Search

The accuracy of a predictive system is critical for predictive maintenance and to support the right decisions at the right times. Statistical models, such as ARIMA and SARIMA, are unable to describe the stochastic nature of the data. Neural networks, such as long short-term memory (LSTM) and the gated recurrent unit (GRU), are good predictors for univariate and multivariate data. The present paper describes a case study where the performances of long short-term memory and gated recurrent units are compared, based on different hyperparameters. In general, gated recurrent units exhibit better performance, based on a case study on pulp paper presses. The final result demonstrates that, to maximize the equipment availability, gated recurrent units, as demonstrated in the paper, are the best options.

In simple dyadic games such as rock, paper, scissors (RPS), people exhibit peculiar sequential dependencies across repeated interactions with a stable opponent. These regularities seem to arise from a mutually adversarial process of trying to outwit their opponent. What underlies this process, and what are its limits? Here, we offer a novel framework for formally describing and quantifying human adversarial reasoning in the rock, paper, scissors game. We first show that this framework enables a precise characterization of the complexity of patterned behaviors that people exhibit themselves, and appear to exploit in others. This combination allows for a quantitative understanding of human opponent modeling abilities. We apply these tools to an experiment in which people played 300 rounds of RPS in stable dyads. We find that although people exhibit very complex move dependencies, they cannot exploit these dependencies in their opponents, indicating a fundamental limitation in people’s capacity for adversarial reasoning. Taken together, the results presented here show how the rock, paper, scissors game allows for precise formalization of human adaptive reasoning abilities.

This research studied the strategies that players use in sequential adversarial games. We took the Rock-Paper-Scissors (RPS) game as an example and ran players in two experiments. The first experiment involved two humans, who played the RPS together for 100 times. Importantly, our payoff design in the RPS allowed us to differentiate between participants who used a random strategy from those who used a Nash strategy. We found that participants did not play in agreement with the Nash strategy, but rather, their behavior was closer to random. Moreover, the analyses of the participants’ sequential actions indicated heterogeneous cycle-based behaviors: some participants’ actions were independent of their past outcomes, some followed a well-known win-stay/lose-change strategy, and others exhibited the win-change/lose-stay behavior. To understand the sequential patterns of outcome-dependent actions, we designed probabilistic computer algorithms involving specific change actions (i.e., to downgrade or upgrade according to the immediate past outcome): the Win-Downgrade/Lose-Stay (WDLS) or Win-Stay/Lose-Upgrade (WSLU) strategies. Experiment 2 used these strategies against a human player. Our findings show that participants followed a win-stay strategy against the WDLS algorithm and a lose-change strategy against the WSLU algorithm, while they had difficulty in using an upgrade/downgrade direction, suggesting humans’ limited ability to detect and counter the actions of the algorithm. Taken together, our two experiments showed a large diversity of sequential strategies, where the win-stay/lose-change strategy did not describe the majority of human players’ dynamic behaviors in this adversarial situation.

To explore whether the brain contains pattern differences in the rock–paper–scissors (RPS) imagery task, this paper attempts to classify this task using fNIRS and deep learning. In this study, we designed an RPS task with a total duration of 25 min and 40 s, and recruited 22 volunteers for the experiment. We used the fNIRS acquisition device (FOIRE-3000) to record the cerebral neural activities of these participants in the RPS task. The time series classification (TSC) algorithm was introduced into the time-domain fNIRS signal classification. Experiments show that CNN-based TSC methods can achieve 97% accuracy in RPS classification. CNN-based TSC method is suitable for the classification of fNIRS signals in RPS motor imagery tasks, and may find new application directions for the development of brain–computer interfaces (BCI).

Spectrum sensing is an important function in radio frequency spectrum management and cognitive radio networks. Spectrum sensing is used by one wireless system (e.g., a secondary user) to detect the presence of a wireless service with higher priority (e.g., a primary user) with which it has to coexist in the radio frequency spectrum. If the wireless signal is detected, the second user system releases the given frequency to maintain the principle of not interfering. This paper proposes a machine learning implementation of spectrum sensing using the entropy measure as a feature vector. In the training phase, the information about the activity of the wireless service with higher priority is gathered, and the model is formed. In the classification phase, the wireless system compares the current sensing report to the created model to calculate the posterior probability and classify the sensing report into either the presence or absence of wireless service with higher priority. This paper proposes the novel application of the Fluctuation Dispersion Entropy (FDE) measure recently introduced in the research community as a feature vector to build the model and implement the classification. An improved implementation of the FDE (IFDE) is used to enhance the robustness to noise. IFDE is further enhanced with an adaptive method (AIFDE) to automatically select the hyper-parameter introduced in IFDE. Then, this paper combines the machine learning approach with the entropy measure approach, which are both recent developments in spectrum sensing research. The approach is compared to similar approaches in literature and the classical energy detection method using a generated radar signal data set with different conditions of SNR(dB) and fading conditions. The results show that the proposed approach is able to outperform the approaches from literature based on other entropy measures or the Energy Detector (ED) in a consistent way across different levels of SNR and fading conditions.

Detection of small targets in aerial images is still a difficult problem due to the low resolution and background-like targets. With the recent development of object detection technology, efficient and high-performance detector techniques have been developed. Among them, the YOLO series is a representative method of object detection that is light and has good performance. In this paper, we propose a method to improve the performance of small target detection in aerial images by modifying YOLOv5. The backbone is was modified by applying the first efficient channel attention module, and the channel attention pyramid method was proposed. We propose an efficient channel attention pyramid YOLO (ECAP-YOLO). Second, in order to optimize the detection of small objects, we eliminated the module for detecting large objects and added a detect layer to find smaller objects, reducing the computing power used for detecting small targets and improving the detection rate. Finally, we use transposed convolution instead of upsampling. Comparing the method proposed in this paper to the original YOLOv5, the performance improvement for the mAP was 6.9% when using the VEDAI dataset, 5.4% when detecting small cars in the xView dataset, 2.7% when detecting small vehicle and small ship classes from the DOTA dataset, and approximately 2.4% when finding small cars in the Arirang dataset.

The registration of optical imagery and 3D Light Detection and Ranging (LiDAR) point data continues to be a challenge for various applications in photogrammetry and remote sensing. In this paper, the framework employs a new registration primitive called virtual point (VP) that can be generated from the linear features within a LiDAR dataset including straight lines (SL) and curved lines (CL). By using an auxiliary parameter (λ), it is easy to take advantage of the accurate and fast calculation of the one-step registration transformation model. The transformation model parameters and λs can be calculated simultaneously by applying the least square method recursively. In urban areas, there are many buildings with different shapes. Therefore, the boundaries of buildings provide a large number of SL and CL features and selecting properly linear features and transforming into VPs can reduce the errors caused by the semi-discrete random characteristics of the LiDAR points. According to the result shown in the paper, the registration precision can reach the 1~2 pixels level of the optical images.

Using a sensor in variable lighting conditions, especially very low-light conditions, requires the application of image enhancement followed by denoising to retrieve correct information. The limits of such a process are explored in the present paper, with the objective of preserving the quality of enhanced images. The LIP (Logarithmic Image Processing) framework was initially created to process images acquired in transmission. The compatibility of this framework with the human visual system makes possible its application to images acquired in reflection. Previous works have established the ability of the LIP laws to perform a precise simulation of exposure time variation. Such a simulation permits the enhancement of low-light images, but a denoising step is required, realized by using a CNN (Convolutional Neural Network). A main contribution of the paper consists of using rigorous tools (metrics) to estimate the enhancement reliability in terms of noise reduction, visual image quality, and color preservation. Thanks to these tools, it has been established that the standard exposure time can be significantly reduced, which considerably enlarges the use of a given sensor. Moreover, the contribution of the LIP enhancement and denoising step are evaluated separately.

Concealed objects detection in terahertz imaging is an urgent need for public security and counter-terrorism. So far, there is no public terahertz imaging dataset for the evaluation of objects detection algorithms. This paper provides a public dataset for evaluating multi-object detection algorithms in active terahertz imaging. Due to high sample similarity and poor imaging quality, object detection on this dataset is much more difficult than on those commonly used public object detection datasets in the computer vision field. Since the traditional hard example mining approach is designed based on the two-stage detector and cannot be directly applied to the one-stage detector, this paper designs an image-based Hard Example Mining (HEM) scheme based on RetinaNet. Several state-of-the-art detectors, including YOLOv3, YOLOv4, FRCN-OHEM, and RetinaNet, are evaluated on this dataset. Experimental results show that the RetinaNet achieves the best mAP and HEM further enhances the performance of the model. The parameters affecting the detection metrics of individual images are summarized and analyzed in the experiments.

Unmanned surface vehicles (USVs) are receiving increasing attention in recent years from both academia and industry. To make a high-level autonomy for USVs, the environment situational awareness is a key capability. However, due to the richness of the features in marine environments, as well as the complexity of the environment influenced by sun glare and sea fog, the development of a reliable situational awareness system remains a challenging problem that requires further studies. This paper, therefore, proposes a new deep semantic segmentation model together with a Simple Linear Iterative Clustering (SLIC) algorithm, for an accurate perception for various maritime environments. More specifically, powered by the SLIC algorithm, the new segmentation model can achieve refined results around obstacle edges and improved accuracy for water surface obstacle segmentation. The overall structure of the new model employs an encoder–decoder layout, and a superpixel refinement is embedded before final outputs. Three publicly available maritime image datasets are used in this paper to train and validate the segmentation model. The final output demonstrates that the proposed model can provide accurate results for obstacle segmentation.

This paper proposes a measurement of risk (MOR) method to recognize risky driving behavior based on the trajectory data extracted from surveillance videos. Three types of risky driving behavior are studied in this paper, i.e., speed-unstable driving, serpentine driving, and risky car-following driving. The risky driving behavior recognition model contains an MOR-based risk evaluation model and an MOR threshold selection method. An MOR-based risk evaluation model is established for three types of risky driving behavior based on driving features to quantify collision risk. Then, we propose two methods, i.e., the distribution-based method and the boxplot-based method, to determine the threshold value of the MOR to recognize risky driving behavior. Finally, the trajectory data extracted from UAV videos are used to validate the proposed model. The impact of vehicle types is also taken into consideration in the model. The results show that there are significant differences between threshold values for cars and heavy trucks when performing speed-unstable driving and risky car-following driving. In addition, the difference between the proportion of recognized risky driving behavior in the testing dataset compared with that in the training dataset is limited to less than 3.5%. The recognition accuracy of risky driving behavior with the boxplot- and distribution-based methods are, respectively, 91% and 86%, indicating the validation of the proposed model. The proposed model can be widely applied to risky driving behavior recognition in video-based surveillance systems.

This paper is the result of the first-phase, inter-disciplinary work of a multi-disciplinary research project (“Urban pop-up housing environments and their potential as local innovation systems”) consisting of energy engineers and waste managers, landscape architects and spatial planners, innovation researchers and technology assessors. The project is aiming at globally analyzing and describing existing pop-up housings (PUH), developing modeling and assessment tools for sustainable, energy-efficient and socially innovative temporary housing solutions (THS), especially for sustainable and resilient urban structures. The present paper presents an effective application of hierarchical agglomerative clustering (HAC) for analyses of large datasets typically derived from field studies. As can be shown, the method, although well-known and successfully established in (soft) computing science, can also be used very constructively as a potential urban planning tool. The main aim of the underlying multi-disciplinary research project was to deeply analyze and structure THS and PUE. Multiple aspects are to be considered when it comes to the characterization and classification of such environments. A thorough (global) web survey of PUH and analysis of scientific literature concerning descriptive work of PUH and THS has been performed. Moreover, out of several tested different approaches and methods for classifying PUH, hierarchical clustering algorithms functioned well when properly selected metrics and cut-off criteria were applied. To be specific, the ‘Minkowski’-metric and the ‘Calinski-Harabasz’-criteria, as clustering indices, have shown the best overall results in clustering the inhomogeneous data concerning PUH. Several additional algorithms/functions derived from the field of hierarchical clustering have also been tested to exploit their potential in interpreting and graphically analyzing particular structures and dependencies in the resulting clusters. Hereby, (math.) the significance ‘S’ and (math.) proportion ‘P’ have been concluded to yield the best interpretable and comprehensible results when it comes to analyzing the given set (objects n = 85) of researched PUH-objects together with their properties (n > 190). The resulting easily readable graphs clearly demonstrate the applicability and usability of hierarchical clustering- and their derivative algorithms for scientifically profound building classification tasks in Urban Planning by effectively managing huge inhomogeneous building datasets.

In this paper, the robot arm Dobot Magician and the Raspberry Pi development platform were used to integrate image processing and robot-arm drawing. For this system, the Python language built into Raspberry Pi was used as the working platform, and the real-time image stream collected by the camera was used to determine the contour pixel coordinates of image objects. We then performed gray-scale processing, image binarization, and edge detection. This paper proposes an edge-point sequential arrangement method, which arranges the edge pixel coordinates of each object in an orderly manner and places them in a set. Orderly arrangement means that the pixels in the set are arranged counterclockwise to the closed curve of the object shape. This arrangement simplifies the complexity of subsequent image processing and calculation of the drawing path. The number of closed curves represents the number of strokes in the drawing of the manipulator. In order to reduce the complexity of the drawing of the manipulator, a fewer number of closed curves will be necessary. To achieve this goal, we not only propose the 8-NN (abbreviation for eight-nearest-neighbor) search, but also use to the 16-NN search and the 24-NN search methods. Drawing path points are then converted into drawing coordinates for the Dobot Magician through the Raspberry Pi platform. The structural design of the Dobot reduces the complexity of the experiment, and its attitude and positioning control can be accurately carried out through the built-in API function or the underlying communication protocol, which is more suitable for drawing applications than other fixed-point manipulators. Experimental results show that the 24-NN search method can effectively reduce the number of closed curves and the number of strokes drawn by the manipulator.

The number of consumers playing virtual reality games is booming. To speed up product iteration, the user experience team needs to collect and analyze unsatisfying experiences in time. In this paper, we aim to detect the unsatisfying experiences hidden in online reviews of virtual reality exergames using a deep learning method and find out the unmet psychological needs of users based on self-determination theory. Convolutional neural networks for sentence classification (textCNN) are used in this study to classify online reviews with unsatisfying experiences. For comparison, we set eXtreme gradient boosting (XGBoost) with lexical features as the baseline of machine learning. Term frequency-inverse document frequency (TF-IDF) is used to extract keywords from every set of classified reviews. The micro-F1 score of textCNN classifier is 90.00, which is better than 82.69 of XGBoost. The top 10 keywords of every set of reviews reflect relevant topics of unmet psychological needs. This paper explores the potential problems causing unsatisfying experiences and unmet psychological needs in virtual reality exergames through text mining and makes a supplement for experimental studies about virtual reality exergames.

Recent years have seen a rapid development of the Internet of Things (IoT) and the growth of autonomous robotic applications which are using network communications. Accordingly, an increasing advancement of intelligent devices with wireless sensors (that means autonomous robotic platforms) operating in challenging environments makes robots a tangible reality in the near future. Unfortunately, as a result of technical development, security problems emerge, especially when considering human–robot collaboration. Two abnormalities often compromise the basic security of collaborative robotic fleets: (a) Information faults and (b) system failures. This paper attempts to describe the methodology of a control framework design for secure robotic systems aided by the Internet of Things. The suggested concept represents a control system structure using blocks as the components. The structure is designed for the robots expected to interact with humans safely and act connected by communication channels. The properties of the components and relations between them are briefly described. The novelty of the proposed concept concerns the security mechanisms. The paper also categorizes two different modes of network attacks summarizing their causal effects on the human–robot collaboration systems. The issue of standardization is also raised. In particular, the works of the National Institute of Standards and Technology (NIST) and European Parliament (EP) on the security templates for communication channels are commented.

RGB–IR cross modality person re-identification (RGB–IR Re-ID) is an important task for video surveillance in poorly illuminated or dark environments. In addition to the common challenge of Re-ID, the large cross-modality variations between RGB and IR images must be considered. The existing RGB–IR Re-ID methods use different network structures to learn the global shared features associated with multi-modalities. However, most global shared feature learning methods are sensitive to background clutter, and contextual feature relationships are not considered among the mined features. To solve these problems, this paper proposes a dual-path attention network architecture MFCNet. SGA (Spatial-Global Attention) module embedded in MFCNet includes spatial attention and global attention branches to mine discriminative features. First, the SGA module proposed in this paper focuses on the key parts of the input image to obtain robust features. Next, the module mines the contextual relationships among features to obtain discriminative features and improve network performance. Finally, extensive experiments demonstrate that the performance of the network architecture proposed in this paper is better than that of state-of-the-art methods under various settings. In the all-search mode of the SYSU and RegDB data sets, the rank-1 accuracy reaches 51.64% and 69.76%, respectively.

Tsunamis are some of the most destructive natural disasters. Some proposed tsunami measurement and arrival prediction systems use a limited number of instruments, then judge the occurrence of the tsunami, forecast its arrival time, location and scale. Since there are a limited number of measurement instruments, there is a possibility that large prediction errors will occur. In order to solve this problem, a long-distance tsunami measurement system based on the binocular stereo vision principle is proposed in this paper. The measuring range is 4–20 km away from the system deployment site. In this paper, we will focus on describing the stereo matching method for the proposed system. This paper proposes a two-step matching method. It first performs fast sparse matching, and then complete high precision dense matching based on the results of the sparse matching. A matching descriptor based on the physical features of sea waves is proposed to solve the matching difficulty caused by the similarity of sea surface image textures. The relationship between disparity and the y coordinate is built to reduce the matching search range. Experiments were conducted on sea surface images with different shooting times and distances, the results verify the effectiveness of the presented method.

Outdoor vision sensing systems often struggle with poor weather conditions, such as snow and rain, which poses a great challenge to existing video desnowing and deraining methods. In this paper, we propose a novel video desnowing and deraining model that utilizes the salience information of moving objects to address this problem. First, we remove the snow and rain from the video by low-rank tensor decomposition, which makes full use of the spatial location information and the correlation between the three channels of the color video. Second, because existing algorithms often regard sparse snowflakes and rain streaks as moving objects, this paper injects salience information into moving object detection, which reduces the false alarms and missed alarms of moving objects. At the same time, feature point matching is used to mine the redundant information of moving objects in continuous frames, and a dual adaptive minimum filtering algorithm in the spatiotemporal domain is proposed by us to remove snow and rain in front of moving objects. Both qualitative and quantitative experimental results show that the proposed algorithm is more competitive than other state-of-the-art snow and rain removal methods.

Recent achievements on CNN (convolutional neural networks) and DNN (deep neural networks) researches provide a lot of practical applications on computer vision area. However, these approaches require construction of huge size of training data for learning process. This paper tries to find a way for continual learning which does not require prior high-cost training data construction by imitating a biological memory model. We employ SDR (sparse distributed representation) for information processing and semantic memory model, which is known as a representation model of firing patterns on neurons in neocortex area. This paper proposes a novel memory model to reflect remembrance of morphological semantics of visual input stimuli. The proposed memory model considers both memory process and recall process separately. First, memory process converts input visual stimuli to sparse distributed representation, and in this process, morphological semantic of input visual stimuli can be preserved. Next, recall process can be considered by comparing sparse distributed representation of new input visual stimulus and remembered sparse distributed representations. Superposition of sparse distributed representation is used to measure similarities. Experimental results using 10,000 images in MNIST (Modified National Institute of Standards and Technology) and Fashion-MNIST data sets show that the sparse distributed representation of the proposed model efficiently keeps morphological semantic of the input visual stimuli.

Existing wearable systems that use G-sensors to identify daily activities have been widely applied for medical, sports and military applications, while body temperature as an obvious physical characteristic that has rarely been considered in the system design and relative applications of HAR. In the context of the normalization of COVID-19, the prevention and control of the epidemic has become a top priority. Temperature monitoring plays an important role in the preliminary screening of the population for fever. Therefore, this paper proposes a wearable device embedded with inertial and temperature sensors that is used to apply human behavior recognition (HAR) to body surface temperature detection for body temperature monitoring and adjustment by evaluating recognition algorithms. The sensing system consists of an STM 32-based microcontroller, a 6-axis (accelerometer and gyroscope) sensor, and a temperature sensor to capture the original data from 10 individual participants under 4 different daily activity scenarios. Then, the collected raw data are pre-processed by signal standardization, data stacking and resampling. For HAR, several machine learning (ML) and deep learning (DL) algorithms are implemented to classify the activities. To compare the performance of different classifiers on the seven-dimensional dataset with temperature sensing signals, evaluation metrics and the algorithm running time are considered, and random forest (RF) is found to be the best-performing classifier with 88.78% recognition accuracy, which is higher than the case of the absence of temperature data (&lt, 78%). In addition, the experimental results show that participants’ body surface temperature in dynamic activities was lower compared to sitting, which can be associated with the possible missing fever population due to temperature deviations in COVID-19 prevention. According to different individual activities, epidemic prevention workers are supposed to infer the corresponding standard normal body temperature of a patient by referring to the specific values of the mean expectation and variance in the normal distribution curve provided in this paper.

Non-traditional machining (NTM) has gained significant attention in the last decade due to its ability to machine conventionally hard-to-machine materials. However, NTMs suffer from several disadvantages such as higher initial cost, lower material removal rate, more power consumption, etc. NTMs involve several process parameters, the appropriate tweaking of which is necessary to obtain economical and suitable results. However, the costly and time-consuming nature of the NTMs makes it a tedious and expensive task to manually investigate the appropriate process parameters. The NTM process parameters and responses are often not linearly related and thus, conventional statistical tools might not be enough to derive functional knowledge. Thus, in this paper, three popular machine learning (ML) methods (viz. linear regression, random forest regression and AdaBoost regression) are employed to develop predictive models for NTM processes. By considering two high-fidelity datasets from the literature on electro-discharge machining and wire electro-discharge machining, case studies are shown in the paper for the effectiveness of the ML methods. Linear regression is observed to be insufficient in accurately mapping the complex relationship between the process parameters and responses. Both random forest regression and AdaBoost regression are found to be suitable for predictive modelling of NTMs. However, AdaBoost regression is recommended as it is found to be insensitive to the number of regressors and thus is more readily deployable. Web of Science 9 11 art. no. 2015

Due to device limitations, small networks are necessary for some real-world scenarios, such as satellites and micro-robots. Therefore, the development of a network with both good performance and small size is an important area of research. Deep networks can learn well from large amounts of data, while manifold networks have outstanding feature representation at small sizes. In this paper, we propose an approach that exploits the advantages of deep networks and shallow Grassmannian manifold networks. Inspired by knowledge distillation, we use the information learned from convolutional neural networks to guide the training of the manifold networks. Our approach leads to a reduction in model size, which addresses the problem of deploying deep learning on resource-limited embedded devices. Finally, a series of experiments were conducted on four remote sensing scene classification datasets. The method in this paper improved the classification accuracy by 2.31% and 1.73% on the UC Merced Land Use and SIRIWHU datasets, respectively, and the experimental results demonstrate the effectiveness of our approach.

In many developed countries, the usage of artificial intelligence (AI) and machine learning (ML) has become important in paving the future path in how data is managed and secured in the small and medium enterprises (SMEs) sector. SMEs in these developed countries have created their own cyber regimes around AI and ML. This knowledge is tested daily in how these countries’ SMEs run their businesses and identify threats and attacks, based on the support structure of the individual country. Based on recent changes to the UK General Data Protection Regulation (GDPR), Brexit, and ISO standards requirements, machine learning cybersecurity (MLCS) adoption in the UK SME market has become prevalent and a good example to lean on, amongst other developed nations. Whilst MLCS has been successfully applied in many applications, including network intrusion detection systems (NIDs) worldwide, there is still a gap in the rate of adoption of MLCS techniques for UK SMEs. Other developed countries such as Spain and Australia also fall into this category, and similarities and differences to MLCS adoptions are discussed. Applications of how MLCS is applied within these SME industries are also explored. The paper investigates, using quantitative and qualitative methods, the challenges to adopting MLCS in the SME ecosystem, and how operations are managed to promote business growth. Much like security guards and policing in the real world, the virtual world is now calling on MLCS techniques to be embedded like secret service covert operations to protect data being distributed by the millions into cyberspace. This paper will use existing global research from multiple disciplines to identify gaps and opportunities for UK SME small business cyber security. This paper will also highlight barriers and reasons for low adoption rates of MLCS in SMEs and compare success stories of larger companies implementing MLCS. The methodology uses structured quantitative and qualitative survey questionnaires, distributed across an extensive participation pool directed to the SMEs’ management and technical and non-technical professionals using stratify methods. Based on the analysis and findings, this study reveals that from the primary data obtained, SMEs have the appropriate cybersecurity packages in place but are not fully aware of their potential. Secondary data collection was run in parallel to better understand how these barriers and challenges emerged, and why the rate of adoption of MLCS was very low. The paper draws the conclusion that help through government policies and processes coupled together with collaboration could minimize cyber threats in combatting hackers and malicious actors in trying to stay ahead of the game. These aspirations can be reached by ensuring that those involved have been well trained and understand the importance of communication when applying appropriate safety processes and procedures. This paper also highlights important funding gaps that could help raise cyber security awareness in the form of grants, subsidies, and financial assistance through various public sector policies and training. Lastly, SMEs’ lack of understanding of risks and impacts of cybercrime could lead to conflicting messages between cross-company IT and cybersecurity rules. Trying to find the right balance between this risk and impact, versus productivity impact and costs, could lead to UK SMES getting over these hurdles in this cyberspace in the quest for promoting the usage of MLCS. UK and Wales governments can use the research conducted in this paper to inform and adapt their policies to help UK SMEs become more secure from cyber-attacks and compare them to other developed countries also on the same future path.

Rapid and accurate grape leaf disease diagnosis is of great significance to its yield and quality of grape. In this paper, aiming at the identification of grape leaf diseases, a fast and accurate detection method based on fused deep features, extracted from a convolutional neural network (CNN), plus a support vector machine (SVM) is proposed. In the research, based on an open dataset, three types of state-of-the-art CNN networks, three kinds of deep feature fusion methods, seven species of deep feature layers, and a multi-class SVM classifier were studied. Firstly, images were resized to meet the input requirements of the CNN network, then, the deep features of the input images were extracted via the specific deep feature layer of the CNN network. Two kinds of deep features from different networks were then fused using different fusion methods to increase the effective classification feature information. Finally, a multi-class SVM classifier was trained with the fused deep features. The experimental results on the open dataset show that the fused deep features with any kind of fusion method can obtain a better classification performance than using a single type of deep feature. The direct concatenation of the Fc1000 deep feature extracted from ResNet50 and ResNet101 can achieve the best classification result compared with the other two fusion methods, and its F1 score is 99.81%. Furthermore, the SVM classifier trained using the proposed method can achieve a classification performance comparable to that of using the CNN model directly, but the training time is less than 1 s, which has an advantage over spending tens of minutes training a CNN model. The experimental results indicate that the method proposed in this paper can achieve fast and accurate identification of grape leaf diseases and meet the needs of actual agricultural production.

Like many other brain productions, language is a complex tool that helps individuals to communicate with each other. Many studies from computational linguistics aim to exhibit and understand the structures and content production. At present, a large list of contributions can describe and manage it with different levels of precision and applicability, but there is still a requirement for generative purposes. This paper is focused on stating the roots to understand language production from a combination of entropy and fractals. It is part of a larger work on seven rules that are intended to help build sentences automatically, in the context of dialogs with humans. As part of the scope of this paper, a set of dialogs are outlined and pre-processed. Three of the thermodynamic rules of language production are introduced and applied. Also, the communication implications and statistical evaluation are presented. From the results, a final analysis suggests that the exploration of fractals explanations of the entropy and entropy perspectives could provide a prospective insight for automatic sentence generation in natural language.

Self-driving cars, i.e., fully automated cars, will spread in the upcoming two decades, according to the representatives of automotive industries; owing to technological breakthroughs in the fourth industrial revolution, as the introduction of deep learning has completely changed the concept of automation. There is considerable research being conducted regarding object detection systems, for instance, lane, pedestrian, or signal detection. This paper specifically focuses on pedestrian detection while the car is moving on the road, where speed and environmental conditions affect visibility. To explore the environmental conditions, a pedestrian custom dataset based on Common Object in Context (COCO) is used. The images are manipulated with the inverse gamma correction method, in which pixel values are changed to make a sequence of bright and dark images. The gamma correction method is directly related to luminance intensity. This paper presents a flexible, simple detection system called Mask R-CNN, which works on top of the Faster R-CNN (Region Based Convolutional Neural Network) model. Mask R-CNN uses one extra feature instance segmentation in addition to two available features in the Faster R-CNN, called object recognition. The performance of the Mask R-CNN models is checked by using different Convolutional Neural Network (CNN) models as a backbone. This approach might help future work, especially when dealing with different lighting conditions.

This paper presents the empirical evaluation of the path-tracking accuracy of a three-wheeled omnidirectional mobile robot that is able to move in any direction while simultaneously changing its orientation. The mobile robot assessed in this paper includes a precise onboard LIDAR for obstacle avoidance, self-location and map creation, path-planning and path-tracking. This mobile robot has been used to develop several assistive services, but the accuracy of its path-tracking system has not been specifically evaluated until now. To this end, this paper describes the kinematics and path-planning procedure implemented in the mobile robot and empirically evaluates the accuracy of its path-tracking system that corrects the trajectory. In this paper, the information gathered by the LIDAR is registered to obtain the ground truth trajectory of the mobile robot in order to estimate the path-tracking accuracy of each experiment conducted. Circular and eight-shaped trajectories were assessed with different translational velocities. In general, the accuracy obtained in circular trajectories is within a short range, but the accuracy obtained in eight-shaped trajectories worsens as the velocity increases. In the case of the mobile robot moving at its nominal translational velocity, 0.3 m/s, the root mean square (RMS) displacement error was 0.032 m for the circular trajectory and 0.039 m for the eight-shaped trajectory, the absolute maximum displacement errors were 0.077 m and 0.088 m, with RMS errors in the angular orientation of 6.27° and 7.76°, respectively. Moreover, the external visual perception generated by these error levels is that the trajectory of the mobile robot is smooth, with a constant velocity and without perceiving trajectory corrections.

The information of tomato young fruits acquisition has an important impact on monitoring fruit growth, early control of pests and diseases and yield estimation. It is of great significance for timely removing young fruits with abnormal growth status, improving the fruits quality, and maintaining high and stable yields. Tomato young fruits are similar in color to the stems and leaves, and there are interference factors, such as fruits overlap, stems and leaves occlusion, and light influence. In order to improve the detection accuracy and efficiency of tomato young fruits, this paper proposes a method for detecting tomato young fruits with near color background based on improved Faster R-CNN with an attention mechanism. First, ResNet50 is used as the feature extraction backbone, and the feature map extracted is optimized through Convolutional Block Attention Module (CBAM). Then, Feature Pyramid Network (FPN) is used to integrate high-level semantic features into low-level detailed features to enhance the model sensitivity of scale. Finally, Soft Non-Maximum Suppression (Soft-NMS) is used to reduce the missed detection rate of overlapping fruits. The results show that the mean Average Precision (mAP) of the proposed method reaches 98.46%, and the average detection time per image is only 0.084 s, which can achieve the real-time and accurate detection of tomato young fruits. The research shows that the method in this paper can efficiently identify tomato young fruits, and provides a better solution for the detection of fruits with near color background.

Individual identification plays an important part in disease prevention and control, traceability of meat products, and improvement of agricultural false insurance claims. Automatic and accurate detection of cattle face is prior to individual identification and facial expression recognition based on image analysis technology. This paper evaluated the possibility of the cutting-edge object detection algorithm, RetinaNet, performing multi-view cattle face detection in housing farms with fluctuating illumination, overlapping, and occlusion. Seven different pretrained CNN models (ResNet 50, ResNet 101, ResNet 152, VGG 16, VGG 19, Densenet 121 and Densenet 169) were fine-tuned by transfer learning and re-trained on the dataset in the paper. Experimental results showed that RetinaNet incorporating the ResNet 50 was superior in accuracy and speed through performance evaluation, which yielded an average precision score of 99.8% and an average processing time of 0.0438 s per image. Compared with the typical competing algorithms, the proposed method was preferable for cattle face detection, especially in particularly challenging scenarios. This research work demonstrated the potential of artificial intelligence towards the incorporation of computer vision systems for individual identification and other animal welfare improvements.

This paper considers the use of a post metadata-based approach to identifying intentionally deceptive online content. It presents the use of an inherently explainable artificial intelligence technique, which utilizes machine learning to train an expert system, for this purpose. It considers the role of three factors (textual context, speaker background, and emotion) in fake news detection analysis and evaluates the efficacy of using key factors, but not the inherently subjective processing of post text itself, to identify deceptive online content. This paper presents initial work on a potential deceptive content detection tool and also, through the networks that it presents for this purpose, considers the interrelationships of factors that can be used to determine whether a post is deceptive content or not and their comparative importance.

The tremendous progress made in the field of deep learning allows us to accurately predict precipitation and avoid major and long-term disruptions to the entire socio-economic system caused by floods. This paper presents an LSTM–CP combined model formed by the Long Short-Term Memory (LSTM) network and Chebyshev polynomial (CP) as applied to the precipitation forecast of Yibin City. Firstly, the data are fed into the LSTM network to extract the time-series features. Then, the sequence features obtained are input into the BP (Back Propagation) neural network with CP as the excitation function. Finally, the prediction results are obtained. By theoretical analysis and experimental comparison, the LSTM–CP combined model proposed in this paper has fewer parameters, shorter running time, and relatively smaller prediction error than the LSTM network. Meanwhile, compared with the SVR model, ARIMA model, and MLP model, the prediction accuracy of the LSTM–CP combination model is significantly improved, which can aid relevant departments in making disaster response measures in advance to reduce disaster losses and promote sustainable development by providing them data support.

In the industrial field, the anthropomorphism of grasping robots is the trend of future development, however, the basic vision technology adopted by the grasping robot at this stage has problems such as inaccurate positioning and low recognition efficiency. Based on this practical problem, in order to achieve more accurate positioning and recognition of objects, an object detection method for grasping robot based on improved YOLOv5 was proposed in this paper. Firstly, the robot object detection platform was designed, and the wooden block image data set is being proposed. Secondly, the Eye-In-Hand calibration method was used to obtain the relative three-dimensional pose of the object. Then the network pruning method was used to optimize the YOLOv5 model from the two dimensions of network depth and network width. Finally, the hyper parameter optimization was carried out. The simulation results show that the improved YOLOv5 network proposed in this paper has better object detection performance. The specific performance is that the recognition precision, recall, mAP value and F1 score are 99.35%, 99.38%, 99.43% and 99.41% respectively. Compared with the original YOLOv5s, YOLOv5m and YOLOv5l models, the mAP of the YOLOv5_ours model has increased by 1.12%, 1.2% and 1.27%, respectively, and the scale of the model has been reduced by 10.71%, 70.93% and 86.84%, respectively. The object detection experiment has verified the feasibility of the method proposed in this paper.

As is well-known, defects precisely affect the lives and functions of the machines in which they occur, and even cause potentially catastrophic casualties. Therefore, quality assessment before mounting is an indispensable requirement for factories. Apart from the recognition accuracy, current networks suffer from excessive computing complexity, making it of great difficulty to deploy in the manufacturing process. To address these issues, this paper introduces binary networks into the area of surface defect detection for the first time, for the reason that binary networks prohibitively constrain weight and activation to +1 and −1. The proposed Bi-ShuffleNet and U-BiNet utilize binary convolution layers and activations in low bitwidth, in order to reach comparable performances while incurring much less computational cost. Extensive experiments are conducted on real-life NEU and Magnetic Tile datasets, revealing the least OPs required and little accuracy decline. When classifying the defects, Bi-ShuffleNet yields comparable results to counterpart networks, with at least 2× inference complexity reduction. Defect segmentation results indicate similar observations. Some network design rules in defect detection and binary networks are also summarized in this paper.

As bridge inspection becomes more advanced and more ubiquitous, artificial intelligence (AI) techniques, such as machine and deep learning, could offer suitable solutions to the nation’s problems of overdue bridge inspections. AI coupling with various data that can be captured by unmanned aerial vehicles (UAVs) enables fully automated bridge inspections. The key to the success of automated bridge inspection is a model capable of detecting failures from UAV data like images and films. In this context, this paper investigates the performances of state-of-the-art convolutional neural networks (CNNs) through transfer learning for crack detection in UAV-based bridge inspection. The performance of different CNN models is evaluated via UAV-based inspection of Skodsberg Bridge, located in eastern Norway. The low-level features are extracted in the last layers of the CNN models and these layers are trained using 19,023 crack and non-crack images. There is always a trade-off between the number of trainable parameters that CNN models need to learn for each specific task and the number of non-trainable parameters that come from transfer learning. Therefore, selecting the optimized amount of transfer learning is a challenging task and, as there is not enough research in this area, it will be studied in this paper. Moreover, UAV-based bridge inception images require specific attention to establish a suitable dataset as the input of CNN models that are trained on homogenous images. However, in the real implementation of CNN models in UAV-based bridge inspection images, there are always heterogeneities and noises, such as natural and artificial effects like different luminosities, spatial positions, and colors of the elements in an image. In this study, the effects of such heterogeneities on the performance of CNN models via transfer learning are examined. The results demonstrate that with a simplified image cropping technique and with minimum effort to preprocess images, CNN models can identify crack elements from non-crack elements with 81% accuracy. Moreover, the results show that heterogeneities inherent in UAV-based bridge inspection data significantly affect the performance of CNN models with an average 32.6% decrease of accuracy of the CNN models. It is also found that deeper CNN models do not provide higher accuracy compared to the shallower CNN models when the number of images for adoption to a specific task, in this case crack detection, is not large enough; in this study, 19,023 images and shallower models outperform the deeper models.

In this paper, a keyword detection scheme is proposed based on deep convolutional neural networks for personal information protection in document images. The proposed scheme is composed of key character detection and lexicon analysis. The first part is the key character detection developed based on RetinaNet and transfer learning. To find the key characters, RetinaNet, which is composed of convolutional layers featuring a pyramid network and two subnets, is exploited to detect key characters within the region of interest in a document image. After the key character detection, the second part is a lexicon analysis, which analyzes and combines several key characters to find the keywords. To train the model of RetinaNet, synthetic image generation and data augmentation are exploited to yield a large image dataset. To evaluate the proposed scheme, many document images are selected for testing, and two performance measurements, IoU (Intersection Over Union) and mAP (Mean Average Precision), are used in this paper. Experimental results show that the mAP rates of the proposed scheme are 85.1% and 85.84% for key character detection and keyword detection, respectively. Furthermore, the proposed scheme is superior to Tesseract OCR (Optical Character Recognition) software for detecting the key characters in document images. The experimental results demonstrate that the proposed method can effectively localize and recognize these keywords within noisy document images with Mandarin Chinese words.

As one of the most devastating disasters to pine forests, pine wilt disease (PWD) has caused tremendous ecological and economic losses in China. An effective way to prevent large-scale PWD outbreaks is to detect and remove the damaged pine trees at the early stage of PWD infection. However, early infected pine trees do not show obvious changes in morphology or color in the visible wavelength range, making early detection of PWD tricky. Unmanned aerial vehicle (UAV)-based hyperspectral imagery (HI) has great potential for early detection of PWD. However, the commonly used methods, such as the two-dimensional convolutional neural network (2D-CNN), fail to simultaneously extract and fully utilize the spatial and spectral information, whereas the three-dimensional convolutional neural network (3D-CNN) is able to collect this information from raw hyperspectral data. In this paper, we applied the residual block to 3D-CNN and constructed a 3D-Res CNN model, the performance of which was then compared with that of 3D-CNN, 2D-CNN, and 2D-Res CNN in identifying PWD-infected pine trees from the hyperspectral images. The 3D-Res CNN model outperformed the other models, achieving an overall accuracy (OA) of 88.11% and an accuracy of 72.86% for detecting early infected pine trees (EIPs). Using only 20% of the training samples, the OA and EIP accuracy of 3D-Res CNN can still achieve 81.06% and 51.97%, which is superior to the state-of-the-art method in the early detection of PWD based on hyperspectral images. Collectively, 3D-Res CNN was more accurate and effective in early detection of PWD. In conclusion, 3D-Res CNN is proposed for early detection of PWD in this paper, making the prediction and control of PWD more accurate and effective. This model can also be applied to detect pine trees damaged by other diseases or insect pests in the forest.

Human-robot interaction has received a lot of attention as collaborative robots became widely utilized in many industrial fields. Among techniques for human-robot interaction, collision identification is an indispensable element in collaborative robots to prevent fatal accidents. This paper proposes a deep learning method for identifying external collisions in 6-DoF articulated robots. The proposed method expands the idea of CollisionNet, which was previously proposed for collision detection, to identify the locations of external forces. The key contribution of this paper is uncertainty-aware knowledge distillation for improving the accuracy of a deep neural network. Sample-level uncertainties are estimated from a teacher network, and larger penalties are imposed for uncertain samples during the training of a student network. Experiments demonstrate that the proposed method is effective for improving the performance of collision identification.

Much applied research uses expert judgment as a primary or additional data source, thus the problem solved in this publication is relevant. Despite the expert’s experience and competence, the evaluation is subjective and has uncertainty in it. There are various reasons for this uncertainty, including the expert’s incomplete competence, the expert’s character and personal qualities, the expert’s attachment to the opinion of other experts, and the field of the task to be solved. This paper presents a new way to use the Bayesian method to reduce the uncertainty of an expert judgment by correcting the expert’s evaluation by the a posteriori mean function. The Bayesian method corrects the expert’s evaluation, taking into account the expert’s competence and accumulated long-term experience. Since the paper uses a continuous case of the Bayesian formula, perceived as a continuous approximation of experts’ evaluations, this is not only the novelty of this work, but also a new result in the theory of the Bayesian method and its application. The paper investigates various combinations of the probability density functions of a priori information and expert error. The results are illustrated by the example of the evaluation of distance learning courses.

When the shearer is cutting, the sound signal generated by the cutting drum crushing coal and rock contains a wealth of cutting status information. In order to effectively process the shearer cutting sound signal and accurately identify the cutting mode, this paper proposed a shearer cutting sound signal recognition method based on an improved complete ensemble empirical mode decomposition with adaptive noise (ICCEMDAN) and an improved grey wolf optimizer (IGWO) algorithm-optimized support vector machine (SVM). First, the approach applied ICEEMDAN to process the cutting sound signal and obtained several intrinsic mode function (IMF) components. It used the correlation coefficient to select the characteristic component. Meanwhile, this paper calculated the composite multi-scale permutation entropy (CMPE) of the characteristic components as the eigenvalue. Then, the method introduced a differential evolution algorithm and nonlinear convergence factor to improve the GWO algorithm. It used the improved GWO algorithm to realize the adaptive selection of SVM parameters and established a cutting sound signal recognition model. According to the proportioning plan, the paper made several simulation coal walls for cutting experiments and collected cutting sound signals for cutting pattern recognition. The experimental results show that the method proposed in this paper can effectively process the cutting sound signal of the shearer, and the average accuracy of the cutting pattern recognition model reached 97.67%.

Facing the big data wave, this study applied artificial intelligence to cite knowledge and find a feasible process to play a crucial role in supplying innovative value in environmental education. Intelligence agents of artificial intelligence and natural language processing (NLP) are two key areas leading the trend in artificial intelligence, this research adopted NLP to analyze the research topics of environmental education research journals in the Web of Science (WoS) database during 2011–2020 and interpret the categories and characteristics of abstracts for environmental education papers. The corpus data were selected from abstracts and keywords of research journal papers, which were analyzed with text mining, cluster analysis, latent Dirichlet allocation (LDA), and co-word analysis methods. The decisions regarding the classification of feature words were determined and reviewed by domain experts, and the associated TF-IDF weights were calculated for the following cluster analysis, which involved a combination of hierarchical clustering and K-means analysis. The hierarchical clustering and LDA decided the number of required categories as seven, and the K-means cluster analysis classified the overall documents into seven categories. This study utilized co-word analysis to check the suitability of the K-means classification, analyzed the terms with high TF-IDF wights for distinct K-means groups, and examined the terms for different topics with the LDA technique. A comparison of the results demonstrated that most categories that were recognized with K-means and LDA methods were the same and shared similar words, however, two categories had slight differences. The involvement of field experts assisted with the consistency and correctness of the classified topics and documents.

This paper focuses on generating distributed flocking strategies via imitation learning. The primary motivation is to improve the swarm robustness and achieve better consistency while respecting the communication constraints. This paper first proposes a quantitative metric of swarm robustness based on entropy evaluation. Then, the graph importance consistency is also proposed, which is one of the critical goals of the flocking task. Moreover, the importance-correlated directed graph convolutional networks (IDGCNs) are constructed for multidimensional feature extraction and structure-related aggregation of graph data. Next, by employing IDGCNs-based imitation learning, a distributed and scalable flocking strategy is obtained, and its performance is very close to the centralized strategy template while considering communication constraints. To speed up and simplify the training process, we train the flocking strategy with a small number of agents and set restrictions on communication. Finally, various simulation experiments are executed to verify the advantages of the obtained strategy in terms of realizing the swarm consistency and improving the swarm robustness. The results also show that the performance is well maintained while the scale of agents expands (tested with 20, 30, 40 robots).

The literature on coronaviruses counts more than 300,000 publications. Finding relevant papers concerning arbitrary queries is essential to discovery helpful knowledge. Current best information retrieval (IR) use deep learning approaches and need supervised training sets with labeled data, namely to know a priori the queries and their corresponding relevant papers. Creating such labeled datasets is time-expensive and requires prominent experts’ efforts, resources insufficiently available under a pandemic time pressure. We present a new self-supervised solution, called SUBLIMER, that does not require labels to learn to search on corpora of scientific papers for most relevant against arbitrary queries. SUBLIMER is a novel efficient IR engine trained on the unsupervised COVID-19 Open Research Dataset (CORD19), using deep metric learning. The core point of our self-supervised approach is that it uses no labels, but exploits the bibliography citations from papers to create a latent space where their spatial proximity is a metric of semantic similarity, for this reason, it can also be applied to other domains of papers corpora. SUBLIMER, despite is self-supervised, outperforms the Precision@5 (P@5) and Bpref of the state-of-the-art competitors on CORD19, which, differently from our approach, require both labeled datasets and a number of trainable parameters that is an order of magnitude higher than our.

Face detection, which is an effortless task for humans, is complex to perform on machines. The recent veer proliferation of computational resources is paving the way for frantic advancement of face detection technology. Many astutely developed algorithms have been proposed to detect faces. However, there is little attention paid in making a comprehensive survey of the available algorithms. This paper aims at providing fourfold discussions on face detection algorithms. First, we explore a wide variety of the available face detection algorithms in five steps, including history, working procedure, advantages, limitations, and use in other fields alongside face detection. Secondly, we include a comparative evaluation among different algorithms in each single method. Thirdly, we provide detailed comparisons among the algorithms epitomized to have an all-inclusive outlook. Lastly, we conclude this study with several promising research directions to pursue. Earlier survey papers on face detection algorithms are limited to just technical details and popularly used algorithms. In our study, however, we cover detailed technical explanations of face detection algorithms and various recent sub-branches of the neural network. We present detailed comparisons among the algorithms in all-inclusive and under sub-branches. We provide the strengths and limitations of these algorithms and a novel literature survey that includes their use besides face detection.

The identification of a person is a natural way to gain access to information or places. A face image is an essential element of visual validation. In this paper, we present the Card3DFace application, which captures a single-shot image of a person’s face. After reconstructing the 3D model of the head, the application generates several images from different perspectives, which, when printed on a card with a layer of lenticular lenses, produce a 3D visualization effect of the face. The image acquisition is achieved with a regular consumer 3D camera, either using plenoptic, stereo or time-of-flight technologies. This procedure aims to assist and improve the human visual recognition of ID cards and travel documents through an affordable and fast process while simultaneously increasing their security level. The whole system pipeline is analyzed and detailed in this paper. The results of the experiments performed with polycarbonate ID cards show that this end-to-end system is able to produce cards with realistic 3D visualization effects for humans.

This article belongs to the Section Computer and Engineering Science and Symmetry/Asymmetry. Speech intelligibility is a crucial element in oral communication that can be influenced by multiple elements, such as noise, channel characteristics, or speech disorders. In this paper, we address the task of speech intelligibility classification (SIC) in this last circumstance. Taking our previous works, a SIC system based on an attentional long short-term memory (LSTM) network, as a starting point, we deal with the problem of the inadequate learning of the attention weights due to training data scarcity. For overcoming this issue, the main contribution of this paper is a novel type of weighted pooling (WP) mechanism, called saliency pooling where the WP weights are not automatically learned during the training process of the network, but are obtained from an external source of information, the Kalinli’s auditory saliency model. In this way, it is intended to take advantage of the apparent symmetry between the human auditory attention mechanism and the attentional models integrated into deep learning networks. The developed systems are assessed on the UA-speech dataset that comprises speech uttered by subjects with several dysarthria levels. Results show that all the systems with saliency pooling significantly outperform a reference support vector machine (SVM)-based system and LSTM-based systems with mean pooling and attention pooling, suggesting that Kalinli’s saliency can be successfully incorporated into the LSTM architecture as an external cue for the estimation of the speech intelligibility level. The work leading to these results has been supported by the Spanish Ministry of Economy, Industry and Competitiveness through TEC2017-84395-P (MINECO) and TEC2017-84593-C2-1-R (MINECO) projects (AEI/FEDER, UE), and the Universidad Carlos III de Madrid under Strategic Action 2018/00071/001.

As a complex task, robot sorting has become a research hotspot. In order to enable robots to perform simple, efficient, stable and accurate sorting operations for stacked multi-objects in unstructured scenes, a robot multi-object sorting system is built in this paper. Firstly, the training model of rotating target detection is constructed, and the placement state of five common objects in unstructured scenes is collected as the training set for training. The trained model is used to obtain the position, rotation angle and category of the target object. Then, the instance segmentation model is constructed, and the same data set is made, and the instance segmentation network model is trained. Then, the optimized Mask R-CNN instance segmentation network is used to segment the object surface pixels, and the upper surface point cloud is extracted to calculate the normal vector. Then, the angle obtained by the normal vector of the upper surface and the rotation target detection network is fused with the normal vector to obtain the attitude of the object. At the same time, the grasping order is calculated according to the average depth of the surface. Finally, after the obtained object posture, category and grasping sequence are fused, the performance of the rotating target detection network, the instance segmentation network and the robot sorting system are tested on the established experimental platform. Based on this system, this paper carried out an experiment on the success rate of object capture in a single network and an integrated network. The experimental results show that the multi-object sorting system based on deep learning proposed in this paper can sort stacked objects efficiently, accurately and stably in unstructured scenes.

Video synthetic aperture radar (ViSAR) can provide long-time surveillance of a region of interest (ROI), which is one of the hotspot directions in the SAR field. In order to better display ViSAR, a high resolution and high frame rate are needed. Azimuth integration angle and sub-aperture overlapping ratio, which determine the image resolution and frame rate, respectively, are analyzed in depth in this paper. For SAR imaging algorithm, polar format algorithm (PFA) is applied, which not only has high efficiency but is also easier to integrate with autofocus algorithms. Due to sensitivity to motion error, it is very difficult to obtain satisfactory focus quality, especially for SAR systems with a high carrier frequency. The three-step motion compensation (MOCO) proposed in this paper, which combines GPS-based MOCO, map-drift (MD) and phase gradient autofocus (PGA), can effectively compensate for motion error, especially for short wavelengths. In ViSAR, problems such as jitter, non-uniform grey scale and low image signal noise ratio (SNR) between different aspects images also need to be considered, so a ViSAR generation method is proposed to solve the above problems. Finally, the results of ViSAR in THz and Ku band demonstrate the effectiveness and practicability of the proposed method.

The rise of deep learning technology has broadly promoted the practical application of artificial intelligence in production and daily life. In computer vision, many human-centered applications, such as video surveillance, human-computer interaction, digital entertainment, etc., rely heavily on accurate and efficient human pose estimation techniques. Inspired by the remarkable achievements in learning-based 2D human pose estimation, numerous research studies are devoted to the topic of 3D human pose estimation via deep learning methods. Against this backdrop, this paper provides an extensive literature survey of recent literature about deep learning methods for 3D human pose estimation to display the development process of these research studies, track the latest research trends, and analyze the characteristics of devised types of methods. The literature is reviewed, along with the general pipeline of 3D human pose estimation, which consists of human body modeling, learning-based pose estimation, and regularization for refinement. Different from existing reviews of the same topic, this paper focus on deep learning-based methods. The learning-based pose estimation is discussed from two categories: single-person and multi-person. Each one is further categorized by data type to the image-based methods and the video-based methods. Moreover, due to the significance of data for learning-based methods, this paper surveys the 3D human pose estimation methods according to the taxonomy of supervision form. At last, this paper also enlists the current and widely used datasets and compares performances of reviewed methods. Based on this literature survey, it can be concluded that each branch of 3D human pose estimation starts with fully-supervised methods, and there is still much room for multi-person pose estimation based on other supervision methods from both image and video. Besides the significant development of 3D human pose estimation via deep learning, the inherent ambiguity and occlusion problems remain challenging issues that need to be better addressed.

The Fourth Industrial Revolution drives industries from traditional manufacturing to the smart manufacturing approach. In this transformation, existing equipment, processes, or devices are retrofitted with some sensors and other cyber-physical systems (CPS), and adapted towards digital production, which is a blend of critical enabling technologies. In the current scenario of Industry 4.0, industries are shaping themselves towards the development of customized and cost-effective processes to satisfy customer needs with the aid of a digital twin framework, which enables the user to monitor, simulate, control, optimize, and identify defects and trends within, ongoing process, and reduces the chances of human prone errors. This paper intends to make an appraisal of the literature on the digital twin (DT) framework in the domain of smart manufacturing with the aid of critical enabling technologies such as data-driven systems, machine learning and artificial intelligence, and deep learning. This paper also focuses on the concept, evolution, and background of digital twin and the benefits and challenges involved in its implementation. The Scopus and Web of Science databases from 2016 to 2021 were considered for the bibliometric analysis and used to study and analyze the articles that fall within the research theme. For the systematic bibliometric analysis, a novel approach known as Proknow-C was employed, including a series of procedures for article selection and filtration from the existing databases to get the most appropriate articles aligned with the research theme. Additionally, the authors performed statistical and network analyses on the articles within the research theme to identify the most prominent research areas, journal/conference, and authors in the field of a digital twin. This study identifies the current scenarios, possible research gaps, challenges in implementing DT, case studies and future research goals within the research theme.

Machine learning-based systems are gaining interest in the field of medicine, mostly in medical imaging and diagnosis. In this paper, we address the problem of automatic cerebral microbleeds (CMB) detection in magnetic resonance images. It is challenging due to difficulty in distinguishing a true CMB from its mimics, however, if successfully solved, it would streamline the radiologists work. To deal with this complex three-dimensional problem, we propose a machine learning approach based on a 2D Faster RCNN network. We aimed to achieve a reliable system, i.e., with balanced sensitivity and precision. Therefore, we have researched and analysed, among others, impact of the way the training data are provided to the system, their pre-processing, the choice of model and its structure, and also the ways of regularisation. Furthermore, we also carefully analysed the network predictions and proposed an algorithm for its post-processing. The proposed approach enabled for obtaining high precision (89.74%), sensitivity (92.62%), and F1 score (90.84%). The paper presents the main challenges connected with automatic cerebral microbleeds detection, its deep analysis and developed system. The conducted research may significantly contribute to automatic medical diagnosis.