-Net again, but using the Dice loss, to segment the bounding box region detected in the location step on a patch-by-patch basis. We have evaluated the proposed NPC-Seg algorithm on the StructSeg-NPC dataset. Our algorithm achieves the Dice similarity coefficient (DSC) of 62.88 ± 8.12 % on 50 training data in the ten-fold cross-validation, substantially outperforming three existing deep learning methods, and also achieves an average DSC of 61.81% on the testing dataset in the online validation. [ABSTRACT FROM AUTHOR]