Search

The paper recommender system has provided great convenience for people to obtain knowledge of what they are interested in. Existing recommendation works mainly rely on learning entities' feature representations (network embedding) from the Heterogeneous Information Networks (HINs) through meta-paths. Most of them suffer from the difficulty of designing and selecting meta-paths in HIN. Moreover, changes in papers' captive audiences and users' reading favor in the recent trends are always covered up by long-term trends. To this end, we introduce a multi-attention aware recommendation method, named MARec. Specifically, to efficiently design and select meta-paths, we propose an auxiliary node-driven HIN construction method and combine it with the graph attention network (GAT) to learn the entities' feature representations. To endow the recommendation results with timeliness, we design an attention-aware Bi-directional LSTM (Bi-LSTM) and a compensation mechanism to incorporate the recent trend changes. Furthermore, to take advantage of the importance of different papers/users, we utilize an attention mechanism to capture their importance by adaptively fusing the interactive historical data of those paper/user entities. We use extensive experiments on three real-world datasets to show that our proposed MARec outperforms current representative baselines by 11.49%, 11.40%, and 13.37% on average on three main metrics. • A general heterogeneous information network towards simplifying meta-path design. • Embedding with attention mechanism avoids painstakingly selecting meta-paths. • Feature extraction under multi-time scale greatly improves the recommended results. • MARec demonstrates superior performance on different real-world domain datasets. [ABSTRACT FROM AUTHOR]

• Newbie study using standard ML techniques with HAR Application and Discussions. • Activities found in Literature with the corresponding reference. • Co-occurrences between activities and sensors with the corresponding reference. • Summary and comparison among the different datasets found in Literature. • Summary of the experimentation settings with performance scores found in Literature. Human Activity Recognition (HAR) is an essential area of research related to the ability of smartphones to retrieve information through embedded sensors and recognize the activity that humans are performing. Researchers have recognized people's activities by processing the data received from the sensors with Machine Learning Models. This work is intended to be a hands-on survey with practical's tables capable of guiding the reader through the sensors used in modern smartphones and highly cited developed machine learning models that perform human activity recognition. Several papers in the literature have been studied, paying attention to the preprocessing, feature extraction, feature selection, and classification techniques of the HAR system. In addition, several summary tables illustrating HAR approaches have been provided: most popular human activities in the literature with paper references, the most popular datasets available for download (Analyzing their characteristics, such as the number of subjects involved, the activities recorded, and the sensors with online-availability), co-occurrences between activities and sensors, and a summary table showing the performance obtained by researchers. =The paper's goal is to recommend, through the discussion phase and thanks to the tables, the current state of the art on this topic. [ABSTRACT FROM AUTHOR]

The detection of underwater fish targets is critical for ecological monitoring and marine biodiversity research. However, underwater fish detection is typically constrained by problems including low image quality and variable underwater surroundings. On behalf of further improving the underwater fish detection accuracy in complex underwater environments, this paper proposes a dual-path (DP) Pyramid Vision Transformer (PVT) feature extraction network named DP-FishNet. The backbone network DP-PVT composed from the PVT Network is made up of two feature extraction paths. The first represents the Vision Transformer path, which extracts global features to enhance the distinction between the foreground and background of underwater images. The second is the convolutional neural network path, which enhances the accuracy of detecting small targets by extracting local features. Additionally, to more effectively utilize the feature information extracted by the network, this paper provides a promising solution to employ the content-aware reassembly of features (Carafe) in the feature pyramid network (FPN). The seesaw loss is utilized as a classification loss to address the problem of unbalanced samples caused by the gap in the number of fish populations. According to the experimental findings, the AP and AP50 of the DP-FishNet are 76.0% and 95.2%, respectively. In comparison to currently available advanced two-stage detection algorithms, the quantity of computation and parameters is reduced by approximately 40%. DP-FishNet strengthens the ability to extract global and local features from underwater images and enhances feature reuse. DP-FishNet can be utilized to detect fish targets in actual and complicated underwater habitats. [ABSTRACT FROM AUTHOR]

Edge detection is of critical importance for middle-level and high-level tasks in computer vision. Existing edge detection methods usually use VGG16 as the encoding network and achieve exceptional performance through transfer learning, which has the characteristics of high parameters and high computational cost. Researchers have implemented edge detection by designing decoding networks. Unlike existing methods, this paper is inspired by the effective mechanisms of edge detection in the parallel pathway of biological vision and proposes a new lightweight encoding–decoding structure for edge detection networks, which we refer to as XYW-Net. In the encoding network, we drew inspiration from the receptive field properties of X-type cells, Y-type cells, and W-type cells involved in the parallel pathway, and developed more meticulous convolutional models. Based on the structure of the parallel pathway, the three cell modules are combined into a novel encoding network and excellent feature extraction performance is obtained. Within the decoding network, this paper draws inspiration from the feature integration ability of the inferotemporal cortex (IT) and introduces a novel feature integration module to constitute the decoding network. Experiments show that the proposed network in this paper obtains the Optimal Dataset Scale (ODS) = 0.812 on the BSDS500 dataset with only 0.79 M parameters. Additionally, the model has exhibited outstanding ODS performance on various publicly available edge detection datasets, including NYUD, BIPEDv1, and Multicue. Moreover, future research could concentrate on implementing more efficient antagonistic mechanisms and devising networks grounded in higher-level visual cortical mechanisms. This could further augment the performance and efficiency of edge detection networks. We encourage researchers to delve into these directions within the field. The codes are available at https://github.com/PXinTao/XYW-Net. [ABSTRACT FROM AUTHOR]

• Pinch force strength is accurately predicted using sEMG. • Different pinch movements are accurately classified using sEMG. • Similarity of EMG features is evaluated. • An EMG feature-to-image method is proposed. The movement of intelligent EMG-driven prosthesis mainly relies on the synergy of different fingers to achieve function of grasping objects. The paper proposes a novel scheme for force prediction and movement classification about pinch between different fingers based on surface electromyography (sEMG) using machine learning. The pinch force and sEMG signals are recorded synchronously by a data acquisition device. Eight features are extracted, which are proven to have better performance in the estimation of sEMG-to-force. We present a novel feature selection method that uses the one-dimensional time series similarity assessment based on Manhattan distance to eliminate the repetitive information between features. Three optimal features carrying less repetitive information are retained. Seven machine learning algorithms are used to predict force strength. The results show that the Long Short Term Memory (LSTM) has the best performance of force prediction, achieving a R 2 of 0.9517 and RMSE of 3.2723. The paper proposes a novel method of converting EMG feature sequences to the normalized gray image in order to classify the finger movement. Five classifiers based on image feature extraction and the Convolutional Neural Network (CNN) are developed respectively. The experimental results indicate that the CNN performs best, achieving an accuracy of 97.66%. In this way, it not only realizes the accurate force prediction, but also realizes the movement classification between different fingers. The proposed methodology has the potential to realize simultaneous force and movement control of prosthetic hand. [ABSTRACT FROM AUTHOR]

Identifying the abnormal samples from a data set and determining their type are two key tasks of anomaly detection. However, the existing anomaly detection algorithms are generally faced with the defects of weak generalization ability and insufficient interpretation, the core reason of which is that they cannot mine specific features for different abnormal types. In this paper, a new anomaly detection algorithm aiming at feature selection for different abnormal types is developed. Inspired by self-supervised learning, we take the stationarity of variance changes of abnormal score similarity as a pretext task and combine it with wrapped search method. Then, the features and the corresponding parameters for different abnormal types can be screened to apply to the downstream task of anomaly integration detection. To verify the efficiency of the new algorithm, we conduct two sets of experiments to compare the new algorithm with 11 classical anomaly detection and 3 clustering anomaly detection algorithms on the data sets WDBC, WPBC and Wilt from DAMI database with the evaluation measures P@n , Adj- P@n , AP, Adj-AP and AUC. The experiment results show that, both in the identification and classification on abnormal samples, all performance measures of the new algorithm are explicitly better than that of the contrastive algorithms. Also, we apply the new algorithm to the Chinese auto insurance market, and find that the results can help managers to identify the main patterns of fraudulent claims and to summarize the feature combinations of fraud behaviors. In general, the new algorithm developed in this paper has the following advantages compared with traditional algorithms: 1) It can directly capture abnormal features and realize effective recognition of abnormal types, which effectively bridge the gap between abnormal judgement and feature screening. 2) Automatic screening of abnormal features can be completed under the condition of self-updating learning optimal strategy. 3) Only a few features are extracted from all features to reveal the abnormal characteristics, which significantly improves the interpretability and generalization ability of the algorithm and its results. In a word, through the novel self-supervised design method, feature screening is skillfully integrated into the anomaly detection task, which may provide a new way for anomaly detection research. [ABSTRACT FROM AUTHOR]

The Coronavirus Disease (COVID-19) caused a lot of mortality. The high mortality rate occurred because of the physicians' wrong or late identification of COVID-19 severity. So, developing Computer-Aided Design (CAD) systems using Artificial Intelligence (AI) techniques is critical to help physicians correctly identify the severity of COVID-19 in the early stages of the pandemic and then decrease the COVID-19 mortality percentage. In this paper, we develop a new green CAD system using a new hybrid handcrafted feature extraction algorithm and two-stage neural network architecture to grade the COVID-19 patient based on Computed Tomography (CT) scan images as having a moderate, severe, or critical infection. Because the proposed system uses handcrafted feature extraction algorithms, it consumes minimum resources and time than recent works. The proposed system consists of three phases: lesion segmentation, feature extraction, and diagnosis. Firstly, lesions from the CT scan image are manually segmented, and then three schemes are applied to extract salient features from the segmented lesions. These schemes are the Histogram of Oriented Gradients (HOG), Speeded Up Robust Features (SURF), and a new hybrid method that consists of cascading Discrete Wavelet Transform (DWT) and Gray-Level Co-Occurrence Matrix (GLCM). Then, the Cumulative Distribution Function (CDF) is computed for each scheme to extract the statistical markers. In the grading phase, a two-stage neural network approach is used. First, the extracted features are individually trained and tested for each scheme in the first neural network stage, and then the results of the first stage are combined to train and test each patient in the second neural network stage. The performance of the proposed system was assessed on a CT image dataset of 300 COVID-19-positive patients collected from the Cancer Imaging Archive (TCIA) website. The experimental results showed that our proposed system achieved 100% accuracy and kappa when the dataset was partitioned into 80% for training and 20% for testing. Also, it achieved 95.67% ± 0.47, 99.33% ± 0.77, and 100% ± 0 accuracies and 93.48% ± 0.74, 98.997% ± 1.16, and 100% ± 0 kappa values when the data was organized using 2, 4, and 10 folds, respectively. A green complexity algorithm analysis shows that this proposed system takes O(n) time complexity and 1 h and 20 min to train and test all cases. The performed green complexity analysis shows that the proposed system consumes 117.80 g Carbon Dioxide Equivalent (CO2e), 130.80 Wh, and 0.13 tree months for the carbon footprint, the energy needed, and the carbon sequestration, respectively. These results show that the proposed work consumes fewer resources and provides a green CAD system. • This paper proposes a CAD system for COVID-19 severity prediction. • The proposed system uses a two-stage neural network architecture. • The proposed system consists of two phases: feature extraction and severity identification. • A green complexity analysis is performed to prove the greenness of the proposed system. [ABSTRACT FROM AUTHOR]

Epilepsy is one of the most paramount neurological diseases, affecting about 1% of the world's population. Seizure detection and classification are difficult tasks and are ongoing challenges in biomedical signal processing to enhance medical diagnosis. This paper presents and highlights the unique frequency and amplitude information found within multiple seizure types, including their morphologies, to aid the development of future seizure classification algorithms. Whilst many published works in the literature have reported on seizure detection using electroencephalogram (EEG), there has yet to be an exhaustive review detailing multi-seizure type classification using EEG. Therefore, this paper also includes a detailed review of multi-seizure type classification performance based on the Temple University Hospital Seizure Corpus (TUSZ) dataset for focal and generalised classification, and multi-seizure type classification. Deep learning techniques have a higher overall average performance for focal and generalised classification compared to machine learning techniques, whereas hybrid deep learning approaches have the highest overall average performance for multi-seizure type classification. Finally, this paper also highlights the limitations of the TUSZ dataset and suggests some future work, including the curation of a standardised training and testing dataset from the TUSZ that would allow a proper comparison of classification methods and spur advancement in the field. [ABSTRACT FROM AUTHOR]

Automatic fault diagnosis in many systems is performed via the analysis of vibration data in the time or frequency domain. In the literature, many approaches for extracting featured from signals for fault diagnosis have been proposed. In this paper, we apply a variety of transform functions including Fourier, Wavelet, etc. to extract features from the vibration data, which includes the statistical features and some features automatically extracted via Convolutional Neural Networks (CNNs). For each of these feature extraction approaches, a learning algorithm is trained to diagnose the faults and their results are aggregated in an ensemble machine learning algorithm. The weights of the base learner algorithms are optimized via an evolutionary algorithm to achieve the best-weighted voting scheme. The architecture of CNNs has a significant effect on the performance of the algorithm, thus, in this paper, an evolutionary algorithm is proposed to find the best architecture for CNNs in fault diagnosis. The CNNs are trained via gradient descent algorithms which suffer from getting stuck in local optima. To manage this, we propose an evolutionary algorithm that benefits from the speed of gradient descent and the global search of evolutionary algorithms. The proposed algorithm is tested on a number of benchmark problems and the experimental results are presented. • An ensemble algorithm is proposed for fault diagnosis. • Different feature extractors are exploited. • An evolutionary training algorithm is proposed. [ABSTRACT FROM AUTHOR]

There are lots of research papers for ASC, and in recent years it is rapidly increasing. DCASE also provides different types of competition for the submission of several papers to solve the various tasks of ASC and it is the opportunity for the research scholars either to participate in those competitions or to provide the enhanced model for ASC. This paper provides details about the various recent approaches along with the block diagram used for pre-processing required before model development for ASC. It also includes a description of different recent techniques used for the classification of different sounds for ASC tasks. The comparative analysis for different recent available techniques both for pre-processing and classification has been done and summarized in this paper. It also describes the contributions towards the survey on ASC by comparing this paper with some existing survey papers based on several parameters like functionality described separately, results from description with quantifiable value, a dataset with proper quantifiable analysis, and pictorial representation of model discussed, etc. Finally, considering the benefits for eminent research scholars, this paper has also focused on the details for future directions for both pre-processing and classification for ASC. [Display omitted] • A systematic review on pre-processing and classification techniques for acoustic scenes. • Comparative analysis of pre-processing and classification techniques. • Discussion on different processing techniques for classification. • Identifies the future directions for further research in the area. [ABSTRACT FROM AUTHOR]

• The feature representation of the network was extracted from a randomly searched sequence of vertices. • In networks of scientific documents, representation learning needs to consider similarity and citation relationships. • This paper proposed a method to search and learn multiple paths connecting similar documents in a scientific document network. • From scientific papers and patent datasets, we found empirical evidence that our algorithm is suitable for representation learning. • This paper can contribute to preparing the basis for inducing innovation in science and technology from scientific documents. A network is a structure that can represent an organic relationship of observations. Network representation learning has the advantage of extracting latent features in a network. In recent years, various algorithms have been developed to embed network representation in a low-dimensional feature space. Conventional algorithms learn a sequence of vertices that are randomly searched from a network and convert a vertex into a vector. This study focused on the representation learning of document networks. Thus, this study proposes a scalable network representation learning algorithm that preserves the structure and content of the document network and reduces data sparsity. To do this, we used the shortest walks to extract a sequence of vertices that reflected semantic similarity from a network. The shortest walk is a method for finding several shortest paths that connect similar vertices. After this, the sequence of vertices was inputted into the language model, assuming that the sequence was a sentence. Scientific papers and patent datasets provided empirical evidence for the validity of our approach. The former, the bibliography network of scientific paper, consists of 17,716 nodes and 105,734 edges. The latter is a patent describing a technology for energy, representing a network of 24,894 nodes and 28,314 edges. The results showed that the proposed algorithm exhibited the lowest error of the tested methods in multi-class classification and link prediction tasks. The results of these tasks provided evidence that our algorithm is suitable for extracting feature representations of document networks. [ABSTRACT FROM AUTHOR]

Feature extraction networks play a crucial role in classification algorithms. However, most convolutional neural networks use the form of residual block stacking to extract downstream features of images, and this simple form is not sufficient to obtain better classification performance. Therefore, this paper proposes a YOLOX detection framework based on an improved ConvNeXt which incorporates many advantages of Swin-Transformer with efficient classification performance. To address the problems of unstable IOU convergence and inaccurate regression coordinates of the algorithm, this paper adopts Manhattan distance as the new penalty term, and the new loss function combines the advantages of Euclidean distance and Manhattan distance, which has no gradient explosion in the early training period and can drop smoothly to a stable point in the late training period. The proposed algorithm is robust to negative effects such as dust, light, shadows, and occlusions, and is well-suited for masked-face detection in industrial and public scenarios. The proposed model is tested on the VOC dataset and the AIZOO open-source masked-face dataset, obtaining 5.58% AP improvement on the VOC dataset and reaching 96.665% AP on the masked-face dataset, which is better than other algorithms. • Designing a deep neural network to use for masked-face detection. • Anchor-free model is used to avoid time-consuming post-processing. • The improved ConvNeXt network improves classification performance. • Adopting Manhattan distance-based localization loss makes training stable. • The Mosaic data augmentation strategy alleviates the overfitting problem. [ABSTRACT FROM AUTHOR]

Image segmentation is one of the vital steps in medical image analysis. The segmentation precision and delicacy of targets affect the diagnosis accuracy to a great extent. Deep learning-based technology is currently a hot research direction in the field of medical image analysis. Compared with manual labeling, deep convolutional neural networks (DCNN) can extract valuable deep features that are difficult for humans to identify. However, some existing DCNN methods tend to only focus on extracting deep features and ignore the semantics loss in the process of continuous pooling and convolution. In addition, medical image datasets are characterized by small-scale, mass noise, and imperceptible differences between foreground and background. How to construct sufficient contextual dependency and feature compensation from multiple perspectives to improve the segmentation performance is still faced with great challenges. In this paper, we propose a deep attention network with multi-perspective feature compensation, called Multi-Perspective Feature Compensation Network (MPFC-Net), to supplement semantics loss during continuous pooling and convolution as much as possible and establish strong contextual dependency. Merits of MPFC-Net are as follows: (1) Serializing Multi-scale Aggregation Blocks (SMAB) are constructed to extract features with different scale receptive fields, capture semantics loss in convolutions, and enhance the nonlinear fitting capability of the network. (2) S pace and C hannel Compound M ultidimensional I nteraction Attention (SCMI Attention) is designed to consider the weights of feature maps of space and channel and build long-term information interactions between different dimensions. (3) Multi-Perspective Feature Compensation (MPFC) is proposed to establish dense context interaction and compensate for semantics loss in continuous pooling and convolution. (4) Deep Feature Reuse (DFR) module is adopted to strengthen the propagation and advanced features reuse. We conduct extensive experiments on the 2018 Data Science Bowl (2018DSB), the ISIC-2018 Challenge (ISIC2018), the JSUAH-Cerebellum, and the DRIVE dataset, and the results show that MPFC-Net achieves competitive performance with simultaneous increases in various metrics. [ABSTRACT FROM AUTHOR]

Telco players are constantly evaluating their loss in customer revenue or customer churn, especially on the prevention, control, and risk assessment of churn. In this paper, a domain-order based shallow fusion deep learning model, the Fully Connected Layer Convolutional Neural Network - Long Short-Term Memory (FCLCNN-LSTM) is designed and applied to a decision support system to assess the risk of customer churn. First, a majority least absolute shrinkage and selection operator (Maj-LASSO) algorithm is proposed to identify churn predictors that have an important impact on classification. The proposed Maj-LASSO algorithm is different from previous studies that solved the study of unbalanced data and feature selection separately, but solved both as one problem to give the feature selection method under unbalanced data. Then, a fully connected layer based on multiple ReLU neurons is used to balance the importance of the features. Doing so addresses the problem where some key features are overlooked due to their unequal contribution to the risk assessment of telco customer churn. Third, a 2D convolutional neural network based on domain information and an LSTM model based on ordinal information are combined to learn the complex mapping between the features and labels, to improve the accuracy and generalizability of the predictive model. The proposed FCLCNN-LSTM model is different from the previous single spatial model or temporal model, but improves the network classification feature extraction capability from both temporal and spatial dimensions, thus innovatively solving the problem of model development for large-scale and high-dimensional data, and applying it in the field of telecommunication customer churn prediction. Finally, using three public customer churn datasets from the Kaggle and UCI platform, the FCLCNN-LSTM model is compared against other classification models, namely, Logistic Regression, Support Vector Machine, Random Forest, eXtreme Gradient Boosting (XGBoost), Convolutional Neural Network, and Long Short-Term Memory. The experimental results inform that the accuracy of the FCLCNN-LSTM is 3.43% higher than that of the other comparison models, while the AUC is 4.84% higher. Thus, the FCLCNN-LSTM model provides a better decision-making reference for telco players to identify potential churners. [ABSTRACT FROM AUTHOR]

• To develop a hybrid deep learning model for student behavior analysis. • To develop an Opposition-based Food Foraging-Pelican Optimizer (O-FFPO) algorithm. • To carry out optimal feature selection by utilizing developed O-FFPO algorithm. • To develop an O-FFPO to enhance the classification performance of developed model. • To adopt a Radian Basis-Recurrent Neural Networks (RB-RNN) deep learning model. With the development of the internet, e-learning is rapidly growing in worldwide, which pave more attention to the students. Further, the learning materials like video lectures, shared assignments as well as lecture slides are performed during the online courses. Moreover, these activities among the students exhibit different behaviors. Hence, the analysis of student behavior becomes complicated in the existing models. This research focuses on developing a deep learning model for analyzing student behaviour in an accurate manner. This paper aims to develop a new student behavior prediction model for E-learning by evaluating brain signals and chatbot interaction. This suggested model collects data from both chatbots and the Brain waves of students. The collected textual data from the chatbot are forwarded to the feature extraction phase by word2vector and bag-of-n grams. Then, the optimal features are selected using an Opposition-based Food Foraging-Pelican Optimizer (O-FFPO). Similarly, the second set of Brain waves of students are forwarded to the phase feature extraction stage by Short-time Fourier transforms (STFT), and then, deep features are extracted using Convolutional Neural Network (CNN). These gathered data can also be forwarded to the optimal feature selection process using the same O-FFPO. Finally, these fused features are fed to Radian Basis-Recurrent Neural Networks (RB-RNN) to get student behaviors, where the constraints are optimized by the same O-FFPO. Various performance metrics are compared and validated with the existing approaches. Moreover, the developed model shows 95% and 93% concerning accuracy and F1-score. Finally, experimental evaluation was carried out with the estimation of both brain waves and chatbot queries of students. [ABSTRACT FROM AUTHOR]

Detecting Sulphur hexafluoride (SF6) gas leaks using infrared video images is the latest technology. However, the current algorithm for autonomous real-time detection is imprecise and incapable of performing the real-time monitoring task effectively. In response to the aforementioned issues, we present a temporal–spatial feature fusion network (TSFF-Net) specifically designed for video real-time detection of SF6 gas leaks. Compared with the existing SF6 gas leak detection model based on deep learning, the biggest feature of TSFF-Net is the introduction of foreground pixel images with SF6 gas movement characteristics, which greatly enhances the model's ability to detect SF6 gas leaks. Simultaneously, by designing a new attention feature fusion module, introducing the Neck module of motion features, and the Backbone module to enhance small target feature extraction, we achieve high-precision SF6 gas leak detection in various complex scenarios, such as sparse leaks, complex backgrounds, camera shake, etc. The mean average precision (mAP) of TSFF-Net in the validation set was 56.3%, which is 13.2% higher than that of Yolov8s. In addition, the model proposed in this paper has parameters that are only 2.6 M, making it suitable for deployment on mobile or embedded devices. The code and test results are at https://github.com/1jianyao/TSFF-Net/tree/main. • A feature-fusion network is presented for the detection of SF6 gas leaks. • The two-branch features are fused by the attention mechanism. • The Neck module to enhance moving target detection is devised. • The Backbone module to enhance tiny object detection is proposed. [ABSTRACT FROM AUTHOR]

The metro system is indispensable for alleviating traffic congestion in the urban transportation system. Precise metro passenger flow (MPF) prediction is crucial in ensuring smooth operations of the metro system. Recently, the graph convolutional network (GCN), which is effective in the spatial feature extraction, has been applied in traffic prediction. However, most existing GCN-based methods construct the empirical graphs based on distance and adjacency, which cannot fully express the correlations of metro stations. This paper proposes a novel MPF prediction method consisting of three parts: K-means-based metro station functional clustering (KMSFC), external feature fusion, and dual-view recurrent GCN (DVRGCN). The KMSFC identifies the metro stations both having similar MPF changing tendencies and being located in similar urban functional areas. Furthermore, the DVRGCN is designed to simultaneously capture the spatiotemporal and external features. The dual-view GCN module in the DVRGCN captures both explicit and implicit spatial features of the metro traffic network. To demonstrate the capability for making accurate MPF predictions, the experiments using a real-world metro traffic dataset are conducted. The ablation experiments are also performed to prove the contribution of each module in the proposed method. The experimental results show that the proposed method outperforms other state-of-the-art traffic prediction methods. • Precise clustering of passenger travel patterns at different metro stations. • Comprehensive consideration of spatial and temporal features of metro networks. • Informative presentation of passenger flow volumes for different metro stations. • Accurate predictions of future passenger flows of all metro stations. [ABSTRACT FROM AUTHOR]

Chronic Kidney Disease (CKD) and liver diseases are progressive diseases with high morbidity and large fatal rates occurring generally in the adult population exclusively those people affected with hypertension and diabetes. Thus this paper develops a deep learning method named snake optimization-based BiGRU model for predicting kidney and liver disease earlier and precisely. Gathered medical data of the patients using the various IoT sensor devices stored under the patient log in the cloud. These data are abundant in quantity. The pre-processing is initially carried out to refill the missing details of the patient. The Convolutional Neural Network ResNet and snake optimization (SO) are included for extracting features and selecting optimal features. The disease prediction is performed using the BiGRU neural network which learns data from two various sources such as forward and backward for providing extremely precise prediction output. For evaluating the performance we compared the recall, MCC, accuracy, specificity, and Kappa score values of the method with the four existing disease prediction methods of LDA-GA-SVM, CNN-RF, Decision Tree, and XGBoost. The proposed method achieved a higher accuracy of 98.9 %, recall of 98.9 %, sensitivity of 98.9 %, and an F1-score of 98.2 %. [ABSTRACT FROM AUTHOR]

• The global and multi-scale local causal features are extracted to interpret the model and improve diagnosis accuracy. • The performance of the initial model can be improved by adopting an extensible framework to add enhancement nodes. • The sample incremental learning is developed to update incrementally based on the previous models without retraining. • Two experiments on bearing components validate the effectiveness and superiority of the proposed method. Intelligent fault diagnosis (IFD) plays a significant function in ensuring the reliable operation of mechanical equipment. However, most existing IFD methods are trained in batch learning way and the feature extraction process is unexplainable and it is challenging to satisfy the complex diagnosis requirements. Thus, this paper proposes a causal broad learning model (CBLM) guided by global and multi-scale local causal features for incremental machinery fault diagnosis. Firstly, rich global and multi-scale local causal features are extracted for training CBLM. Then, incremental learning capability is developed to update model by expanding or modifying the original weights, considerably reducing the time consumption and greatly improving the computational efficiency. When the initial model has inadequate nodes for feature learning, CBLM performs structural incremental learning by adding extra nodes to improve the diagnostic performance. As new samples with various fault degrees are entered, sample incremental learning is built to rapidly update based on previous model without retraining. Finally, two experimental results on bearings indicate that CBLM improves the testing accuracy of the initial model by 0.13% to 52.22% and 1.95% to 57.64%, respectively, and remains above 97.22% and 99.77% for the updated models. The training time consumption is reduced by 25.06 s to 133.0 s, 20.78 s and 141.9 s, respectively, and the subsequent models update time are only about 12 s and 10 s. Consequently, the proposed CBLM is an effective online IFD method because it is obviously superior to the existing IFD methods in regards to time consumption and diagnosis accuracy. [ABSTRACT FROM AUTHOR]

• The proposed method had fused multi-level information compensation and U-Net. • The module designed for multi-level information compensation and compression. • The method had fused compressed features and correlation features. Recently, Deep Convolutional Neural Networks have demonstrated high-quality reconstruction in image super-resolution procedure. In this paper, we propose improved image super-resolution reconstruction via multi-level information compensation and U-Net network to address the problem that the image super-resolution reconstruction algorithm based on deep neural networks tends to lose feature information in the feature extraction process, resulting in the lack of texture and edge details in the reconstructed image. Firstly, we design the U-net like network for image super-resolution reconstruction, which performs multi-level feature extraction and channel compression for the input features through the down-channel branch. It fuses the compressed features and extracts the correlation features of different channels through the bottom module, and performs multi-level feature extraction and channel recovery for the compressed correlation features through the up-channel branch. The multi-level information compensation model is then designed to compensate for the information lost in the channel compression process and the information that is difficult to recover in the channel recovery process of U-net like network. The experimental results can show that the proposed algorithm achieves a significant improvement in Peak Signal-to-Noise Ratio and Structure Similarity Index and visual effect compared with the state-of-arts algorithms. The average experimental results of PSNR from proposed method had improved by 1.63 dB, 1.53 dB, 0.97 dB and 0.94 dB compared to SRCNN, HAT, DAT and CARN, respectively. [ABSTRACT FROM AUTHOR]

In human–computer interaction (HCI) and especially speech signal processing, emotion recognition is one of the most important and challenging tasks due to multi-modality and limited data availability. Nowadays, an intelligent system is required for real-world applications to efficiently process and understand the speaker's emotional state and to enhance the analytical abilities to assist communication by a human-machine interface (HMI). Designing a reliable and robust Multimodal Speech Emotion Recognition (MSER) to efficiently recognize emotions through multi-modality such as speech and text is necessary. This paper propose a novel MSER model with a deep feature fusion technique using a multi-headed cross-attention mechanism. The proposed model utilize audio and text cues to predict the emotion label accordingly. Our proposed model process the raw speech signal and text by CNN and feeds to corresponding encoders for discriminative and semantic feature extractions. The cross-attention mechanism is applied to both features to enhance the interaction between text and audio cues by crossway to extract the most relevant information for emotion recognition. Finally, combining the region-wise weights from both encoders enables interaction among different layers and paths by the proposed deep feature fusion scheme. The authors evaluate the proposed system using the IEMOCAP and MELD datasets and conduct extensive experiments that obtain state-of-the-art (SOTA) results and show a 4.5% improved recognition rate, respectively. Our model secured a significant improvement over SOTA methods, which shows the robustness and effectiveness of the proposed MSER model. [ABSTRACT FROM AUTHOR]

The restoration technology of non-uniform blurred images is a challenging open topic. Most of the existing algorithms fail to effectively fuse multi-scale feature extraction with a self-attention mechanism, and also ignore the potential contribution of image frequency domain information to image restoration. Frequency domain features play an important role in restoring high-quality images and ignoring this property often leads to over-smoothing of the restoration results. In response to these problems, an image deblurring method based on self-attention and residual wavelet transform is proposed in this paper. Based on a single U-Net network, the multi-scale feature cross-fusion strategy and self-attention mechanism are combined to make the network pay more attention to different degrees of blurred regions so that relatively robust blurred features can be extracted for image deblurring. Meanwhile, considering the important role of frequency domain features for image restoration, the wavelet transform is embedded into the depth residual network to convert spatial domain features to wavelet domain, and the sharp details such as edge contours of blurred features are restored by making full use of the texture structure information possessed by high-frequency sub-bands, which further improves the image restoration performance. Experimental results of quantitative and qualitative comparison with other state-of-the-art methods show that the image deblurring effect of the proposed network performs favorably in terms of peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) metrics. The codes and models are available at https://github.com/BingY998/MRDNet. [ABSTRACT FROM AUTHOR]

Predictions of the oil price are critical to support intelligent decision-making for individual investors, governments, and corporations, but a challenging task since there are complex nonlinear and random fluctuations involved in series. Additionally, most existing methods mainly focus on the data-layer fitting, and inherit uninterpretable design flowcharts and unclear layer functions. Based on the fact that living things evolve numerous advanced cognition processes after long-term interactions with environments, it is a promising problem to design bionic deep prediction networks. In this paper, an a uditory-inspired m ulti- f eature c ollaboration network (AMFC-Net) is explored to predict the oil price, and includes the feature extraction, brain memory, and comprehensive prediction blocks. Specifically, through imitating that complex sounds are captured by auditory canals and converted into electrical impulses, the first block adopts four typical activation functions to extract multi-group complementary features for detecting nonlinear changes. Through imitating that electrical impulses are transformed to the left and right hemispheres for handling and analyzing, the second block realizes gating and cooperation mechanisms to learn long short-term dependencies and highlight core information. Through imitating that information is sent to higher cerebral cortices for sensing the environment, the third block maps relationships from features to targets for producing final predictions. In summary, multi-channel convolution operations establish the imitations of both form (clear functional layer) and spirit (layer-by-layer collaboration), which not only ensure the prediction effectiveness but also improve the AMFC-Net interpretability. Three experiments (comparison with 12 machine learning methods), three experiments (comparison with 8 hybrid methods under decomposition plus ensemble framework), and seven discussions under two real-world oil price datasets all indicate that the proposed AMFC-Net is superior and suitable to predict the oil price. [ABSTRACT FROM AUTHOR]

Optical Character Recognition (OCR) is a method to convert a scanned photo of"handwritten character recognition (HCR) or printed character recognition (PCR)" into a form of digital text. HCR is a version of OCR, which is remarkably modeled to identify handwritten text, while PCR aims at printed text identification. The identification of handwritten characters and digits is more complicated as compared to PCR because of the diversities in human writing styles, stokes, thickness, and curves of characters. Similarly, achievements in several computer vision tasks consider the Convolutional Neural Networks (CNN) to give an end-to-end solution for HCR with huge success. However, the process of significant feature learning for the identification of images is complicated with little data. Hence, this paper aims to develop a new handwritten character and digit recognition model with the incorporation of a deep learning strategy. Initially, the data related to Indian languages are collected from the standard benchmark datasets. Then, the collected data are given into the feature extraction phase 1, where the ResNet 151 is used for extracting the feature set 1. Similarly, the data gathered are considered in the feature extraction phase 2, where the Optimal Ensemble Pattern extraction approach is developed with Local Binary Pattern (LBP), Local Gradient Patterns (LGP), Local Tetra Pattern (LTrP), and Local Vector Pattern (LVP) for extracting the significant patterns from the language data. These extracted patterns are given into the ResNet 151 for getting the feature set 2. Here, the features from ResNet 151 get optimized with the enhanced optimization algorithm with Fitness-based Sail Fish Optimizer (F-SFO). The obtained feature set 1 and optimal feature set 2 are concatenated for performing final recognition. At last, the HCR is done with the help of developed Bi-LSTM-DNN to achieve the enhanced and accurate recognition of handwritten characters of the Indian languages. The performances of character recognition are further improved with the parameter optimization in Bi-LSTM-DNN with the same enhanced F-SFO. Overall result analysis, the accuracy of the designed method attains 95.12%. [ABSTRACT FROM AUTHOR]

For the purpose of managing financial risk and making investment decisions, interval stock price forecasting is essential. Currently, decomposition integration frameworks are widely used in point-valued stock price forecasting studies, mainly focusing on mining internal information. However, point forecasts are difficult to adequately capture price uncertainty and may suffer from loss of volatile information. Therefore, this paper proposes an enhanced interval-valued decomposition integration model for stock price prediction based on comprehensive feature extraction and optimized deep learning. Firstly, the interval variational modal decomposition with feedback mechanism (FIVMD) is proposed to extract internal features and can decompose interval values into interval trend and residual. FIVMD not only solves the interval decomposition challenge, but also helps to improve the internal feature extraction capability. Secondly, while considering the influencing factors more comprehensively, appropriate feature selection and compression techniques can effectively achieve external feature extraction, obtain the best influencing factors, and improve the modeling capability of high-dimensional data. Finally, the final prediction results are obtained by modeling the interval trend and residuals separately through the optimization algorithm and deep learning model to improve the prediction accuracy. The results of the empirical analysis reveal that the proposed interval decomposition integrated model has the smallest of the three evaluation metrics, where the values of interval mean average percentage errors (IMAPE) are 1.8188%, 1.1244%, 1.9001%, and 2.1542% respectively. This shows that the model is significantly more accurate and stable than the other comparative models, and it is a successful model for predicting interval-valued stock prices. [ABSTRACT FROM AUTHOR]

This paper demonstrates a novel approach for segmenting complex vessel-like structures from images of retinal vessels, surface cracks, and roadmaps, a challenging task due to nuisance variations in width, curvature, and branching patterns, as well as cluttered backgrounds caused by adverse imaging conditions. We introduce the Spectral Transformer (SpecFormer), a Transformer built from the frequency domain to segment the elongated and linear structured content of images. The idea behind SpecFormer is to take full advantage of the ability of low-frequency components in the Fourier domain to represent the overall structure, global patterns, and smooth variations. Specifically, a Sparse Spectral Neural Operator (SSNO) is proposed to modulate the sparse frequency-concentrated spectrum via learnt frequency-specific filtering, which can well represent the vessel-like structure in the Fourier domain. This operator, as the core component of Dual Attention Block (DAB), is designed in a dual-path way, i.e. , self- and scaling-attention paths, to simultaneously capture the long-range dependencies and contextual information of the feature. The complete form of the SpecFormer is built with multiple DABs and modules for patch manipulations and feature fusion. We evaluated the SpecFormer on a wide range of publicly available datasets and achieved consistent improvements over the state-of-the-art (SOTA) methods. Code is available at https://github.com/LouisNUST/Spectral_Transformer. • SSNO consisting of feed-forward layer, sparse spectral layer and residual layer. • DAB with SSNO-based scaling attention and self-attention for feature extraction. • U-shaped Spectral Transformer by stacking and arranging multiple DABs. • SpecFormer achieves new state-of-the-art performance on several datasets. [ABSTRACT FROM AUTHOR]

Models based on extremely deep convolutional networks and attention mechanisms now have powerful feature extraction abilities, improving text recognition performance in natural scenes. However, the very deep network model built with stacked layers brings massive parameters, which limits the application of the text recognition algorithm on storage-constrained devices. In this paper, we propose a lightweight and effective backbone called the Recursive Residual Transformer Network (RRTrN) for scene text recognition. Specifically, by leveraging recursive learning and a combination of convolutional layers and a transformer unit, RRTrN achieves powerful feature extraction while significantly reducing the number of parameters. This reduction in parameters promotes the deployment of our text recognition algorithm on storage-constrained devices, making it more accessible for practical applications. Furthermore, a recursive distillation strategy is presented to balance the recursive learning inference time and performance, enhancing the practicality and efficiency of RRTrN. Extensive experiments on mainstream benchmarks and popular models verify the generalization of RRTrN and achieve state-of-the-art recognition performance on five datasets. Notably, the classical STR model based on RRTrN can achieve a 3 percentage point increase in recognition accuracy or reduce the number of parameters by 80%. • Propose a lightweight and efficient backbone, RRTrN, for scene text recognition. • Present a simple and effective strategy, Recursive Distillation, for computational efficiency issues. • Good generalization on several classical and advanced models. [ABSTRACT FROM AUTHOR]

Feature extraction is one of the fundamental problems in pattern recognition research. For image recognition, extracting effective image features is the key to accomplish the recognition task. In this paper, a partial differential equations-based multi-rank two-dimensional canonical correlation analysis (PDEs-MR2DC2A) is proposed for multi-view feature extraction and pattern classification. Unlike most of the previous researches on multi-view algorithms that work directly on the original 2D representation, in our approach, we first utilize the evolution process of PDEs to extract the feature matrix of per-view. In addition, we employ multi-rank left and right projecting matrices to maximize the correlation. The computational complexity of PDEs-MR2DC2A is also analyzed. To evaluate the effectiveness of the proposed algorithm, we conducted a series of performance comparisons with some existing methods on several popular datasets. The experimental results showed that our proposed algorithm performed very well on these datasets and outperformed the existing related methods on some metrics. • A novel multi-rank multi-view algorithm based on PDEs is proposed. • This approach can extract rotation and translation invariant features. • Numerous experimental results demonstrate the effectiveness of our proposed method. [ABSTRACT FROM AUTHOR]

When a doe is approaching delivery, it often exhibits some specific behaviors such as eating less, sleeping less and pawing (scraping the floor with the hoof). Automatic recognition of these behaviors is valuable for farmers to determine whether does are nearing parturition or not. However, it is difficult for humans to recognize the behavior of pawing because movements are subtle and the foreleg is small. In this paper, we proposed a method to recognize the pawing behavior of prepartum does based on semantic segmentation, motion history image (MHI), and multi-scale histogram of oriented gradient (MS-HOG) features. Firstly, the foreleg region of the doe is segmented using YOLO-v5 and Unet. Next, MHI features are extracted from consecutive frames to represent the motion information of foreleg. Since MHI is insufficient to describe the local textures and shape information of the foreleg movements, we extracted MS-HOG based on the MHI features. A support vector machine (SVM) classifier was trained on the extracted features to determine whether or not a given frame contains pawing behavior. Furthermore, a word frequency vector was employed to characterize the proportion of pawing frames in a video episode with a length of one second to facilitate the recognition of pawing behavior. The proposed method was validated using both shorter video episodes and long videos. For the video episodes, our method achieved a classification accuracy of 93.14% and an average error rate of 9.72% in estimating the duration of the behavior. For the long videos, our method achieved a recall of 87.6% and precision of 77.8% when considering a temporal Intersection over Union (t I o U ≥ 0. 5). In summary, this study can be further combined with other characteristic behaviors to predict the impending delivery time of the does. • An automated method for identifying prepartum doe's pawing behavior. • Segment small target (foreleg) using a two-stage semantic segmentation method. • Reduce the computational overhead of feature extraction using multi-scale HOG. • Achieved 93.14% accuracy in pawing behavior recognition for video episodes. • Achieved 77.8% precision, 87.6% recall in long video considering tIOU ≥ 0.5. [ABSTRACT FROM AUTHOR]

In this research paper, we introduce a novel approach to enhance the performance of 2D feature matching and pose estimation through the integration of a hierarchical attention mechanism and knowledge distillation. Our proposed hierarchical attention mechanism operates at multiple scales, enabling both global context awareness and precise matching of 2D features, which is crucial for various computer vision tasks. To further improve our model's performance, we incorporate insights from an existing model PixLoc (Sarlin et al., 2021) through knowledge distillation, effectively acquiring its behavior and capabilities by ignoring dynamic objects. SAM-Net outperforms state-of-the-art methods, validated on both indoor and outdoor public datasets. For the indoor dataset, our approach achieves remarkable AUC (5 ° / 10 ° / 20 °) scores of 55.31/71.70/83.37. Similarly, for the outdoor dataset, we demonstrate outstanding AUC values of 26.01/46.44/63.61. Furthermore, SAM-Net achieves top ranking among published methods in two public visual localization benchmarks, highlighting the real benefits of the proposed method. The code and test suite can be accessed at link. 1 1 https://benjaminkelenyi.github.io/samnet/. • Geometric key-point feature extraction for 2D vision with spatial transformers. • Feature extraction with Knowledge distillation and self-attention. • Robust position estimation from multi-view camera systems. [ABSTRACT FROM AUTHOR]

Endoscopic ultrasonography (EUS) is a valuable imaging modality for diagnosing gastrointestinal submucosal tumors (SMTs). However, inherent content variations in EUS images due to gastrointestinal tract mobility and handheld ultrasound instability challenge the saliency of SMTs' visual features. The presence of fine-grained inter-class and large intra-class differences further complicates EUS-based diagnosis. To address these issues, this paper presents a novel Multi-Attribute Guided Contextual Attention Network (MAG-CA-Net) for interpretable SMT recognition in EUS. Inspired by endoscopists' clinical diagnosis expertise, our framework initially localizes abnormal areas based on echo attributes and subsequently determines tumor categories using contextual semantics. Experimental results demonstrate the effectiveness of MAG-CA-Net, exhibiting improved recognition recall and precision rates for gastrointestinal stromal tumor, leiomyoma, and pancreatic rest. Specifically, the MAG network facilitates abnormal area localization, while the CA mechanism guides the model to focus on the most discriminative tumor-context-associated regions. The proposed method achieved an average classification accuracy of 93.16%, an average precision of 93.17%, a weighted recall of 93.16%, and an average F1-score of 93.15 % for the three disease categories. The proposed approach provides crucial guidelines for data collection standards and model development in the clinical diagnosis process of SMTs under EUS. Its interpretability analysis enhances the credibility of clinical physicians towards assisted diagnostic methods based on deep learning. The source code will be publicly available at https://gitee.com/HangbinZheng/mag-ca-net. • Hierarchical identification strategy combining clinical knowledge and data-driven models for submucosal tumor detection. • Contextual attention mechanism enhances tumor recognition using gastrointestinal layered features. • Improved adaptability to endoscopic ultrasound image variation and fine-grained feature extraction compared to state-of-the-art models. • Visual interpretability close to experienced endoscopists enhances method's interpretability. [ABSTRACT FROM AUTHOR]

In this paper, we propose a 3D object detection method called MMAF-Net that is based on the multi-view and multi-stage adaptive fusion of RGB images and LiDAR point cloud data. This is an end-to-end architecture, which combines the characteristics of RGB images, the front view of point clouds based on reflection intensity, and the bird's eye view of point clouds. It also adopts a multi-stage fusion approach of "data-level fusion + feature-level fusion" to fully exploit the strength of multimodal information. Our proposed method addresses key challenges found in current 3D object detection methods for autonomous driving, including insufficient feature extraction from multimodal data, rudimentary fusion techniques, and sensitivity to distance and occlusion. To ensure the comprehensive integration of multimodal information, we present a series of targeted fusion methods. Firstly, we propose a novel input form that encodes dense point cloud reflectivity information into the image to enhance its representational power. Secondly, we design the Region Attention Adaptive Fusion module utilizing an attention mechanism to guide the network in adaptively adjusting the importance of different features. Finally, we extend the 2D DIOU (Distance Intersection over Union) loss function to 3D and develop a joint regression loss based on 3D_DIOU and SmoothL1 to optimize the similarity between detected and ground truth boxes. The experimental results on the KITTI dataset demonstrate that MMAF-Net effectively addresses the challenges posed by highly obscured or crowded scenes while maintaining real-time performance and improving the detection accuracy of smaller and more difficult objects that are occluded at far distances. • A novel object detection method is proposed integrating RGB images and LiDAR data. • This study tests the method on the KITTI dataset. • This method tackles highly obscured or crowded scenes for autonomous driving. • This method improves the detection accuracy of smaller and occluded targets far away. [ABSTRACT FROM AUTHOR]

Remote sensing scene understanding is crucial for extracting valuable information from high-resolution images, including object detection and classification. Traditional object detection methods face challenges in handling the diverse scales, orientations, and complex backgrounds present in remote sensing data. In this paper, we propose a novel remote sensing scene understanding system called multiscale-attention R-CNN (MSA R-CNN), which incorporates a super multiscale feature extraction network (SMENet) for enhanced feature extraction from multiscale images, an adaptive dynamic inner lateral (ADIL) connection module to tackle information loss in feature pyramid networks (FPN), and a distributed lightweight attention module (DLAM) to refine feature information processing. Furthermore, a new dataset combining the DIOR and DOTA datasets is introduced to extract the background information of detected objects and evaluate the proposed system's performance. MSA R-CNN achieved an mAP of 74.37% on the DIOR dataset when the gamma value was set to 0.2 and 81.97% on the DOTA dataset when the gamma value was set to 0.1 with the same learning rate, outperforming state-of-the-art models on both datasets. The proposed system demonstrates significant improvements in both object detection and background information extraction, providing a comprehensive solution for remote sensing scene understanding. • MSA R-CNN is proposed for remote sensing scene understanding. • SMENet enhances multi-scale feature extraction in MSA R-CNN. • ADIL module tackles FPN information loss. • DLAM refines feature information processing in FPN. • Achieved state-of-the-art performance on DIOR & DOTA dataset. [ABSTRACT FROM AUTHOR]

Grape leaf disease is a primary factor affecting grape yield and impeding the growth of the industrial economy. Currently, high-quality datasets are required for deep recognition models to effectively identify grape leaf diseases. In reality, the acquired image inevitably has the appearance of being low quality and unclear due to the influence of equipment and environment. The super-resolution method is an effective way to solve the problem. However, due to the complex model structure and learning strategies, existing super-resolution methods are computationally expensive and neglect the balance between global and local feature extraction. In this paper, a novel multiple attention transformer super-resolution (MATSR) method is proposed to solve these problems. Specifically, the method uses convolution to extract low-level features, which are then fed into the multiple attention shuffle transformer (MAST) module. The module consists of the shuffle transformer for global feature extraction, the constructed channel spatial attention block (CSAB) for local feature extraction, and the built reduisal feature block (RFB) for aggregating these two features. Meanwhile, considering the importance of the balance of global and local features for grape leaf disease recognition, a dynamic learning strategy is designed to balance these two features. Further, the abstract feature extraction layer combined the learned features with the low-level features. Finally, these abstracted features are fed into the pixels shuffle layer for super-resolution upsampling. Numerous comparative experiments are carried out on the collected high-resolution grape leaf disease images as well as the apple leaf pathological images. The results demonstrate the proposed method achieved the best PSNR of 38.152 dB and SSIM of 0.96 at scale ×2, and 35.896 dB and SSIM of 0.934 at scale ×3, respectively. Furthermore, four recognition models are experimented on the super-resolution grape disease dataset, among which ShuffleNet and Inception achieve the best accuracy of 98.28 % and 98.53 %, respectively. Therefore, this study can effectively aid grape growers in accurately identifying diseases and taking proactive measures. [ABSTRACT FROM AUTHOR]

Fire disaster causes damage to ships, pollute the environment, and threatens people's lives, so making early detection and reasonable decision is essential for avoiding catastrophic accidents. Smoke, as the main feature of fire, plays a significant role in early fire identification. However, challenged by the complex engine room (E/R) environment and shooting issues, the recognition performances of current smoke detection methods based on machine vision are unsatisfactory at the early fire stage. To address the issues, this paper proposes a proactive machine vision model based on the fusion of the transfer learning method and proactive perception technology for smoke detection. Firstly, a smoke images database is established based on similar environments (indoor and ship fire scenes) for the transfer learning module training, where more smoke features are extracted and learned from complex scenes. Afterward, real-time images are input into the trained model for local significance analysis which is applied as a triggering criterion for active smoke perception. When the significance indicator reaches the trigger condition, the proactive perception technology adopts reinforcement learning and proactive vision for further identifying detailed information about scenarios. By this processing, the internal and external parameters of the camera are adjusted to narrow and focus the targets of interest in the current scene. Finally, the ship E/R workshop and distribution box fire cases are selected to validate the proposed method. The experiment results indicate that the proposed model outperforms the existing techniques in accuracy as it has the potential to detect earlier smoke features. [ABSTRACT FROM AUTHOR]

Smoke detection is essential for disaster management and timely response to fires. However, popular deep learning-based algorithms ignore key traditional features such as color and texture, causing limitations in detection. Therefore, a lightweight edge-guided smoke detection network (ESmokeNet) is proposed in this paper. First, edge cues are refined from the low-level maps and integrated into the network features according to the color channels. Then, to increase the smoke feature extraction capability, we design a mutual context embedding module (MCE). It enhances multiscale context and multilevel semantic information based on attention and chiral features. Finally, the classification and localization tasks are decoupled to pursue the usability of network features directed by smoke edges, and an EdgeDet head is proposed. Extensive experimental and visualization results have confirmed that ESmokeNet can successfully capture smoke edges and that it has significant superiority over the existing state-of-the-art methods and specialized smoke detection algorithms. In addition, the test data show that our ESmokeNet is only 5.42 M and is a lightweight network suitable for smoke detection tasks. Code is available at https://github.com/jingjing-maker/ESmokeNet. • Propose a lightweight edge-guided smoke detection network, termed ESmokeNet. • Feature extraction is enhanced by a mutual context embedding module (MCE). • An anchor-free EdgeDet head pursues an effective fusion of edge cues and network. • ESmokeNet is only 5.42M and can be embedded on resource-constrained devices. [ABSTRACT FROM AUTHOR]

Hazy weather conditions feed serious haze effects, such as color and visibility degradations in the captured videos. Extensive research on video dehazing has been carried out. However, they lack to balance the spatial and high-level contextual information during restoration. Being a complex task in this paper, we propose a cross-stage recurrent feature sharing network to balance the competing objective. In the proposed network, a cross-stage feature merging module operates on learned features of the previous video frame using the proposed encoder decoder based generator architecture. Learned features fed into the same encoder decoder framework for the current video frame feature learning. This module benefit the network while restoring the current video frame. It shares the feature information between two consecutive frames as they have a very minute information change. Meanwhile, the design of our encoder decoder generator architecture is based on a multi-receptive resolution module, feature fusion module, and pixel-wise spatial attention module. These modules mainly focus on acquiring broad contextual information, fusing intermediate-stage features to increase the learning ability of the network, and capturing robust features while discarding unnecessary features. Experimental evaluation results and ablation study show that the proposed network is superior to existing state-of-the-art methods for video dehazing. • Cross-stage Recurrent Feature sharing Network is proposed for video dehazing. • Effective feature extraction strategy to aggregate contextual and spatial features. • Feature aggregation strategy to obtain robust combination of spatial features. • Cross stage feature sharing to balance the spatial and contextual information. [ABSTRACT FROM AUTHOR]

Sentiment analysis based on product online reviews is an important measure in grasping customer's satisfaction with products, which has the problem of inadequate feature extraction and difficulty in identifying the sentiment polarity of texts with irony. To solve the problem, this paper proposes a novel method of product feature sentiment analysis based on gate recurrent units and capsule neural network (GRU-CAP). Firstly, product online reviews are preprocessed to build a review data set. The feature extraction model of GRU-CAP is then used to extract the product features contained in the data set and generate the corresponding product feature index. After combining the feature index with the data set to form new input data, the recognition of positive/negative emotions, sarcastic/non-sarcastic semantics is carried out through the sentiment analysis model of GRU-CAP. Finally, the recognition results are integrated to achieve fine-grained product feature sentiment analysis. The effectiveness and practicability of the method is validated by an example product-mobile phone on an e-commerce platform. The results show that the method can extract product features and identify sentiment polarity from product online review texts with irony more accurately, so as to provide better product design support for enhancing customer satisfaction with products more effectively. [ABSTRACT FROM AUTHOR]

Accurate and robust localization is one of the most basic functional modules of mobile robots. A recent promising approach is to employ a camera that captures ground textures to achieve high-precision localization. However, low robustness and real-time performance are disadvantages of the system due to the time-consuming front-end feature extraction and search methods. To solve these problems, a comprehensive system that has high robustness and real-time performance is proposed, including offline mapping, global localization, local localization and map update. This paper summarizes a method for selecting the optimal front-end for different textures. A fast global localization method based on bag of words is proposed, which can quickly and effectively determine the local positioning area. Experiments are carried out on public datasets and actual indoor scenes to verify the practicability of the system. [ABSTRACT FROM AUTHOR]

Deep learning-based automatic sleep staging methods have been widely applied for sleep scoring and sleep diagnosis. However, most methods consider only a single temporal scale when dealing with sleep signals. Furthermore, these methods are limited to target only a single-age group or single dataset. In this paper, we propose a multi-scale temporally focused sleep staging model, MAGSleepNet, which can be used for multi-age groups simultaneously. MAGSleepNet consists of (1) a group age classification (GAC) module that can offer a preliminary screening on epidemic estimation of multiple age, (2) a dimensional expansion module (DEM) that can expand the dimension of the input signals, (3) a sequential multi-scale convolutional neural network (SMCNN) that extracts multi-scale features and short-time temporal information, and (4) sequence temporal encoder (STE) that extracts sequential temporal information. In addition, two auxiliary tasks are used to complement the short-time temporal information and to reassign the probabilities of different age groups to enhance the model robustness, respectively. The MAGSleepNet is evaluated in adult, child and infant, tested on MASS, CHAT, and CHFU datasets with accruacy of 86.7%, 80.1% and 66.5%, outperforming the state-of-the-art methods. Based on the excellent performance of the proposed method, it is expected to pave the way for automatic sleep staging methods with strong generalizability for multiple age groups. • A novel model is proposed for simultaneous sleep staging in multi-age groups. • The hierarchical architecture of model ensures the effective feature extraction. • An auxiliary task is used to provide temporal information on different scales. • Numerous experiments are conducted to evaluate the proposed model. [ABSTRACT FROM AUTHOR]

Accurate photovoltaic (PV) power forecast is crucial for carbon neutrality. Current researches on PV power forecast mainly focus on using temporal information from single PV station, and the spatial information in multiple PV power stations are often neglected. To address this problem, this paper introduces Moran index to verify the spatial autocorrelation of PV power for the first time, uses Granger causality test and transfer entropy to reveal the spatial information gain in PV power forecast for the first time, and proposes a novel spatial–temporal probabilistic PV forecast method using deep Graph Gated Recurrent Unit (GraphGRU) network-based spatial–temporal multi-task learning and Kernel Density Estimation (KDE). Deep GraphGRU combines the advantages of Graph Convolutional Network (GCN) in spatial feature extraction and the advantages of Gated Recurrent Unit (GRU) network in temporal feature extraction, and thus has strong ability to extract spatial–temporal information in historical data of multiple different PV power stations. Through GraphGRU, temporal dependency information extracted from historical data of multiple PV stations can promote each other to improve the forecast accuracy of each PV stations. KDE is used for estimating the joint probabilistic density function and giving the spatial–temporal probabilistic confidence interval of PV power. Experiments were performed in the five-year actual PV power data from 11 provinces of Belgium and the three-year solar irradiation data from 12 places in China to verify the superiorities of the proposed method. Comparison with conventional spatial–temporal and temporal forecast methods show that the proposed GraphGRU-based spatial–temporal forecast method can extract the spatial–temporal information from multiple PV power stations well and significantly outperform conventional temporal forecast methods. [ABSTRACT FROM AUTHOR]

Point cloud classification has emerged as a vital research area in several emerging applications, including robotics and autonomous driving. However, discriminative feature learning has long been a challenging issue owing to the irregularity and disorder of point clouds. Recently, although Transformer-based methods can achieve high accuracy in point cloud learning, plenty of Transformer layers bring huge computation and memory consumption. This paper presents a novel hierarchical local–global framework based on Transformer network for point clouds, named TNPC. TNPC contains two serial Stages implementing downsampling operation, and each Stage is composed of a local feature extracting (LFE) block and a global feature extracting (GFE) block, which can reduce computation and memory consumption obviously. LFE block consists of two parallel branches, namely Transformer branch and shared multilayer perceptron (SMLP) branch, which are designed to learn relevant feature of any two points and local high-dimensional semantic features of each point between sampling centroid and its neighborhoods, respectively. The proposed two parallel branches not only improve the feature extraction effect, but also reduce the computation and memory consumption. GFE block consists of a center point contact (CPC) module and a global point cloud transformer layer (PCTL) module, which can improve the effect of global features extracting without adding the number of parameters and computation. The performance of our method is validated experimentally on ModelNet40 and ScanObjectNN datasets. Our method improves the mean accuracy on each category (m A c c) to 91.6% and 79.8% on the ModelNet40 dataset and ScanObjectNN dataset, respectively. In terms of efficiency, our method leads to a significant reduction, with only 4.73MB parameters and only 1.91GB floating-point operations (FLOPs). Experimental results demonstrate that the proposed method achieves state-of-the-art performance on classification accuracy and efficiency. • A novel hierarchical local–global framework is proposed for 3D point clouds. • LFE block contains parallel Transformer branch and SMLP branch. • LFE block can learn local high-dimensional semantic features of each point. • GFE block is designed to extract the global feature of point cloud effectively. [ABSTRACT FROM AUTHOR]

With the rapid advancements in computer vision object detection networks, several detection networks based on remote sensing data have been proposed for oriented object detection. However, existing oriented networks have limitations in terms of feature extraction and oriented information extraction. To address these limitations, in this paper, we proposed an advanced adaptive points-based anchor-free oriented detection network called RA 2 DC-Net which includes a residual augment-convolutions (Res-AugConvs) module that uses residual structures and attention mechanisms to enhance feature focus information and minimize the loss of focus feature information. In addition, we proposed a method called adaptive deformable convolution (ADConv) to fine-tune custom convolution weights according to the features and generate high-quality adaptive offsets for accurate orientation information. Furthermore, we evaluated the performance of RA 2 DC-Net on two large-scale datasets: the Dataset for Object deTection in Aerial images (DOTA), the rotated object DetectIon in Optical Remote sensing images (DIOR-R) and the High-Resolution Ship Collection 2016 (HRSC2016). Experimental results demonstrated that RA 2 DC-Net achieved mAP values of 76.25%, 68.73% and 89.75% on the three datasets, respectively, with high AP values for various object classes on DOTA. • Decreased the continuous convolutions losses of oriented information in oriented object detection. • Enabling adaptive orientations of object features improves the precision of offset calculations. • Devising a new head structure facilitates focused oriented object features. [ABSTRACT FROM AUTHOR]

During recent years, deep convolutional neural networks (CNNs) have become increasingly popular in sensor-based human activity recognition (HAR), due to powerful feature extraction capacity. However, the required high computational cost has hindered their practical deployment in real-world applications. To address this issue, this paper presents a novel approach named ABCSearch to prune channels through an artificial bee colony (ABC) algorithm, which is a meta-heuristic optimization algorithm inspired by the foraging behavior of honeybees. Unlike traditional pruning strategies that prioritize channel importance, the proposed method seeks to identify the optimal number of channels at each layer, i.e., the optimal pruning structure. To lessen human interference, we transform the search problem of the optimal pruning structure into a simple optimization problem and solve it automatically through the ABC algorithm, which can significantly decrease both model size and computational cost without compromising performance. Specifically, we validate the effectiveness of ABCSearch for a standard five-layer CNN backbone architecture, which achieves 97.05%, 98.45%, 91.71%, 78.89% classification accuracies on four HAR benchmarks, i.e., UCI-HAR, WISDM, PAMAP2, UniMiB-SHAR. Our approach can yield 87.44%, 89.41%, 89.06%, 89.59% FLOPs reductions, and 89.54%, 88.70%, 88.38%, 90.45% less memory consumptions, as well as 4.4 × , 5.0 × , 5.8 × , 6.1 × practical speedups on real hardware devices, while maintaining comparable accuracy. [ABSTRACT FROM AUTHOR]

When performing data classification tasks, it often occurs to them the curse of dimensionality problem. To address the issue, a manifold learning method termed locally conjugated orthogonal subspace (LCOS) is put forward for dimensionality reduction or feature extraction in this paper. Note that point to feature space (P2S) distance contributes to mining local geometry information, both a local margin characterizing data apartness and a locally conjugated orthogonal constraint beneficial to removing data redundancy are well studied from the P2S distance metric. They are all exploited to model the proposed LCOS. Then, a low dimensional subspace can be explored by maximizing the P2S distance induced local margin under the constraint. Compared with some other related dimensionality reduction methods, experimental results on benchmark face and object data sets validate the performance of the proposed method. [ABSTRACT FROM AUTHOR]

Accurately predicting support resistance in coal mine working faces is a pressing matter in the field of intelligent mining research, which holds immense importance in enhancing the safety of coal production and the overall level of intelligent mining. The current research in this field primarily focuses on utilizing deep temporal sequence models for prediction. While there have been promising advancements in this area, concerns persist regarding the dependability of predicted outcomes. In fact, the movement of overburden in mining faces varies over time and space, and the changes in the support resistance that reflect the instability of overburden movement also have temporal and spatial characteristics. Therefore, in order to improve the reliability of model prediction, this paper proposes the application of deep spatiotemporal sequence models for support resistance prediction, and identifies the PredRNN and PredRNN++ models as suitable options. The exploration on the application of deep spatiotemporal sequence models for predicting the support resistance has been conducted. The proposed application involves the establishment of a fully working face area grid model with support number and mining cycle as coordinate units. Different from deep temporal sequence models that take the studied grid features as observations, the 25 adjacent spatiotemporal feature grids corresponding to the studied grid are considered as observations for the deep spatiotemporal sequence models. Based on the predicted three-dimensional tensor, the required predicted support resistance data can be obtained. Based on the complete support resistance data from 9 working faces in a coal mine in Datong mining area from 2009 to 2018, a comprehensive dataset was established to experimentally analyze the predictive performance and reliability of the LSTM, PredRNN, and PredRNN++ models. The experimental results demonstrate significant improvements in predictive performance and reliability of deep spatiotemporal sequence models compared to the current mainstream deep temporal sequence models. In terms of predictive performance, the PredRNN and PredRNN++ models have improved their feature extraction ability by incorporating the convolution operations to extract the spatial correlation of the input data. Furthermore, through the incorporation of gated units and nonlinear neurons into the storage units, coupled with the integration of vertical memory state transitions, the depth of the network has significantly increased, resulting in the enhancement of the sequence modeling capabilities. In terms of reliability of prediction results: The acceptable threshold for prediction error was set at 1000 kN. According to the average prediction results of all samples in the test set, it was found that the deep temporal sequence model was unable to accurately predict support resistance ahead. While the deep spatiotemporal sequence model demonstrated a predictable sequence length of 8. According to the prediction results of all samples in the test set, the accurate prediction rate increased from 30 %∼35 % to 65 %∼85 % and the maximum MAE value decreased from 10000 ∼ 12500kN to 6500 ∼ 9000kN in more than 20,000 test samples. The PredRNN model with an input sequence length of 50 achieves an accurate prediction rate of 85.48 % and a maximum MAE value of 6698kN, making it the most reliable prediction model. [ABSTRACT FROM AUTHOR]

Crowd density estimation is a task of intelligent applications, and its operation efficiency is very important. However, to obtain a better density estimation performance, most of the existing works often design larger and more complex network structures, which will result in them occupying considerable memory, time and other resources at runtime, and require the support of high-performance hardware platforms, which are difficult to apply in practice. In this paper, to overcome the above problems, we propose a lightweight dense crowd estimation method based on channel attention multi-scale feature fusion. Specifically, in the process of feature extraction, an efficient and lightweight convolution module (L-weight) is designed to extract crowd features in stages, which reduces the amount of network parameters and computing costs, and we capture multi-scale crowd information through the feature extraction network of pyramid structure, which solves the problem of uneven crowd scale in video images. In the process of feature fusion, a channel attention fusion module is designed, which weights and fuses the feature information of different scales, effectively fuses multi-scale information and suppresses useless information. In addition, we design a new loss function, which enhances the sensitivity of the crowd through the pixel space loss (L 2), counting loss (L C) and structural similarity loss (L S), to ensure the counting accuracy. Extensive experiments on four mainstream datasets demonstrate that compared with other state-of-the-art methods, our method achieves an optimal trade-off between counting performance and running speed, and is suitable for low-performance computing platforms such as embedded. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) technology has significant potential in Earth sciences, particularly in mineral identification for industrial exploration, geological mapping, and archaeological research. However, traditional methods are time-consuming, expensive, and complex. And existing mineral identification methods based on mineral photos face several critical challenges, including lack of consideration for natural image features captured in real environments, limitations of single-label classification which does not align with multi-mineral occurrences in nature, and growing computational complexity as the number of identifiable mineral labels increases. Therefore, this paper proposes an efficient mineral identification model based on multi-label image classification, focusing on natural environmental features. First, realistic feature datasets are created by simulating mineral photos in real environments. Then, the model uses the query-label (Query2Label) framework, with MaxViT-T (Multi-Axis Vision Transformer-Tiny) as the feature extraction network and the asymmetric loss function. Knowledge distillation is employed to improve identification accuracy while reducing computational complexity. The proposed model achieves an impressive average identification accuracy of 84.74% on a dataset of 495,756 mineral photos, surpassing existing models like ResNet-101, ML-GCN (Multi-Label Graph Convolutional Network), and SRN (Spatial Regularization Net). It maintains a lower parameter count and computational complexity. In the end, ablation experiments demonstrate the effectiveness of each optimization scheme. • Selecting image noise and color shift methods for realistic mineral data simulation. • Proposing an enhanced two-stage image multi-Label identification model framework. • Applying knowledge distillation for multi-label classification subtask decomposition. • Being with low computational complexity and high mineral identification accuracy. [ABSTRACT FROM AUTHOR]

• Generalization of binary sequences to multivariate improves data representation. • The weighted-probability of pattern occurrence is constructed based on complexity. • The MDS based on the novel dissimilarity measure is constructed. • The MDS creates better clustering results when compared with other methods. • The method has an acceptable running time and possesses robustness to noisy data. In this paper, a new dissimilarity measure for more accurate feature extraction and clustering is put forward. The method is proposed from the perspective of the weighted-probability distribution of dispersion patterns and their rank order statistics, where an effective quantization procedure is provided and the loss of information can be reduced. The proposed dissimilarity is applied in the multidimensional scaling (MDS) method to investigate simulated and reality-based signals. The comparative experiment shows that the clustering results of the proposed technique are clearer and more appropriate. State-of-the-art techniques and conventional methods are both included in the comparative experiments. In particular, for the heartbeat signals, it is discovered that the distribution of the weighted-probabilities of the dispersion patterns can discriminate subjects with different physiological conditions and exhibit visible changes of the subject's dynamical features when aging and disease attacks are taken place, which can be regarded as a microscopic insight of the dynamical mechanisms. [ABSTRACT FROM AUTHOR]

Recently meta-learning-based few-shot learning methods have been widely used for relation classification. Previous work reveals that meta-learning performs poorly in scenarios where the edge probability distribution of the target domain dataset appears to be significantly different from the source domain. In this paper, we enhance the meta-learning framework with high-dimensional semantic feature extraction and hyperplane projection metrics for meta-tasks. First, we enhance the focus of BERT on entity words by adding entity markers and vector pooling. After that, the high-dimensional semantic features of the support set are extracted and transformed into hyperplanes. Finally, we obtain the classification results by calculating the projection distance between the query sample and the hyperplane. In addition, we design a auxiliary function with a plane correction factor, which can better amplify the plane spacing and reduce the degree of category confusion, which is important for solving the problem of metric spatial loss. Experiments on two real-world few-shot datasets show that our model HPN is more effective in classifying few-shot relations in the same domain and domain-adapted scenarios. And HPN is more stable on NOTA tasks. • A hyperplane projection network is proposed for few shot relation classification. • Propose an effective entity encoder that extracts better features. • First apply a novel hyperplane module to generate class-level feature. [ABSTRACT FROM AUTHOR]