Search

The existing plethora of document classification techniques exploits different data sources either from the content or metadata of research articles. Various journal publishers like Springer, Elsevier, IEEE, etc., do not provide open access to the content of research articles, whereas metadata is freely available there. Metadata like title, keyword, and abstract can serve as a better alternative to the content in various scenarios. In the current literature, researchers have assessed the role of some of the metadata individually. We believe that the collective contribution of metadata parameters can play a significant role in classifying research papers. This paper presents a comprehensive evaluation of the role of metadata, individually as well as in combinations to achieve the objective of research paper classification. Moreover, we have classified the research articles into ACM hierarchy root categories (e.g. general literature, hardware, software, etc.). In this comprehensive evaluation, we have assessed all the possible combinations of metadata features against different classifiers such as Random Forest, K Nearest Neighbor, and Decision Tree. The results of this research reveal that the title & keywords combination outperforms other combinations with an F-measure score of 0.88.

To improve the accuracy of user implicit rating prediction, we combine the traditional latent factor model (LFM) and bidirectional gated recurrent unit neural network (BiGRU) model to propose a hybrid model that deeply mines the latent semantics in the unstructured content of the text and generates a more accurate rating matrix. First, we utilize the user’s historical behavior (favorites records) to build a user rating matrix and decompose the matrix to obtain the latent factor vectors of users and literature. We also apply the BERT model for word embedding of the research papers to obtain the sequence of word vectors. Then, we apply the BiGRU with the user attention mechanism to mine the research paper textual content and to generate the new literature latent feature vectors that are used to replace the original literature latent factor vectors decomposed from the rating matrix. Finally, a new rating matrix is generated to obtain users’ ratings of noninteractive research papers and to generate the recommendation list according to the user latent factor vector. We design experiments on the real datasets and verify that the research paper recommendation model is superior to traditional recommendation models in terms of precision, recall, F1-value, coverage, popularity and diversity.

Most existing studies and public datasets for handwritten Chinese text recognition are based on the regular documents with clean and blank background, lacking research reports for handwritten text recognition on challenging areas such as educational documents and financial bills. In this paper, we focus on examination paper text recognition and construct a challenging dataset named examination paper text (SCUT-EPT) dataset, which contains 50 000 text line images (40 000 for training and 10 000 for testing) selected from the examination papers of 2 986 volunteers. The proposed SCUT-EPT dataset presents numerous novel challenges, including character erasure, text line supplement, character/phrase switching, noised background, nonuniform word size, and unbalanced text length. In our experiments, the current advanced text recognition methods, such as convolutional recurrent neural network (CRNN) exhibits poor performance on the proposed SCUT-EPT dataset, proving the challenge and significance of the dataset. Nevertheless, through visualizing and error analysis, we observe that humans can avoid vast majority of the error predictions, which reveal the limitations and drawbacks of the current methods for handwritten Chinese text recognition (HCTR). Finally, three popular sequence transcription methods, connectionist temporal classification (CTC), attention mechanism, and cascaded attention-CTC are investigated for HCTR problem. It is interesting to observe that although the attention mechanism has been proved to be very effective in English scene text recognition, its performance is far inferior to the CTC method in the case of HCTR with large-scale character set.

For current steganalysis of the image content-adaptive steganography, there are multiple problems to be improved, such as high difficulty and low accuracy, when detecting images with various contents and textures. For this problem, an improved steganalytic method is proposed in this paper based on the pre-classification and feature selection. First, using the features extracted based on the dependency analysis of image adjacent data, the images with various content and texture complexities are pre-classified as multiple clusters by the K-means algorithms. Then, the performance of existing various steganalytic features are analyzed for different clusters of images, and the optimal features for each cluster are selected for final classification. The experimental results show that the detection accuracy could be improved by the proposed method, and the rationality and availability are also verified. At the same time, the analysis and experimental results in this paper also show that the images with rich content and complex texture should be paid more attention both in steganography and in steganalysis.

In the process of digital image acquisition, color inconsistency is a common issue due to the influence of photographic conditions. The inconsistency of color affects the quality of digital orthophoto products and subsequent applications significantly, and this issue remains challenging both in the academia and industry. This paper proposes a color equalization method for aerial image based on block adjustment and an improved color balance model based on the analysis of causes for color deviation. Two nonlinear functions are used to simulate the process of color deviation and correct the color of images. Due to the existence of error, the color corrected images still have minor deviations in the color of the overlapped regions. Therefore, this research introduces an error equation to address this issue and uses the least square principle to calculate the adjustment. The simulation experiments and real experiments are carried out on several groups of aerial image data respectively. The results show that the method not only achieves better visual effect but also has a higher adjustment precision. Finally, the rationality and parameters of the model proposed in this paper are discussed.

Image classification needs to consider image degradations in practice because an image classification network trained with clean images works poorly for degraded images while digital images usually include some degradations such as JPEG compression. To tackle this problem, a common approach is training classification networks on degraded images with various levels of degradation, e.g. various quality factors for JPEG compression. However, the classification networks do not usually have enough accuracy for clean images because the classification networks have been averagely trained for various levels of degradation. This paper aims to construct a classification network of degraded images without sacrificing the performance of clean images. This paper proposes a network to learn the classification of degraded images and degradation levels of degraded images as multi-task learning. Learning the classification of degraded images is the main task of multi-task learning. On the other hand, learning degradation levels is a sub-task of multi-task learning and reinforces the classification ability. In the proposed network, a feature extractor of the classification is trained with image features acquired from a classification network of clean images by using consistency regularization with the cosine similarity loss. The experimental results using different types of degradations, including JPEG compression, Gaussian blur, Gaussian noise, and salt-and-pepper noise, show that our proposed network has enough ability to classify degraded images without sacrificing the performance for clean images.

Distress image retrieval for infrastructure maintenance via self-trained deep metric learning using experts’ knowledge is proposed in this paper. Since engineers take multiple images of a single distress part for inspection of road structures, it is necessary to construct a similar distress image retrieval method considering the input of multiple images to support determination of the level of deterioration. Thus, the construction of an image retrieval method while selecting an effective input from multiple images is described in this paper. The proposed method performs deep metric learning by using a small number of effective images labeled by experts’ knowledge with information about their effectiveness and a large number of unlabeled images via a self-training approach. Specifically, an end-to-end learning approach that performs retraining of the model by assigning pseudo-labels to these unlabeled images according to the output confidence of the model is achieved. Thus, the proposed method can select an effective image from multiple images that are input at the retrieval as a query image. This is the main contribution of this paper. As a result, the proposed method realizes highly accurate retrieval of similar distress images considering the actual situation of inspection in which multiple images of a distress part are input.

A human-centric emotion estimation method based on correlation maximization with consideration of changes with time in visual attention and brain activity when viewing images is proposed in this paper. Owing to the recent developments of many kinds of biological sensors, many researchers have focused on multimodal emotion estimation using both eye gaze data and brain activity data for improving the quality of emotion estimation. In this paper, a novel method that focuses on the following two points is introduced. First, in order to reduce the burden on users, we obtain brain activity data from users only in the training phase by using a projection matrix calculated by canonical correlation analysis (CCA) between gaze-based visual features and brain activity-based features. Next, for considering the changes with time in both visual attention and brain activity, we obtain novel features based on CCA-based projection in each time unit. In order to include these two points, the proposed method analyzes a fourth-order gaze and image tensor for which modes are pixel location, color channel and the changes with time in visual attention. Moreover, in each time unit, the proposed method performs CCA between gaze-based visual features and brain activity-based features to realize human-centric emotion estimation with a high level of accuracy. Experimental results show that accurate human emotion estimation is achieved by using our new human-centric image representation.

Machine reading comprehension is a very challenging task, which aims to determine the answer span based on the given context and question. The newly developed pre-training language model has achieved a series of successes in various natural language understanding tasks with its powerful contextual representation ability. However, these pre-training language models generally lack the downstream processing structure for specific tasks, which limits further performance improvement. In order to solve this problem and deepen the model’s understanding of the question and context, this paper proposes S&I Reader. On the basis of the pre-training model, skimming, intensive reading, and gated mechanism modules are added to simulate the behavior of humans reading text and filtering information. Based on the idea of granular computing, a multi-granularity module for computing context granularity and sequence granularity is added to the model to simulate the behavior of human beings to understand the text from words to sentences, from parts to the whole. Compared with the previous machine reading comprehension model, our model structure is novel. The skimming module and multi-granularity module proposed in this paper are used to solve the problem that the previous model ignores the key information of the text and cannot understand the text with multi granularity. Experiments show that the model proposed in this paper is effective for both Chinese and English datasets. It can better understand the question and context and give a more accurate answer. The performance has made new progress on the basis of the baseline model.

The airway tree is one of the most important part in human respiratory system. Airway segmentation plays a crucial role in pulmonary disease diagnosis, localization and surgical navigation. We propose a novel method to improve airway segmentation in thoracic computed tomography(CT) using deep learning. In order to take into account the multi-scale changes of the airway and achieve accurate airway segmentation, we design an end-to-end Tiny Atrous Convolutional Network (TACNet) based on 3D convolution neural network. In view of the difficulty of classification due to the numerous branches of airway, two evaluation factors, namely, the angle of airway bifurcation and the buffer length of airway bifurcation are designed, which are used for airway classification by combining the centerline extraction. We train TACNet on the inspirator thoracic CT scans with ground truth, which are generated by clinicians and evaluate on our own clinical data sets and EXACT’09 data sets. Compared with state-of-the-art airway segmentation algorithms, proposed algorithm in this paper achieve very competitive results in 20 test datasets of the EXACT’09 challenge. The experimental results show that the algorithm proposed in this paper has high robustness and advantages regardless of airway segmentation or airway classification. In 20 test datasets of the EXACT’09 challenge, the average airway tree length detection rate is the best in the public literatures.

This paper tackles the problem of detecting incongruities between headlines and body text, where a news headline is irrelevant or even in opposition to the information in its body. Our model, called the graph-based hierarchical dual encoder (GHDE), utilizes a graph neural network to efficiently learn the content similarity between news headlines and long body paragraphs. This paper also releases a million-item-scale dataset of incongruity labels that can be used for training. The experimental results show that the proposed graph-based neural network model outperforms previous state-of-the-art models by a substantial margin (5.3%) on the area under the receiver operating characteristic (AUROC) curve. Real-world experiments on recent news articles confirm that the trained model successfully detects headline incongruities. We discuss the implications of these findings for combating infodemics and news fatigue.

In this paper, we propose an efficient attention module for visible and thermal infrared (TIR) matching deep learning networks. This method judges the right or wrong of heterogeneous sensor matching through the proposed deep learning model and increases the matching rate through the attention module using the edge-utilizing structure. This paper contributes to three aspects: 1) The first aspect is Convolutional Neural Network (CNN) structure comparisons for heterogeneous sensor registration. We consider the matching network as a classification problem when stacked heterogeneous sensor data become input of a single CNN network. Therefore, this paper shows result that is related with not only the network designed for heterogeneous sensor matching, but also various deep learning networks used for classification. 2) the second is a consideration for efficient attention module. The experiments show the module can replace lots of convolution blocks and the results achieve more better performance. The attention module uses a $1\times $ k filter and a k $\times $ 1 filter to extract horizontal and vertical edges and convolves two paths using them. 3) The third is suitable deep learning model for aerial complex visible and TIR data registration. To compare the various methods, we describe the calibration process of aerial visible and TIR data obtained directly from a drone. By using the calibrated data, this paper presents an AVIR attention block-based architecture that shows optimal matching results with minimal addition of parameters.

Music is a heavenly way of expressing feelings about the world. The language of music has vast diversity. For centuries, people have indulged in debates to stratisfy between Western and Indian Classical Music. But through this paper, an understanding can be fabricated while differentiating the types of Indian Classical Music. Classical music is one of the essential characteristics of Indian Cultural Heritage. Indian Classical Music is divided into two major parts, i.e. Hindustani and Carnatic. Models have been sculptured and trained to classify between Hindustani and Carnatic Music. In this paper, two approaches are used to implement classification models. MFCCs are used as features and implemented models like DNN (1 Layer, 2 Layers, 3 Layers), CNN (1 Layer, 2 Layers, 3 Layers), RNN-LSTM, SVM (Sigmoid, Polynomial & Gaussian Kernel) as one approach. A 3 channels input is created by merging features like MFCC, Spectrogram and Scalogram and implemented models like VGG-16, CNN (1 Layer, 2 Layers, 3 Layers), ResNet-50 as another approach. 3 Layered CNN and RNN-LSTM model performed best among all the approaches.

An intrusion detection system (IDS) is an important protection instrument for detecting complex network attacks. Various machine learning (ML) or deep learning (DL) algorithms have been proposed for implementing anomaly-based IDS (AIDS). Our review of the AIDS literature identifies some issues in related work, including the randomness of the selected algorithms, parameters, and testing criteria, the application of old datasets, or shallow analyses and validation of the results. This paper comprehensively reviews previous studies on AIDS by using a set of criteria with different datasets and types of attacks to set benchmarking outcomes that can reveal the suitable AIDS algorithms, parameters, and testing criteria. Specifically, this paper applies 10 popular supervised and unsupervised ML algorithms for identifying effective and efficient ML-AIDS of networks and computers. These supervised ML algorithms include the artificial neural network (ANN), decision tree (DT), k-nearest neighbor (k-NN), naive Bayes (NB), random forest (RF), support vector machine (SVM), and convolutional neural network (CNN) algorithms, whereas the unsupervised ML algorithms include the expectation-maximization (EM), k-means, and self-organizing maps (SOM) algorithms. Several models of these algorithms are introduced, and the turning and training parameters of each algorithm are examined to achieve an optimal classifier evaluation. Unlike previous studies, this study evaluates the performance of AIDS by measuring the true positive and negative rates, accuracy, precision, recall, and F-Score of 31 ML-AIDS models. The training and testing time for ML-AIDS models are also considered in measuring their performance efficiency given that time complexity is an important factor in AIDSs. The ML-AIDS models are tested by using a recent and highly unbalanced multiclass CICIDS2017 dataset that involves real-world network attacks. In general, the k-NN-AIDS, DT-AIDS, and NB-AIDS models obtain the best results and show a greater capability in detecting web attacks compared with other models that demonstrate irregular and inferior results.

In this paper, we show that adversarial training time attacks by a few pixel modifications can cause undesirable overfitting in neural networks for both discriminative and generative models. We propose an evolutionary algorithm to search for an optimal pixel attack using a novel cost function inspired by domain adaptation literature to design our training time attack. The proposed cost function explicitly maximizes the generalization gap and domain divergence between clean and corrupted images. Empirical evaluations demonstrate that our adversarial training attack can achieve significantly low testing accuracy (with high training accuracy) on multiple datasets by just perturbing a single pixel in the training images. Even under the use of popular regularization techniques, we identify a significant performance drop compared to clean data training. Our attack is more successful than previous pixel-based training time attacks on state-of-the-art Convolutional Neural Networks (CNNs) architectures, as evidenced by significantly lower testing accuracy. Interestingly, we find that the choice of optimization plays an essential role in robustness against our attack. We empirically observe that Stochastic Gradient Descent (SGD) is resilient to the proposed adversarial training attack, different from adaptive optimization techniques such as the popular Adam optimizer. We identify that such vulnerabilities are caused due to over-reliance on the cross-entropy (CE) loss on highly predictive features. Therefore, we propose a robust loss function that maximizes the mutual information between latent features and input images, in addition to optimizing the CE loss. Finally, we show that the discriminator in Generative Adversarial Networks (GANs) can also be attacked by our proposed training time attack resulting in poor generative performance. Our paper is one of the first works to design attacks for generative models.

Increasing human activities have caused serious disturbance to global forest resources, so how to accurately identify individual trees has become an important task of forest resources investigation. In order to get the accurate number of individual trees, this paper took coniferous forest and mixed coniferous and broad-leaved forest as experimental samples, as well as Digital Orthophoto Map and airborne LiDAR Point Cloud as research data. We propose a deep Learning individual tree segmentation method based on RetinaNet model and PCS algorithm by doing comparative analysis (classical Watershed Algorithm and Layer Stacking Algorithm) at the plots (with high, medium, and low densities). The experimental results show that the method proposed in this paper can solve the problem of individual tree segmentation in high density forest and improve its degree of automation. Compared with Watershed algorithm and Layer stacking algorithm, F-Measure is improved by 6%-29% and 7%-20%, respectively. In other words, the results of individual tree segmentation presented in this paper can not only improve the precision of individual tree segmentation, but also maintain a high detection rate, which can meet the accuracy and high efficiency of individual tree extraction, so as to meet the needs for modern forestry investigation.

The key to an accurate understanding of terrain is to extract the informative features from the multi-modal data obtained from different devices. Sensors, such as RGB cameras, depth sensors, vibration sensors, and microphones, are used as the multi-modal data. Many studies have explored ways to use them, especially in the robotics field. Some papers have successfully introduced single-modal or multi-modal methods. However, in practice, robots can be faced with extreme conditions; microphones do not work well in crowded scenes, and an RGB camera cannot capture terrains well in the dark. In this paper, we present a novel framework using the multi-modal variational autoencoder and the Gaussian mixture model clustering algorithm on image data and audio data for terrain type clustering by forcing the features to be closer together in the feature space. Our method enables the terrain type clustering even if one of the modalities (either image or audio) is missing at the test-time. We evaluated the clustering accuracy with a conventional multi-modal terrain type clustering method and we conducted ablation studies to show the effectiveness of our approach.

As a basic task in the field of computer vision, target detection has been concerned by many researchers. The performance of target detection method is also directly related to the research in many advanced semantic fields. Current general target detection methods are not effective in small target detection, so this paper studies the problem of small target detection and proposes a small target detection method based on deep learning with considerate feature and effectively expanded sample size. Firstly, according to the characteristics of convolutional neural network, we improve the current deep network architecture which performs well in target detection, and introduce considerate multi-feature and multi-scale detection. Then, we transform the high-resolution images obtained on the Internet by combining two groups of sampling method, so that the feature distribution of the high-resolution target is closer to that of the low-resolution target, thus effectively expanding the training data set, solving the problem that small target data is difficult to be labeled and effectively avoiding overfitting. The results show the effectiveness of the improved method in small target detection. In addition, in view of the shortage of small target detection review literature, this paper gives a comprehensive and detailed introduction to the field of small target detection in terms of related work and future work.

Deep convolutional neural networks (DCNNs) have attracted extensive attention in synthetic aperture radar (SAR) image target detection. However, there is usually a lack of labeled data in SAR images which results in difficulties to improve the detection performance with DCNN based methods. Transfer learning borrows generic knowledge from other tasks for the target task and can be an effective solution to the small sample problem, but the features’ generality and specificity across various models are not discussed in target detection. In addition, conventional DCNN based models have too many parameters related to the prior information of the targets which increase the difficulty of transfer learning between different tasks. In this paper, we mainly conduct research in three folds to address these mentioned issues. First, the generality and specificity of features in different network models are analyzed in the field of SAR target detection. Second, a transitive transfer framework is developed for detection under small sample condition. Third, an anchor free rotatable detector with flexible structure is designed for SAR images detection to suit the transfer detection framework proposed in this paper and a reasonable explanation is given for its upper hand with small samples. Experimental results demonstrate that: 1) transitive transfer framework is valid in SAR target detection; 2) the anchor free detector in this framework is superior to the anchor based one; 3) the proposed detector achieves better performance under small sample condition by comparison with state-of-the-art methods.

This paper focuses on how to represent the testing face images for multi-feature face recognition. The choice of feature is critical for face recognition. The different features of the sample contribute differently to face recognition. The joint similar and specific learning (JSSL) has been effectively applied in multi-feature face recognition. In the JSSL, although the representation coefficient is divided into the similar coefficient and the specific coefficient, there is the disadvantage that the training images cannot represent the testing images well, because there are probable expressions, illuminations and disguises in the testing images. We think that the intra-class variations of one person can be linearly represented by those of other people. In order to solve well the disadvantage of JSSL, in the paper, we extend JSSL and propose the extended joint similar and specific learning (EJSSL) for multi-feature face recognition. EJSSL constructs the intra-class variant dictionary to represent the probable variation between the training images and the testing images. EJSSL uses the training images and the intra-class variant dictionary to effectively represent the testing images. The proposed EJSSL method is perfectly experimented on some available face databases, and its performance is superior to many current face recognition methods.

With the rapid development of remote-sensing imaging technology, remote-sensing images have become increasingly diverse, and people are paying more attention to ocean remote-sensing research. Because ocean remote-sensing data are complex, and the ocean environment is diverse, results will differ, even if the same target is detected at different times in the same scene. To obtain more semantic features and better pixel-level prediction capabilities, this paper proposes a pixel-level ocean remote-sensing image algorithm (GLPO-Net) that combines local and global features. First, texture features, color features, and spatial relationship features are extracted. Second, the algorithm constructs a multiscale local cross-attention mechanism strategy to obtain feature weight information in different directions to fully mine the local features of ocean remote-sensing images. Concurrently, an algorithm constructs a multiscale global cross-attention mechanism strategy to obtain global features. Then, the fusion of global features and local features is described in each submodule to obtain more representative deep features. Finally, small-sample ocean remote-sensing is described via image pixel-level prediction. The algorithm proposed in this paper has been tested with three public ocean remote-sensing datasets. The experimental results show that the proposed GLPO-Net algorithm can learn features from small samples of ocean remote-sensing images. Compared to the prediction results of other remote-sensing image algorithms, GLPO-Net exhibits better prediction capabilities.

Near-infrared spectroscopy (NIR) has become one of the most important methods for counterfeit drugs identification for its low cost, non-destructive, and on-site detection. However, it is often invalid for unknown samples beyond the scope of modeling samples, as well as it is not efficient in conditions of insufficient samples (insufficient number of samples within the class), unbalanced samples (large difference in the number of samples between classes), and sensitivity of identification results (different tolerance for different errors in application scenarios). To solve these problems, this paper proposes a general method for on-site identification of counterfeit drugs based on Siamese-network modeling with near-infrared spectroscopy, which especially constructs the train set and test set, learning the general knowledge of spectral differences to identify the different drugs by a costumed one-dimensional convolution neural network (1D-CNN), and finally answered the question of whether the on-site two spectra are pointed to the same drug. Based on experimental modeling samples of 1314 spectra, which are involved 9 drugs produced by 25 manufacturers, this paper has constructed and fully trained its model. Then, not known at the time of modeling, 4 drugs produced by 9 manufacturers are used for testing in the on-site application, and the accuracy rate amounts to 97.3%. For generalizing consideration, randomly divided into training and testing categories, the 32015 spectra of 135 drugs produced by 391 manufacturers in the spectral library are handled by the same processing. The generalization model is equally applicable, and the accuracy is above 97%. Compared with traditional binary classification identification methods such as SVM, PLS-DA, Auto-encoding, and one class (OC) threshold identification algorithms such as SVM-OC, SIMCA, conformity test (CT), the proposed method has the best identification ability for unknown samples in modeling.

We address the problem of depth estimation from a single monocular image in the paper. Depth estimation from a single image is an ill-posed and inherently ambiguous problem. In the paper, we propose an encoder-decoder structure with the feature pyramid to predict the depth map from a single RGB image. More specifically, the feature pyramid is used to detect objects of different scales in the image. The encoder structure aims to extract the most representative information from the original image through a series of convolution operations and to reduce the resolution of the input image. We adopt Res2-50 as the encoder to extract important features. The decoder section uses a novel upsampling structure to improve the output resolution. Then, we also propose a novel loss function that adds gradient loss and surface normal loss to the depth loss, which can predict not only the global depth but also the depth of fuzzy edges and small objects. Additionally, we use Adam as our optimization function to optimize our network and speed up convergence. Our extensive experimental evaluation proves the efficiency and effectiveness of the method, which is competitive with previous methods on the Make3D dataset and outperforms state-of-the-art methods on the NYU Depth v2 dataset.

Dementia diseases are increasing rapidly, according to the World Health Organization (WHO), becoming an alarming problem for the health sector. The electroencephalogram (EEG) is a non-invasive test that records brain electrical activity and has a wide field of applications in the medical area, one of which is the detection of neurodegenerative diseases. The aim of this work is to present the results of a thorough review of the use of EEG systems for the detection of dementia diseases. Around 82 papers published between 2009 and 2020 were reviewed and compared obtaining data such as sampling time, number of electrodes, the most popular processing, classification, and validation techniques, as well as an analysis of the reported results. The relationship of the selected parameters with the efficiency obtained is shown. Some more common combinations in the reviewed papers that demonstrated to have reliability levels greater than 90%, and details to be considered at each stage of the process. An overview of the most commonly used classification tools and processing techniques is also described.

The Internet of Things is based on a communication network to automatically receive various information from the natural world. It uses intelligent objects with perception, communication and computers to automatically receive various information from nature. All independently managed physical objects are connected to each other to achieve integrated perception, reliable transmission and intelligent processing, and to establish intelligent information between people and objects, and objects and object service systems. This article mainly introduces the impact of industrial agglomeration on the regional economy in a simulated intelligent environment based on machine learning. This paper proposes a method to detect industrial agglomeration index to analyze industrial agglomeration. Through the establishment of the industrial agglomeration index system, the level of integration of the manufacturing industry in our city was objectively analyzed, and the impact of the integration of manufacturing industry on our city was tested empirically. Finally, the relationship between industrial integration and regional economic development was tested. The experimental results in this paper show that industrial integration in a smart environment based on simulation learning has a significant positive impact on regional economic development. Among them, the level of cooperation between the manufacturing service industry and the manufacturing industry increased by 1%, and the level of regional economic development increased by 0.025%.

The attention-based SeriesNet (A-SeriesNet) combined augmented attention residual learning module-based convolutional neural network (augmented ARLM-CNN) subnetwork with hidden state attention module-based recurrent neural network (HSAM-RNN) subnetwork for conditional time series prediction with high accuracy. The augmented ARLM-CNN subnetwork has defects in extracting latent features of the multi-condition series. The forecasting accuracy will decrease when the feature dimension of the multi-condition series becomes high. The same problem also occurs in the HSAM-RNN subnetwork of A-SeriesNet. The dual-stage attention recurrent neural network (DA-RNN) proved that the attention-based encoder-decoder framework is an effective model for dealing with the above problem. This paper applies the DA-RNN to the HSAM-RNN subnetwork of A-SeriesNet and presents the triple-stage attention-based recurrent neural network (TA-RNN) subnetworks. Furthermore, this paper considers a CNN-based encoder-decoder structure named dual attention residual learning module-based convolutional neural network (DARLM-CNN) subnetwork to improve the augmented ARLM-CNN subnetwork of A-SeriesNet. Finally, this paper presents the triple-stage attention-based SeriesNet (TA-SeriesNet), which uses a new concatenation method instead of the element-wise multiplication of A-SeriesNet to parallel connect the proposed subnetworks and reduce the dependence of forecasting results on a certain subnetwork. The experimental results show our TA-SeriesNet is superior to other deep learning models in forecasting accuracy evaluation metrics for high feature dimensional time series datasets.

The rapid development of tools for communication such as social networks, tweeting and Whatsapp has generated a large mass of important textual data. Also, the COVID-19 pandemic has inflamed social networks, hence the automatic analysis of opinions has become paramount. The purpose of this paper is to analyze Arabic tweets in terms of positivity, negativity, or neutrality.In analyzing the opinions of the Arabic language, a real challenge is encountered, which lies in the use of different dialects (Egyptian, Saudian, Maghrebian, Gulfian, Levantine, Syrian $\ldots $ ). In this paper, we introduce two major components: The first employs six machine learning (ML) methods, including Decision Trees (DT), Logistic Regression (LR), k Nearest Neighbors (K-NN), Random Forests (RF), Support Vector Machines (SVM), and Nave Bayes (NB), with the TF-IDF method acting as the feature extraction.While, the second part consists of testing three variants of Deep Learning (DL) based on multiplicative Long Short Term Memory (mLSTM), Long Short Term Memory (LSTM), and Gated Recurrent Unit (GRU) by applying word embedding as the input vector. The experimental study was validated using three Arabic language corpora (TEAD, ATSAD, and ASTD) and two learning modes (Hold out and 10-folds cross validation). The obtained results in terms of Accuracy (ACC), Precesion (PREC), Recall (REC), and F1-score (F1) show a clear performance for DL techniques based on a 10-folds strategy compared to the state-of-the-art. The experiments shown in the paper reveal that the proposed DL models accomplished the best results.

Technology forecasting not only helps business managers to make the right decisions but also helps researchers to grasp the direction of technology development. Technology forecasting, which facilitates the identification of the development technologies with high potential, can be an effective tool to support the management and plan for the future research activities. For this purpose, this paper firstly constructs Multi-modal input based on deep learning (MIDL) text classification model to extract relevant SCI papers from Web of Science database from 1996 to 2019 for topic classification, and then apply the Ensemble Empirical Mode Decomposition (EEMD) and Long Short-Term Memory (LSTM) neural networks to build an EEMD-LSTM technology forecasting model to predict the future development of each research field. Besides, we verify the validity of the method by taking robotics as an example in this paper. The results show that the accuracy, recall and F1 of MIDL text classification model are 0.826, 0.822 and 0.824 respectively. As compared with the optimal results of other classification models, the accuracy, recall and F1 are improved by 4.1%, 3.5%, and 3.8% respectively. The mean MAPE of the EEMD-LSTM model is 7%, which is 11% lower than ARIMA, 8% lower than LSTM, and 10% lower than 2-layer-LSTM.

In this paper, we propose a multi-branch neural network architecture named Coupled Modular Neural Network (CMNN). A CMNN is a network consisting of $\beta $ closely coupled sub-networks, where $\beta $ is termed as the branching factor in this paper. We call the whole network a super-graph and each sub-network a sub-graph. Each sub-graph is a stand-alone neural network and shares a common block with other sub-graphs. To effectively leverage the super-graph we propose a simple but easy-to-implement Round-Robin-based learning algorithm. Each training iteration contains two phases. In the first phase, we choose a sub-graph in a Round-Robin fashion and train it using knowledge of the super-graph (distillation). In the second phase, we fine-tune the super-graph based on the updated sub-graphs. This algorithm produces a different copy of the super-graph at each iteration which acts as an improved teacher network for the sub-graph; and a different copy of one of the sub-graphs which functions as a new building block for the super-graph. To validate and test CMNN and the proposed algorithm, we conduct experiments on CIFAR-10, CIFAR-100, Tiny ImageNet and a private On-Road-Risk (ORR) datasets. Empirical results on all these four datasets indicate that we not only obtain a strong sub-graph network, the learning framework can also produce strong ensemble performance which substantiates the diversity introduced throughout the learning framework.

Autonomous driving (AD) provides a reliable solution for safe driving by replacing human drivers responsible for the majority of accidents. The emergence of Machine Learning, specifically Deep Reinforcement Learning (DRL), and its ability to solve complex games proved its potential to address AD challenges. However, model-free methods still suffer from safety-related issues that can be resolved using safe-DRL approaches. The addition of model-based safety filters to the learning-based algorithms provides safety bounds on their performance and constraint satisfaction. In this paper, we investigate the addition of a safety filter based on Model Predictive Control and show an increase in mean testing episode reward by 110% from −75 mean episode reward during testing for 50 episodes for Deep Deterministic Policy Gradient ( $DDPG$ ) to 7.758. We study the impacts of safety filters (7.758 mean reward), heuristic rules, bounded additive noises (0.49% performance increase comparing to noise-free case), and exploration (3.425 mean reward) on the learning algorithm. We compare the effects of filters in the context of simulated exploration and bounded exploration and prove that bounded exploration results in 9.86% increase in mean reward and 12.95% decrease in std comparing to the other method. Additionally, inspired by Deep Internal Learning and biological mechanisms like brain plasticity, we investigate the idea of using each sample for training only once instead of utilizing stochastic batches which increases the mean testing accumulated reward by 1.87% and leads to the best performance (7.942 mean reward and 0.048 std). Finally, the results demonstrate better automotive results for our proposed method than DDPG. Our proposed method, DDPG with safety filter in bounded exploration and adaptive learning under noisy input conditions, has a success rate of 100% under different traffic densities for the simulation environment used in this paper and our assumptions. The proposed method’s automotive results are shown for a braking scenario to avoid collision with other road users.

Choosing optimal Deep Learning (DL) architecture and hyperparameters for a particular problem is still not a trivial task among researchers. The most common approach relies on popular architectures proven to work on specific problem domains led on the same experiment environment and setup. However, this limits the opportunity to choose or invent novel DL networks that could lead to better results. This paper proposes a novel approach for providing general recommendations of an appropriate DL architecture and its hyperparameters based on different configurations presented in thousands of published research papers that examine various problem domains. This architecture can further serve as a starting point of investigating DL architecture for a concrete data set. Natural language processing (NLP) methods are used to create structured data from unstructured scientific papers upon which intelligent models are learned to propose optimal DL architecture, layer type, and activation functions. The advantage of the proposed methodology is multifold. The first is the ability to eventually use the knowledge and experience from thousands of DL papers published through the years. The second is the contribution to the forthcoming novel researches by aiding the process of choosing optimal DL setup based on the particular problem to be analyzed. The third advantage is the scalability and flexibility of the model, meaning that it can be easily retrained as new papers are published in the future, and therefore to be constantly improved.

Anomaly detection has become an indispensable tool for modern society, applied in a wide range of applications, from detecting fraudulent transactions to malignant brain tumours. Over time, many anomaly detection techniques have been introduced. However, in general, they all suffer from the same problem: a lack of data that represents anomalous behaviour. As anomalous behaviour is usually costly (or dangerous) for a system, it is difficult to gather enough data that represents such behaviour. This, in turn, makes it difficult to develop and evaluate anomaly detection techniques. Recently, generative adversarial networks (GANs) have attracted a great deal of attention in anomaly detection research, due to their unique ability to generate new data. In this paper, we present a systematic literature review of the applications of GANs in anomaly detection, covering 128 papers on the subject. The goal of this review paper is to analyze and summarize: (1) which anomaly detection techniques can benefit from certain types of GANs, and how, (2) in which application domains GAN-assisted anomaly detection techniques have been applied, and (3) which datasets and performance metrics have been used to evaluate these techniques. Our study helps researchers and practitioners to find the most suitable GAN-assisted anomaly detection technique for their application. In addition, we present a research roadmap for future studies in this area., Comment: 34 pages,5 figures

Accurate electricity consumption forecasting in the power grids ensures efficient generation and distribution of electricity. Keeping this in mind, the paper introduces a novel deep learning model, termed Gated-FCN, for short-term load forecasting. The key idea is to introduce an automated feature selection and deep learning model for forecasting. The model includes an eight-layered Fully Convolutional Network (FCN-8) in which the hand-crafted feature selection that requires expert domain knowledge is avoided. Furthermore, the model also reduces noise as it learns internal dependencies and the correlation of the time series. Enhanced Bidirectional Gated Recurrent Unit (EBiGRU) is used in combination with FCN-8 to learn long-term temporal dependencies of the time series. Moreover, a weighted averaging mechanism of multiple snapshot models is adopted in the proposed model to assign optimized weights to BiGRU. At the end of FCN-8 and BiGRU, a fully connected dense layer is used that gives final prediction results. Gated-FCN is an end-to-end forecasting model that does not require any other model for enhancing its forecasting efficiency. Different activation functions are initially analyzed to determine how the proposed model learns complex patterns from the time series data. Later, the activation function having the best accuracy is used for forecasting. The proposed model extracts both spatial and temporal features from the data. Furthermore, this paper also provides predictive and exploratory data analyses to assist policymakers in making optimal decisions regarding power production and dispatch. In order to demonstrate the applicability of the proposed technique, the simulations are performed using nine years’ load consumption data taken from Independent System Operators New England (ISO-NE). The comparison with five state-of-the-art techniques is also provided to prove the fact that Gated-FCN gives the best forecasting accuracy as compared to other benchmark techniques in terms of two performance metrics: Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE).

Many studies showed the feasibility of detecting Freezing of Gait (FOG) of Parkinson’s patients by using several numbers of inertial sensors worn on the body and in back-end computing power. This work uses machine learning approaches for analyzing the data of one single body-worn inertial sensor system to classify and detect FOG. Long-Short-Term-Memory (LSTM) is employed as the FOG detection algorithm and the Daphnet (FOG and normal gait) dataset provides the data for model training and testing in this paper. The model considers raw data from three channels of the acceleration sensor mounted on the patient’s shank and ignores all other data from other sensors. The model is patient dependent and uses sensitivity and specificity metrics to evaluate the model’s performance. In this paper, we propose a novel padding method that is applied to the windows of FOG and non-FOG with zero overlaps on the training set and adapts the padding to the individual regions. This method produces windows of only one type of data and label. The proposed padding method reduces the padding amount by two orders of magnitude compared to bigger batch sizes in the sequence splitting method offered by MATLAB 2019a. The padding amount is independent of the batch size. Raw data is fed to the model in the testing mode without any pre-processing or data transformation. The standard rolling window generates fixed-size windows for the test set without overlap and the higher amount of FOG or Normal walking data which defines the label of the individual window. The model for one-second long windows applied in this work outperformed the literature results with a sensitivity of 92.57% and a specificity of 95.62% compared to 82% and 94% reported by Masiala et al.

This paper discusses impact of recovering missing Electroencephalography (EEG) data on classification accuracy of hand movements using tensor-based methods. Improvement in classification accuracy is important for efficient performance of prosthesis. Classification accuracy relies on quality of the observed data. In practice, observed data is usually incomplete because of disconnection of electrodes and other artefacts, which negatively affect the classification accuracy. In this paper, we employ tensor-based imputation methods (canonical/polyadiac decomposition (CPD), weighted optimization version of CPD (CP-WOPT and Nonnegative Matrix Factorization (NMF)) to recover missing data in EEG signals and apply various classifiers to classify hand movements on recovered data. In particular, structured missing data was considered because that is how data gets missed in real-life data acquisition. Percentage of missing data was changed from 10% to 50%. Classifiers are applied on complete, missing and recovered data explicitly to test the performance of our framework. Results show that mean classification accuracy on complete data, missing data and recovered data was 71%, 53% and 64% respectively which shows applicability of our framework.

Laparoscopic surgery is a surgical procedure performed by inserting narrow tubes into the abdomen without making large incisions in the skin. It is done with the aid of a video camera. Laparoscopic videos are affected by various distortions during surgery which lead to loss of visual quality. Identification of these distortions is the primary requisite in automated video enhancement systems used to classify the distortions correctly and accordingly select the proper algorithm to enhance video quality. In addition to high accuracy, the speed of distortion classification should be high, and the system must consider real-time conditions. This paper aims to address the issues faced by similar methods by developing a fast and accurate deep learning model for distortion classification. The dataset proposed by the ICIP2020 conference challenge was used for training and evaluation of the proposed method. This challenging dataset contains videos that have five types of distortions such as noise, smoke, uneven illumination, defocus blur, and motion blur with four levels of intensity. This paper discusses the proposed solution which received the first prize in the ICIP2020 challenge. The solution utilized a transfer learning approach to transfer representation from the domain of natural images to the domain of laparoscopic videos. We used a pre-trained ResNet50 convolutional neural network (CNN) to extract informative features that were mapped by support vector machine (SVM) classifiers to various distortion categories. In this work, the problem of multiple distortions in the same video was formulated as a multi-label distortion classification problem. The approach of transfer learning with decision fusion was applied and was found to outperform other solutions in terms of accuracy (83%), F1 score of a single distortion (94.7%), and F1 score of single and multiple distortions (94.9%). In addition, the proposed solution can run in real time with an inference speed of 20 frames per second (FPS).

The new coronavirus, which has become a global pandemic, has confirmed more than 88 million cases worldwide since the first case was recorded in December 2019, causing over 1.9 million deaths. Since COIVD-19 lesions have clear imaging features on CT images, it is suitable for the auxiliary diagnosis and treatment of COVID-19. Deep learning can be used to segment the lesions areas of COVID-19 in CT images to help monitor the epidemic situation. In this paper, we propose a multi-point supervision network (MPS-Net) for segmentation of COVID-19 lung infection CT image lesions to solve the problem of a variety of lesion shapes and areas. A multi-scale feature extraction structure, a sieve connection structure (SC), a multi-scale input structure and a multi-point supervised training structure were implemented into MPS-Net. In order to increase the ability to segment various lesion areas of different sizes, the multi-scale feature extraction structure and the sieve connection structure will use different sizes of receptive fields to extract feature maps of various scales. The multi-scale input structure is used to minimize the edge loss caused by the convolution process. In order to improve the accuracy of segmentation, we propose a multi-point supervision training structure to extract supervision signals from different up-sampling points on the network. Experimental results showed that the dice similarity coefficient (DSC), sensitivity, specificity and IOU of the segmentation results of our model are 0.8325, 0.8406, 09988 and 0.742, respectively. The experimental results demonstrated that the network proposed in this paper can effectively segment COVID-19 infection on CT images. It can be used to assist the diagnosis and treatment of new coronary pneumonia.

In recent years, a variety of supervised manifold learning techniques have been proposed to outperform their unsupervised alternative versions in terms of classification accuracy and data structure capturing. Some dissimilarity measures have been used in these techniques to guide the dimensionality reduction process. Their good performance was empirically demonstrated; however, the relevant analysis is still missing. This paper contributes to a theoretical analysis on a) how dissimilarity measures affect maintaining manifold neighbourhood structure and b) how supervised manifold learning techniques could contribute to the reduction of classification error. This paper also provides a cross-comparison between supervised and unsupervised manifold learning approaches in terms of structure capturing using Kendall’s Tau coefficients and co-ranking matrices. Four different metrics (including three dissimilarity measures and Euclidean distance) have been considered along with manifold learning methods such as Isomap, ${t}$ -Stochastic Neighbour Embedding ( ${t}$ -SNE), and Laplacian Eigenmaps (LE), in two datasets: Breast Cancer and Swiss-Roll. This paper concludes that although the dissimilarity measures used in the manifold learning techniques can reduce classification error, they do not learn well or preserve the structure of the hidden manifold in the high dimensional space, but instead, they destroy the structure of the data. Based on the findings of this paper, it is advisable to use supervised manifold learning techniques as a pre-processing step in classification. In addition, it is not advisable to apply supervised manifold learning for visualization purposes since the two-dimensional representation using supervised manifold learning does not improve the preservation of data structure.

Intra prediction is the key technology to reduce spatial redundancy in the modern video coding standard. Recently, deep learning based methods that directly generate the intra prediction by neural network achieve superior performance than traditional directional based intra prediction. However, these methods lack the ability to handle complex blocks which contain mixed directional textures or recurrent patterns since they only use the neighboring reference samples of the current one. The other intermediate information denoted as reference priors in this paper generated during the coding process is not exploited. In this paper, a Current Frame Priors assisted Neural Network (CFPNN) is presented to improve the intra prediction efficiency. Specifically, we utilize the local contextual information provided by the neighboring multiple references as the primary inference source. In addition to the neighboring references, we additionally use the other two reference priors within the current frame – the predictor searched by intra block copy (IntraBC) and the corresponding residual component. The IntraBC predictor provides useful nonlocal information to help generate more accurate prediction for complex blocks together with neighboring local information. While the residual component contains unique information that reflects the characteristics of the block to some extent is utilized to reduce the noise contained in the reconstructed reference samples. Moreover, we investigate the best way to integrate the proposed method into the codec. Experimental results demonstrate that compared to HEVC, our proposed CFPNN achieves an average of 4.1% BD-rate reduction for the luma component under the All Intra configuration.

Subjective paper evaluation is a tricky and tiresome task to do by manual labor. Insufficient understanding and acceptance of data are crucial challenges while analyzing subjective papers using Artificial Intelligence (AI). Several attempts have been made to score students’ answers using computer science. However, most of the work uses traditional counts or specific words to achieve this task. Furthermore, there is a lack of curated data sets as well. This paper proposes a novel approach that utilizes various machine learning, natural language processing techniques, and tools such as Wordnet, Word2vec, word mover’s distance (WMD), cosine similarity, multinomial naive bayes (MNB), and term frequency-inverse document frequency (TF-IDF) to evaluate descriptive answers automatically. Solution statements and keywords are used to evaluate answers, and a machine learning model is trained to predict the grades of answers. Results show that WMD performs better than cosine similarity overall. With enough training, the machine learning model could be used as a standalone as well. Experimentation produces an accuracy of 88% without the MNB model. The error rate is further reduced by 1.3% using MNB.

Generally, a few-shot distribution shift will lead to a poor generalization. Furthermore, while the number of instances of each class in the real world may significantly different, the existing few-shot classification methods are based on the assumption that the number of samples in each class is equal, which causes the trained classifier invalid. Moreover, through ResNet and WRN (Wide Residual Network) have achieved great success in the image processing field, the depth and width of CNNs constrain the conventional convolution layer performance. Thus, to overcome the above problems, the model of this paper proposes a novel few-shot classification model that uses learning balance variables to decide how much to learn from the imbalance dataset, which dynamically generates the convolution kernel based on each input. In our model, to extend the decision boundaries and enhance the class representations, this paper uses embedding propagation as a regularizer for manifold smoothing. Manifold smoothing can effectively solve the above problems of transductive learning. The interpolations between neural network features based on similarity graphs are used by embedding propagation. Experiments show that embedding propagation can produce a better embedding manifold and our model in standard few-shot datasets, such as miniImagenet, tieredImagenet, CUB has state-of-the-art results. It significantly outperforms the existing few-shot approaches, which consistently improves the accuracy of the models by about 11%.

It is expected that peer to peer energy trading will constitute a significant share of research in upcoming generation power systems due to the rising demand of energy in smart microgrids. However, the on-demand use of energy is considered a big challenge to achieve the optimal cost for households. This paper proposes a blockchain-based predictive energy trading platform to provide real-time support, day-ahead controlling, and generation scheduling of distributed energy resources. The proposed blockchain-based platform consists of two modules; blockchain-based energy trading and smart contract enabled predictive analytics modules. The blockchain module allows peers with real-time energy consumption monitoring, easy energy trading control, reward model, and unchangeable energy trading transaction logs. The smart contract enabled predictive analytics module aims to build a prediction model based on historical energy consumption data to predict short-term energy consumption. This paper uses real energy consumption data acquired from the Jeju province energy department, the Republic of Korea. This study aims to achieve optimal power flow and energy crowdsourcing, supporting energy trading among the consumer and prosumer. Energy trading is based on day-ahead, real-time control, and scheduling of distributed energy resources to meet the smart grid’s load demand. Moreover, we use data mining techniques to perform time-series analysis to extract and analyze underlying patterns from the historical energy consumption data. The time-series analysis supports energy management to devise better future decisions to plan and manage energy resources effectively. To evaluate the proposed predictive model’s performance, we have used several statistical measures, such as mean square error and root mean square error on various machine learning models, namely recurrent neural networks and alike. Moreover, we also evaluate the blockchain platform’s effectiveness through hyperledger calliper in terms of latency, throughput, and resource utilization. Based on the experimental results, the proposed model is effectively used for energy crowdsourcing between the prosumer and consumer to attain service quality.

Analysis of Brazil’s rainforest fires caused by various factors has become a hot topic nowadays,. Mining of rainforest fire data through learning unlabeled training samples can reveal inherent properties and patterns, providing a clue for fire prevention. Among commonly used mining approaches, clustering algorithms based on density estimation can relatively effectively capture the potential ignition features through probability calculation, while the Gaussian mixture model (GMM) based on Expectation-Maximum (EM) can effectively quantify fire distribution curves and decompose a fire object into different shape clustering problems based on the actual distribution characteristics of fires data, and thus cluster fires more accurately. However, when the discrimination of probability density is not apparent, the clustering effect is susceptible to both the number of parameters used in clustering and the shape of the clustering problem. Therefore, in the present paper, based on a new strategy of selecting and updating the parameters in the GMM, a new hybrid clustering model called Principal Component Analysis-boosted Dynamic Gaussian Mixture Clustering model (PCA-DGM) is developed. Specifically, Principal Component Analysis (PCA) reduces the dimension of fire samples and strengthens key ignition features. Furthermore, a new dynamic distance loss function is developed by dynamically selecting density parameters or distance parameters, whose computing value is utilized as one important parameter of the clustering shape decision of the GMM. Using the PCA-DGM, which can effectively solve clustering problems with various shapes, the causes of forest fires in Brazil are analyzed at both the temporal and geographical levels, and the experimental results demonstrate that the proposed PCA-DGM in this paper has a better clustering effect than the other traditional clustering algorithms.

The epidemic increase in online reviews’ growth made the sentiment classification a fascinating domain in academic and industrial research. The reviews assist several domains, which is complicated to gather annotated training data. Several sentiment classification methodologies are devised for performing the sentiment analysis, but retrieval of information is not accurately performed, less effective, and less convergence speed. In this paper, we propose a sentiment paper proposes a sentiment classification model, namely Spider Monkey Crow Optimization algorithm (SMCA), for training the deep recurrent neural network (DeepRNN). In this method, the telecom review is employed to remove stop words and stemming to eliminate inappropriate data to minimize user’s seeking time. Meanwhile, the feature extraction is performed using SentiWordNet to derive the sentiments from the reviews. The extracted SentiWordNet features and other features, like elongated words, punctuation, hashtag, and numerical values, are employed in the DeepRNN for classifying sentiments. To retrieve the required review, the Fuzzy K-Nearest neighbor (Fuzzy-KNN) is employed to retrieve the review based on a distance measure. With rigorous assessments and experimentation, it is observed that the proposed SMCA-based DeepRNN performs better in terms of accuracy of 97.7%, precision of 95.5%, recall of 94.6%, and F1-score 96.7%, respectively.

Based on the good feature learning ability of the pyramid scene parsing network, a method for extracting the centerline of structured light stripes of weld lines based on the pyramid scene parsing network and Steger algorithm is proposed. This method avoids the traditional complex weld image preprocessing technology, and simplifies the operation steps of extracting the centerline of the structured light stripe of the weld image line. In this paper, the pyramid scene parsing network is used to predict the pixels containing weld feature information. Through the pyramid pooling module, the local and global context feature information is fused to supplement the feature information of the weld edge, and then the Steger algorithm is used to extract the weld feature centerline. The results show that the method in this paper can accurately extract the centerline position of the structured light stripe of the weld line under the interference of reflection, and the average value reaches 86.8% on the accuracy evaluation index mean intersection over union, the 18.93 pixels on the weld extraction accuracy index root mean square error, and the average time of extracting the center line of structured light stripe of weld line is $0.188s$ .

As the number of space debris has increased rapidly in recent years, it poses a major threat to the safety of spacecraft in space, and damage assessment for space debris hypervelocity impact is very important. In order to more comprehensively and accurately describe the damage defects in the infrared inspection data collected by infrared thermal imaging technology, this paper proposes an ultra-high-speed impact damage detection algorithm based on infrared reconstruction image fusion. The algorithm first preprocesses the obtained infrared thermal response image sequence and separates the damage features, and then applies a multi-objective evolutionary optimization algorithm to extract the typical transient thermal response, and then reconstructs the feature infrared images. Finally, image fusion based on guided filtering is performed on the damage reconstruction infrared images. In this paper, several reconstruction images that represent different types of impact damage defect are merged together to improve the detection ability of the algorithm. Infrared detection experiments on damaged specimens obtained from actual hypervelocity impacts verify the effectiveness of the proposed algorithm.

This paper proposed a novel fuzzy best-worst multi-criteria group decision-making method to solve the group decision-making (GDM) problem with multi-granular linguistic approach, which is an effective and promising technique to tackle this issue. In the proposed method, the selectable multi-granularity linguistic term sets (LTS) are firstly provided for experts to expressed their individual assessment information. Then, the improved fuzzy BWM is employed to calculate the weights of criteria with the form of fuzzy numbers. In current several studies using the BWM for group decision-making, only two unified best and worst criteria are given, which cannot reflect the evaluation of the best and worst criteria by different experts, resulting in the omission of information. Moreover, the difference between the best and worst criteria initially given and the experts’ ideas will cause the experts to be inaccurate in the comparison of each criterion. Therefore, in this paper, in order not to omit too much information, each expert will determine the best and the worst criteria. The evaluation information of each expert is integrated into two comparison vectors according to the transformation formula proposed in this paper. What’s more, an improved input-based consistency measurement is proposed, which can provide the DMs with a clear guideline on the revision of the inconsistent judgement(s). Finally, two examples are given to illustrate the effectiveness and applicability of the proposed method.

In order to solve the problems of unbalanced sample data and the lack of consideration of temporal information in existing Siamese-based trackers, this paper proposes a Siamese recurrent neural network and region proposal network (Siamese R-RPN), which can be trained in an end-to-end manner. Siamese R-RPN is consisted of Siamese network, recurrent neural network and region proposal network. Image features extracted by the Siamese network are strengthened by the channel and spatial attention mechanisms, and are sent to the RPN for classification and regression. Temporal information is processed by a recurrent neural network-based Long Short-Term Memory (LSTM) to predict the rough location of the target, it is mapped to the anchor feature map of the RPN for anchor selection. This makes the positive and negative samples participating in the training procedure to become more balanced and representative. Because of the collaborative use of temporal and spatial information, the tracker proposed in this paper has achieved state-of-the-art performance on three large tracking benchmarks—OTB 2015, VOT2016 and VOT 2018—where this verifies its effectiveness.

Hyperspectral and multispectral information processing systems and technologies have demonstrated its usefulness for the improvement of agricultural productivity and practices by providing useful information to farmers and crop managers on the factors affecting crop status and growth. These technologies are widely used in a range of agriculture applications such as crop management, crop yield forecasting, crop disease detection, and the monitoring of agriculture land usage, water, and soil conditions. Hyperspectral information sensing can acquire several hundred spectral bands that cover the electromagnetic spectrum of an observational scene in a single acquisition. The resulting hyperspectral data cube contains a large volume of spatial and spectral information. The hyperspectral sequence of images or video further increases the data generation velocity and volume which lead to the Big data challenges particularly in agricultural remote sensing applications. This paper is structured to first give a comprehensive review of representative studies to provide insights into significant research efforts in agriculture using Big data, machine learning and deep learning with the focus on frameworks or architectures, information processing and analytics with hyperspectral and multispectral data. The potential for utilizing Big data, machine learning and deep learning for hyperspectral and multispectral data in agriculture is very promising. The paper then further explores the potential of using ensemble machine learning and scalable parallel discriminant analysis which takes into consideration the spatial and spectral components for Big data in agriculture. To the best of our knowledge, no similar review study on agriculture with Big data, machine learning and deep learning for hyperspectral and multispectral information processing has been reported. Furthermore, the potential of ensemble machine learning and scalable parallel discriminant analysis has not been explored in agriculture information processing. Experiments and data analytics have been performed on hyperspectral data from agriculture for validation. The results have shown the good performance of our approach.

Due to the complex and changeable surrounding rock geology of the shaft, the design of the shaft blasting is affected by many factors. It is difficult to determine the blasting parameters scientifically and reasonably using traditional methods. Aiming at the above key technical problems, a design method of vertical shaft blasting scheme based on the integration of knowledge base and artificial intelligence (AI) is proposed in this paper. Firstly, the characteristic weight of factors affecting the design of the blasting was obtained based on the analytic hierarchy process (AHP) of particle swarm optimization, and according to the experience of experts, the type of surrounding rock, rock sturdiness coefficient, section shape, and gross diameter were selected as the key indicator for the decision-making of the blasting scheme. Secondly, the expert knowledge base and inference machine are established. This paper puts forward the research method of integrating knowledge base with a relational database, describes and expresses information content with knowledge attributes, and establishes an explosion expert knowledge base based on expert experience, industry norms, and typical cases. The cyclic search algorithm and production rule reasoning strategy were developed, and a hybrid reasoning machine for vertical shaft blasting design was established. Then, a visual intelligent design system for vertical shaft blasting was established by using computer development technology. Finally, the system has been applied in Jinzhuang Coal Mine. Compared with the traditional blasting scheme, the total number of holes in the optimized scheme of the system model is reduced by 8.39%, the single-cycle progress is increased by 11.11%, the hole utilization rate is increased by 10%, and the unit explosive consumption is reduced by 15.57%. The comprehensive effect is good, it provides a new way for vertical shaft blasting design.