Search

Recent technological developments have enabled computers to identify and categorize facial expressions to determine a person's emotional state in an image or a video. This process, called "Facial Expression Recognition (FER)", has become one of the most popular research areas in computer vision. In recent times, deep FER systems have primarily concentrated on addressing two significant challenges: the problem of overfitting due to limited training data availability, and the presence of expression-unrelated variations, including illumination, head pose, image resolution, and identity bias. In this paper, a comprehensive survey is provided on deep FER, encompassing algorithms and datasets that offer insights into these intrinsic problems. Initially, this paper presents a detailed timeline showcasing the evolution of methods and datasets in deep facial expression recognition (FER). This timeline illustrates the progression and development of the techniques and data resources used in FER. Then, a comprehensive review of FER methods is introduced, including the basic principles of FER (components such as preprocessing, feature extraction and classification, and methods, etc.) from the pro-deep learning era (traditional methods using handcrafted features, i.e., SVM and HOG, etc.) to the deep learning era. Moreover, a brief introduction is provided related to the benchmark datasets (there are two categories: controlled environments (lab) and uncontrolled environments (in the wild)) used to evaluate different FER methods and a comparison of different FER models. Existing deep neural networks and related training strategies designed for FER, based on static images and dynamic image sequences, are discussed. The remaining challenges and corresponding opportunities in FER and the future directions for designing robust deep FER systems are also pinpointed. [ABSTRACT FROM AUTHOR]

This research proposes a face detection algorithm named LighterFace, which is aimed at enhancing detection speed to meet the demands of real-time community applications. Two pre-trained convolutional neural networks are combined, namely Cross Stage Partial Network (CSPNet), and ShuffleNetv2. Connecting the optimized network with Global Attention Mechanism (GAMAttention) extends the model to compensate for the accuracy loss caused by optimizing the network structure. Additionally, the learning rate of the detection model is dynamically updated using the cosine annealing method, which enhances the convergence speed of the model during training. This paper analyzes the training of the LighterFace model on the WiderFace dataset and a custom community dataset, aiming to classify faces in real-life community settings. Compared to the mainstream YOLOv5 model, LighterFace demonstrates a significant reduction in computational demands by 85.4% while achieving a 66.3% increase in detection speed and attaining a 90.6% accuracy in face detection. It is worth noting that LighterFace generates high-quality cropped face images, providing valuable inputs for subsequent face recognition models such as DeepID. Additionally, the LighterFace model is specifically designed to run on edge devices with lower computational capabilities. Its real-time performance on a Raspberry Pi 3B+ validates the results. [ABSTRACT FROM AUTHOR]

The accurate classification of landfill waste diversion plays a critical role in efficient waste management practices. Traditional approaches, such as visual inspection, weighing and volume measurement, and manual sorting, have been widely used but suffer from subjectivity, scalability, and labour requirements. In contrast, machine learning approaches, particularly Convolutional Neural Networks (CNN), have emerged as powerful deep learning models for waste detection and classification. This paper analyses VGG-16, InceptionResNetV2, DenseNet121, Inception V3, and MobileNetV2 models to classify real-life waste when trained on pristine and unadulterated materials, versus samples collected at a landfill site. When training on DiversionNet, the unadulterated material dataset with labels required for landfill modelling, classification accuracy was limited to 49.69% in the real environment. Using real-world samples in the newly formed RealWaste dataset showed that practical applications for deep learning in waste classification are possible, with Inception V3 reaching 89.19% classification accuracy on the full spectrum of labels required for accurate modelling. [ABSTRACT FROM AUTHOR]

Cybersecurity in modern vehicles has received increased attention from the research community in recent years. Intrusion Detection Systems (IDSs) are one of the techniques used to detect and mitigate cybersecurity risks. This paper proposes a novel implementation of an IDS for in-vehicle security networks based on the concept of multi-scale histograms, which capture the frequencies of message identifiers in CAN-bus in-vehicle networks. In comparison to existing approaches in the literature based on a single histogram, the proposed approach widens the informative context used by the IDS for traffic analysis by taking into consideration sequences of two and three CAN-bus messages to create multi-scale dictionaries. The histograms are created from windows of in-vehicle network traffic. A preliminary multi-scale histogram model is created using only legitimate traffic. Against this model, the IDS performs traffic analysis to create a feature space based on the correlation of the histograms. Then, the created feature space is given in input to a Convolutional Neural Network (CNN) for the identification of the windows of traffic where the attack is present. The proposed approach has been evaluated on two different public data sets achieving a very competitive performance in comparison to the literature. [ABSTRACT FROM AUTHOR]

This paper presents a novel low-cost integrated system prototype, called School Violence Detection system (SVD), based on a 2D Convolutional Neural Network (CNN). It is used for classifying and identifying automatically violent actions in educational environments based on shallow cost hardware. Moreover, the paper fills the gap of real datasets in educational environments by proposing a new one, called Daily School Break dataset (DSB), containing original videos recorded in an Italian high school yard. The proposed CNN has been pre-trained with an ImageNet model and a transfer learning approach. To extend its capabilities, the DSB was enriched with online images representing students in school environments. Experimental results analyze the classification performances of the SVD and investigate how it performs through the proposed DSB dataset. The SVD, which achieves a recognition accuracy of 95%, is considered computably efficient and low-cost. It could be adapted to other scenarios such as school arenas, gyms, playgrounds, etc. [ABSTRACT FROM AUTHOR]

Biometric technology is fast gaining pace as a veritable developmental tool. So far, biometric procedures have been predominantly used to ensure identity and ear recognition techniques continue to provide very robust research prospects. This paper proposes to identify and review present techniques for ear biometrics using certain parameters: machine learning methods, and procedures and provide directions for future research. Ten databases were accessed, including ACM, Wiley, IEEE, Springer, Emerald, Elsevier, Sage, MIT, Taylor & Francis, and Science Direct, and 1121 publications were retrieved. In order to obtain relevant materials, some articles were excused using certain criteria such as abstract eligibility, duplicity, and uncertainty (indeterminate method). As a result, 73 papers were selected for in-depth assessment and significance. A quantitative analysis was carried out on the identified works using search strategies: source, technique, datasets, status, and architecture. A Quantitative Analysis (QA) of feature extraction methods was carried out on the selected studies with a geometric approach indicating the highest value at 36%, followed by the local method at 27%. Several architectures, such as Convolutional Neural Network, restricted Boltzmann machine, auto-encoder, deep belief network, and other unspecified architectures, showed 38%, 28%, 21%, 5%, and 4%, respectively. Essentially, this survey also provides the various status of existing methods used in classifying related studies. A taxonomy of the current methodologies of ear recognition system was presented along with a publicly available occlussion and pose sensitive black ear image dataset of 970 images. The study concludes with the need for researchers to consider improvements in the speed and security of available feature extraction algorithms. [ABSTRACT FROM AUTHOR]

The musical key serves as a crucial element in a piece, offering vital insights into the tonal center, harmonic structure, and chord progressions while enabling tasks such as transposition and arrangement. Moreover, accurate key estimation finds practical applications in music recommendation systems and automatic music transcription, making it relevant across academic and industrial domains. This paper presents a comprehensive comparison between standard deep learning architectures and emerging vision transformers, leveraging their success in various domains. We evaluate their performance on a specific subset of the GTZAN dataset, analyzing six different deep learning models. Our results demonstrate that DenseNet, a conventional deep learning architecture, achieves remarkable accuracy of 91.64%, outperforming vision transformers. However, we delve deeper into the analysis to shed light on the temporal characteristics of each deep learning model. Notably, the vision transformer and SWIN transformer exhibit a slight decrease in overall performance (1.82% and 2.29%, respectively), yet they demonstrate superior performance in temporal metrics compared to the DenseNet architecture. The significance of our findings lies in their contribution to the field of musical key estimation, where accurate and efficient algorithms play a pivotal role. By examining the strengths and weaknesses of deep learning architectures and vision transformers, we can gain valuable insights for practical implementations, particularly in music recommendation systems and automatic music transcription. Our research provides a foundation for future advancements and encourages further exploration in this area. [ABSTRACT FROM AUTHOR]

Universal adversarial perturbations are image-agnostic and model-independent noise that, when added to any image, can mislead the trained deep convolutional neural networks into the wrong prediction. Since these universal adversarial perturbations can seriously jeopardize the security and integrity of practical deep learning applications, the existing techniques use additional neural networks to detect the existence of these noises at the input image source. In this paper, we demonstrate an attack strategy that, when activated by rogue means (e.g., malware, trojan), can bypass these existing countermeasures by augmenting the adversarial noise at the AI hardware accelerator stage. We demonstrate the accelerator-level universal adversarial noise attack on several deep learning models using co-simulation of the software kernel of the Conv2D function and the Verilog RTL model of the hardware under the FuseSoC environment. [ABSTRACT FROM AUTHOR]

The aim of predictive maintenance, within the field of prognostics and health management (PHM), is to identify and anticipate potential issues in the equipment before these become serious. The main challenge to be addressed is to assess the amount of time a piece of equipment will function effectively before it fails, which is known as remaining useful life (RUL). Deep learning (DL) models, such as Deep Convolutional Neural Networks (DCNN) and Long Short-Term Memory (LSTM) networks, have been widely adopted to address the task, with great success. However, it is well known that these kinds of black box models are opaque decision systems, and it may be hard to explain their outputs to stakeholders (experts in the industrial equipment). Due to the large number of parameters that determine the behavior of these complex models, understanding the reasoning behind the predictions is challenging. This paper presents a critical and comparative revision on a number of explainable AI (XAI) methods applied on time series regression models for PM. The aim is to explore XAI methods within time series regression, which have been less studied than those for time series classification. This study addresses three distinct RUL problems using three different datasets, each with its own unique context: gearbox, fast-charging batteries, and turbofan engine. Five XAI methods were reviewed and compared based on a set of nine metrics that quantify desirable properties for any XAI method. One of the metrics introduced in this study is a novel metric. The results show that Grad-CAM is the most robust method, and that the best layer is not the bottom one, as is commonly seen within the context of image processing. [ABSTRACT FROM AUTHOR]

In recent years, the rapid economic development of China, the increase of the urban population, the continuous growth of private car ownership, the uneven distribution of traffic flow, and the local congestion of the road network have caused traffic congestion. Traffic congestion has become an inevitable problem in the process of urban development, bringing hazards and hidden dangers to citizens' travel and urban development. The management of traffic congestion first lies in the accurate completion of the identification of road traffic status and the need to predict road congestion in the city, so as to improve the use rate of urban infrastructure road facilities and better alleviate road congestion. In this study, a deep spatial and temporal network model (DSGCN) for predicting traffic congestion status is proposed. First, our study divides the traffic network into grids, where each grid represents a different independent region. In this paper, the centroids of the grid regions are abstracted as nodes, and the dynamic correlations between the nodes are expressed in the form of adjacency matrix. Then, Graph Convolutional Neural Network is used to capture the spatial correlation between regions and a two-layer long and short-term feature model (DSTM) is used to capture the temporal correlation between regions. Finally, the DSGCN outperforms other baseline models and has higher accuracy for traffic congestion prediction as demonstrated by experiments on real PeMS datasets. [ABSTRACT FROM AUTHOR]

An electrocardiogram (ECG) is a unique representation of a person's identity, similar to fingerprints, and its rhythm and shape are completely different from person to person. Cloning and tampering with ECG-based biometric systems are very difficult. So, ECG signals have been used successfully in a number of biometric recognition applications where security is a top priority. The major challenges in the existing literature are (i) the noise components in the signals, (ii) the inability to automatically extract the feature set, and (iii) the performance of the system. This paper suggests a beat-based template matching deep learning (DL) technique to solve problems with traditional techniques. ECG beat denoising, R-peak detection, and segmentation are done in the pre-processing stage of this proposed methodology. These noise-free ECG beats are converted into gray-scale images and applied to the proposed deep-learning technique. A customized activation function is also developed in this work for faster convergence of the deep learning network. The proposed network can extract features automatically from the input data. The network performance is tested with a publicly available ECGID biometric database, and the proposed method is compared with the existing literature. The comparison shows that the proposed modified Siamese network authenticated biometrics have an accuracy of 99.85%, a sensitivity of 99.30%, a specificity of 99.85%, and a positive predictivity of 99.76%. The experimental results show that the proposed method works better than the state-of-the-art techniques. [ABSTRACT FROM AUTHOR]

Drowning is one of the top five causes of death for children aged 1–14 worldwide. According to data from the World Health Organization (WHO), drowning is the third most common reason for unintentional fatalities. Designing a drowning detection system is becoming increasingly necessary in order to ensure the safety of swimmers, particularly children. This paper presents a computer vision and deep learning-based early drowning detection approach. We utilized five convolutional neural network models and trained them on our data. These models are SqueezeNet, GoogleNet, AlexNet, ShuffleNet, and ResNet50. ResNet50 showed the best performance, as it achieved 100% prediction accuracy with a reasonable training time. When compared to other approaches, the proposed approach performed exceptionally well in terms of prediction accuracy and computational cost. [ABSTRACT FROM AUTHOR]

Efforts have been dedicated by researchers in the field of natural language processing (NLP) to detecting and combating fake news using an assortment of machine learning (ML) and deep learning (DL) techniques. In this paper, a review of the existing studies is conducted to understand and curtail the dissemination of fake news. Specifically, we conducted a benchmark study using a wide range of (1) classical ML algorithms such as logistic regression (LR), support vector machines (SVM), decision tree (DT), naive Bayes (NB), random forest (RF), XGBoost (XGB) and an ensemble learning method of such algorithms, (2) advanced ML algorithms such as convolutional neural networks (CNNs), bidirectional long short-term memory (BiLSTM), bidirectional gated recurrent units (BiGRU), CNN-BiLSTM, CNN-BiGRU and a hybrid approach of such techniques and (3) DL transformer-based models such as BERT b a s e and RoBERTa b a s e . The experiments are carried out using different pretrained word embedding methods across four well-known real-world fake news datasets—LIAR, PolitiFact, GossipCop and COVID-19—to examine the performance of different techniques across various datasets. Furthermore, a comparison is made between context-independent embedding methods (e.g., GloVe) and the effectiveness of BERT b a s e —contextualised representations in detecting fake news. Compared with the state of the art's results across the used datasets, we achieve better results by solely relying on news text. We hope this study can provide useful insights for researchers working on fake news detection. [ABSTRACT FROM AUTHOR]

In recent years, the presence of malware has been growing exponentially, resulting in enormous demand for efficient malware classification methods. However, the existing machine learning-based classifiers have high false positive rates and cannot effectively classify malware variants, packers, and obfuscation. To address this shortcoming, this paper proposes an efficient deep learning-based method named AIFS-IDL (Atanassov Intuitionistic Fuzzy Sets-Integrated Deep Learning), which uses static features to classify malware. The proposed method first extracts six types of features from the disassembly and byte files and then fuses them to solve the single-feature problem in traditional classification methods. Next, Atanassov's intuitionistic fuzzy set-based method is used to integrate the result of the three deep learning models, namely, GRU (Temporal Convolutional Network), TCN (Temporal Convolutional Network), and CNN (Convolutional Neural Networks), which improves the classification accuracy and generalizability of the classification model. The proposed method is verified by experiments and the results show that the proposed method can effectively improve the accuracy of malware classification compared to the existing methods. Experiments were carried out on the six types of features of malicious code and compared with traditional classification algorithms and ensemble learning algorithms. A variety of comparative experiments show that the classification accuracy rate of integrating multi-feature, multi-model aspects can reach 99.92%. The results show that, compared with other static classification methods, this method has better malware identification and classification ability. [ABSTRACT FROM AUTHOR]

Speech signals carry various bits of information relevant to the speaker such as age, gender, accent, language, health, and emotions. Emotions are conveyed through modulations of facial and vocal expressions. This paper conducts an empirical comparison of performances between the classical classifiers: Gaussian Mixture Model (GMM), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Artificial neural networks (ANN); and the deep learning classifiers, i.e., Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN), and Gated Recurrent Unit (GRU) in addition to the ivector approach for a text-independent speaker verification task in neutral and emotional talking environments. The deep models undergo hyperparameter tuning using the Grid Search optimization algorithm. The models are trained and tested using a private Arabic Emirati Speech Database, Ryerson Audio–Visual Database of Emotional Speech and Song dataset (RAVDESS) database, and a public Crowd-Sourced Emotional Multimodal Actors (CREMA) database. Experimental results illustrate that deep architectures do not necessarily outperform classical classifiers. In fact, evaluation was carried out through Equal Error Rate (EER) along with Area Under the Curve (AUC) scores. The findings reveal that the GMM model yields the lowest EER values and the best AUC scores across all datasets, amongst classical classifiers. In addition, the ivector model surpasses all the fine-tuned deep models (CNN, LSTM, and GRU) based on both evaluation metrics in the neutral, as well as the emotional speech. In addition, the GMM outperforms the ivector using the Emirati and RAVDESS databases. [ABSTRACT FROM AUTHOR]

Due to the absorption and scattering effects of light in water bodies and the non-uniformity and insufficiency of artificial illumination, underwater images often present various degradation problems, impacting their utility in underwater applications. In this paper, we propose a model-based underwater image simulation and learning-based underwater image enhancement method for coping with various degradation problems in underwater images. We first derive a simplified model for describing various degradation problems in underwater images, then propose a model-based image simulation method that can generate images with a wide range of parameter values. The proposed image simulation method also comes with an image-selection part, which helps to prune the simulation dataset so that it can serve as a training set for learning to enhance the targeted underwater images. Afterwards, we propose a convolutional neural network based on the encoder-decoder backbone to learn to enhance various underwater images from the simulated images. Experiments on simulated and real underwater images with different degradation problems demonstrate the effectiveness of the proposed underwater image simulation and enhancement method, and reveal the advantages of the proposed method in comparison with many state-of-the-art methods. [ABSTRACT FROM AUTHOR]

This paper proposes a supervised machine learning system to detect fake news in online sources published in Romanian. Additionally, this work presents a comparison of the obtained results by using recurrent neural networks based on long short-term memory and gated recurrent unit cells, a convolutional neural network, and a Bidirectional Encoder Representations from Transformers (BERT) model, namely RoBERT, a pre-trained Romanian BERT model. The deep learning architectures are compared with the results achieved by two classical classification algorithms: Naïve Bayes and Support Vector Machine. The proposed approach is based on a Romanian news corpus containing 25,841 true news items and 13,064 fake news items. The best result is over 98.20%, achieved by the convolutional neural network, which outperforms the standard classification methods and the BERT models. Moreover, based on irony detection and sentiment analysis systems, additional details are revealed about the irony phenomenon and sentiment analysis field which are used to tackle fake news challenges. [ABSTRACT FROM AUTHOR]

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. [ABSTRACT FROM AUTHOR]

Recently, deep learning to hash has extensively been applied to image retrieval, due to its low storage cost and fast query speed. However, there is a defect of insufficiency and imbalance when existing hashing methods utilize the convolutional neural network (CNN) to extract image semantic features and the extracted features do not include contextual information and lack relevance among features. Furthermore, the process of the relaxation hash code can lead to an inevitable quantization error. In order to solve these problems, this paper proposes deep hash with improved dual attention for image retrieval (DHIDA), which chiefly has the following contents: (1) this paper introduces the improved dual attention mechanism (IDA) based on the ResNet18 pre-trained module to extract the feature information of the image, which consists of the position attention module and the channel attention module; (2) when calculating the spatial attention matrix and channel attention matrix, the average value and maximum value of the column of the feature map matrix are integrated in order to promote the feature representation ability and fully leverage the features of each position; and (3) to reduce quantization error, this study designs a new piecewise function to directly guide the discrete binary code. Experiments on CIFAR-10, NUS-WIDE and ImageNet-100 show that the DHIDA algorithm achieves better performance. [ABSTRACT FROM AUTHOR]

Traditional time series forecasting techniques can not extract good enough sequence data features, and their accuracies are limited. The deep learning structure SeriesNet is an advanced method, which adopts hybrid neural networks, including dilated causal convolutional neural network (DC-CNN) and Long-short term memory recurrent neural network (LSTM-RNN), to learn multi-range and multi-level features from multi-conditional time series with higher accuracy. However, they didn't consider the attention mechanisms to learn temporal features. Besides, the conditioning method for CNN and RNN is not specific, and the number of parameters in each layer is tremendous. This paper proposes the conditioning method for two types of neural networks, and respectively uses the gated recurrent unit network (GRU) and the dilated depthwise separable temporal convolutional networks (DDSTCNs) instead of LSTM and DC-CNN for reducing the parameters. Furthermore, this paper presents the lightweight RNN-based hidden state attention module (HSAM) combined with the proposed CNN-based convolutional block attention module (CBAM) for time series forecasting. Experimental results show our model is superior to other models from the viewpoint of forecasting accuracy and computation efficiency. [ABSTRACT FROM AUTHOR]

The reasonable pricing of options can effectively help investors avoid risks and obtain benefits, which plays a very important role in the stability of the financial market. The traditional single option pricing model often fails to meet the ideal expectations due to its limited conditions. Combining an economic model with a deep learning model to establish a hybrid model provides a new method to improve the prediction accuracy of the pricing model. This includes the usage of real historical data of about 10,000 sets of CSI 300 ETF options from January to December 2020 for experimental analysis. Aiming at the prediction problem of CSI 300ETF option pricing, based on the importance of random forest features, the Convolutional Neural Network and Long Short-Term Memory model (CNN-LSTM) in deep learning is combined with a typical stochastic volatility Heston model and stochastic interests CIR model in parameter models. The dual hybrid pricing model of the call option and the put option of CSI 300ETF is established. The dual-hybrid model and the reference model are integrated with ridge regression to further improve the forecasting effect. The results show that the dual-hybrid pricing model proposed in this paper has high accuracy, and the prediction accuracy is tens to hundreds of times higher than the reference model; moreover, MSE can be as low as 0.0003. The article provides an alternative method for the pricing of financial derivatives. [ABSTRACT FROM AUTHOR]

In recent years, deep learning has been used in various applications including the classification of ship targets in inland waterways for enhancing intelligent transport systems. Various researchers introduced different classification algorithms, but they still face the problems of low accuracy and misclassification of other target objects. Hence, there is still a need to do more research on solving the above problems to prevent collisions in inland waterways. In this paper, we introduce a new convolutional neural network classification algorithm capable of classifying five classes of ships, including cargo, military, carrier, cruise and tanker ships, in inland waterways. The game of deep learning ship dataset, which is a public dataset originating from Kaggle, has been used for all experiments. Initially, the five pretrained models (which are AlexNet, VGG, Inception V3 ResNet and GoogleNet) were used on the dataset in order to select the best model based on its performance. Resnet-152 achieved the best model with an accuracy of 90.56%, and AlexNet achieved a lower accuracy of 63.42%. Furthermore, Resnet-152 was improved by adding a classification block which contained two fully connected layers, followed by ReLu for learning new characteristics of our training dataset and a dropout layer to resolve the problem of a diminishing gradient. For generalization, our proposed method was also tested on the MARVEL dataset, which consists of more than 10,000 images and 26 categories of ships. Furthermore, the proposed algorithm was compared with existing algorithms and obtained high performance compared with the others, with an accuracy of 95.8%, precision of 95.83%, recall of 95.80%, specificity of 95.07% and F1 score of 95.81%. [ABSTRACT FROM AUTHOR]

Face recognition algorithms based on deep learning methods have become increasingly popular. Most of these are based on highly precise but complex convolutional neural networks (CNNs), which require significant computing resources and storage, and are difficult to deploy on mobile devices or embedded terminals. In this paper, we propose several methods to improve the algorithms for face recognition based on a lightweight CNN, which is further optimized in terms of the network architecture and training pattern on the basis of MobileFaceNet. Regarding the network architecture, we introduce the Squeeze-and-Excitation (SE) block and propose three improved structures via a channel attention mechanism—the depthwise SE module, the depthwise separable SE module, and the linear SE module—which are able to learn the correlation of information between channels and assign them different weights. In addition, a novel training method for the face recognition task combined with an additive angular margin loss function is proposed that performs the compression and knowledge transfer of the deep network for face recognition. Finally, we obtained high-precision and lightweight face recognition models with fewer parameters and calculations that are more suitable for applications. Through extensive experiments and analysis, we demonstrate the effectiveness of the proposed methods. [ABSTRACT FROM AUTHOR]

Fall is a prominent issue due to its severe consequences both physically and mentally. Fall detection and prevention is a critical area of research because it can help elderly people to depend less on caregivers and allow them to live and move more independently. Using electrocardiograms (ECG) signals independently for fall detection and activity classification is a novel approach used in this paper. An algorithm has been proposed which uses pre-trained convolutional neural networks AlexNet and GoogLeNet as a classifier between the fall and no fall scenarios using electrocardiogram signals. The ECGs for both falling and no falling cases were obtained as part of the study using eight volunteers. The signals are pre-processed using an elliptical filter for signal noises such as baseline wander and power-line interface. As feature extractors, frequency-time representations (scalograms) were obtained by applying a continuous wavelet transform on the filtered ECG signals. These scalograms were used as inputs to the neural network and a significant validation accuracy of 98.08% was achieved in the first model. The trained model is able to distinguish ECGs with a fall activity from an ECG with a no fall activity with an accuracy of 98.02%. For the verification of the robustness of the proposed algorithm, our experimental dataset was augmented by adding two different publicly available datasets to it. The second model can classify fall, daily activities and no activities with an accuracy of 98.44%. These models were developed by transfer learning from the domain of real images to the medical images. In comparison to traditional deep learning approaches, the transfer learning not only avoids "reinventing the wheel," but also presents a lightweight solution to otherwise computationally heavy problems. [ABSTRACT FROM AUTHOR]

Human activity recognition (HAR) is a classification task that involves predicting the movement of a person based on sensor data. As we can see, there has been a huge growth and development of smartphones over the last 10–15 years—they could be used as a medium of mobile sensing to recognize human activity. Nowadays, deep learning methods are in a great demand and we could use those methods to recognize human activity. A great way is to build a convolutional neural network (CNN). HAR using Smartphone dataset has been widely used by researchers to develop machine learning models to recognize human activity. The dataset has two parts: training and testing. In this paper, we propose a hybrid approach to analyze and recognize human activity on the same dataset using deep learning method on cloud-based platform. We have applied principal component analysis on the dataset to get the most important features. Next, we have executed the experiment for all the features as well as the top 48, 92, 138, and 164 features. We have run all the experiments on Google Colab. In the experiment, for the evaluation of our proposed methodology, datasets are split into two different ratios such as 70–10–20% and 80–10–10% for training, validation, and testing, respectively. We have set the performance of CNN (70% training–10% validation–20% testing) with 48 features as a benchmark for our work. In this work, we have achieved maximum accuracy of 98.70% with CNN. On the other hand, we have obtained 96.36% accuracy with the top 92 features of the dataset. We can see from the experimental results that if we could select the features properly then not only could the accuracy be improved but also the training and testing time of the model. [ABSTRACT FROM AUTHOR]

The atrial fibrillation (AF) is one of the most well-known cardiac arrhythmias in clinical practice, with a prevalence of 1–2% in the community, which can increase the risk of stroke and myocardial infarction. The detection of AF electrocardiogram (ECG) can improve the early detection of diagnosis. In this paper, we have further developed a framework for processing the ECG signal in order to determine the AF episodes. We have implemented machine learning and deep learning algorithms to detect AF. Moreover, the experimental results show that better performance can be achieved with long short-term memory (LSTM) as compared to other algorithms. The initial experimental results illustrate that the deep learning algorithms, such as LSTM and convolutional neural network (CNN), achieved better performance (10%) as compared to machine learning classifiers, such as support vectors, logistic regression, etc. This preliminary work can help clinicians in AF detection with high accuracy and less probability of errors, which can ultimately result in reduction in fatality rate. [ABSTRACT FROM AUTHOR]

Facial beauty prediction (FBP) is a burgeoning issue for attractiveness evaluation, which aims to make assessment consistent with human opinion. Since FBP is a regression problem, to handle this issue, there are data-driven methods for finding the relations between facial features and beauty assessment. Recently, deep learning methods have shown its amazing capacity for feature representation and analysis. Convolutional neural networks (CNNs) have shown tremendous performance on facial recognition and comprehension, which are proved as an effective method for facial feature exploration. Lately, there are well-designed networks with efficient structures investigated for better representation performance. However, these designs concentrate on the effective block but do not build an efficient information transmission pathway, which led to a sub-optimal capacity for feature representation. Furthermore, these works cannot find the inherent correlations of feature maps, which also limits the performance. In this paper, an elaborate network design for FBP issue is proposed for better performance. A residual-in-residual (RIR) structure is introduced to the network for passing the gradient flow deeper, and building a better pathway for information transmission. By applying the RIR structure, a deeper network can be established for better feature representation. Besides the RIR network design, an attention mechanism is introduced to exploit the inner correlations among features. We investigate a joint spatial-wise and channel-wise attention (SCA) block to distribute the importance among features, which finds a better representation for facial information. Experimental results show our proposed network can predict facial beauty closer to a human's assessment than state-of-the-arts. [ABSTRACT FROM AUTHOR]

Determining an object measurement is a challenging task without having a well-defined reference. When a ruler is placed in the same plane of an object being measured it can serve as metric reference, thus a measurement system can be defined and calibrated to correlate actual dimensions with pixels contained in an image. This paper describes a system for non-contact object measurement by sensing and assessing the distinct spatial frequency of the graduations on a ruler. The approach presented leverages Deep Learning methods, specifically Mask Region proposal based Convolutional Neural Networks (R-CNN), for rulers' recognition and segmentation, as well as several other computer vision (CV) methods such as adaptive thresholding and template matching. We developed a heuristic analytical method for calibrating an image by applying several filters to extract the spatial frequencies corresponding to the ticks on a given ruler. We propose an automated in-plane optical scaling calibration system for non-contact measurement. [ABSTRACT FROM AUTHOR]

The problem of forged images has become a global phenomenon that is spreading mainly through social media. New technologies have provided both the means and the support for this phenomenon, but they are also enabling a targeted response to overcome it. Deep convolution learning algorithms are one such solution. These have been shown to be highly effective in dealing with image forgery derived from generative adversarial networks (GANs). In this type of algorithm, the image is altered such that it appears identical to the original image and is nearly undetectable to the unaided human eye as a forgery. The present paper investigates copy-move forgery detection using a fusion processing model comprising a deep convolutional model and an adversarial model. Four datasets are used. Our results indicate a significantly high detection accuracy performance (~95%) exhibited by the deep learning CNN and discriminator forgery detectors. Consequently, an end-to-end trainable deep neural network approach to forgery detection appears to be the optimal strategy. The network is developed based on two-branch architecture and a fusion module. The two branches are used to localize and identify copy-move forgery regions through CNN and GAN. [ABSTRACT FROM AUTHOR]

Human eye movement is one of the most important functions for understanding our surroundings. When a human eye processes a scene, it quickly focuses on dominant parts of the scene, commonly known as a visual saliency detection or visual attention prediction. Recently, neural networks have been used to predict visual saliency. This paper proposes a deep learning encoder-decoder architecture, based on a transfer learning technique, to predict visual saliency. In the proposed model, visual features are extracted through convolutional layers from raw images to predict visual saliency. In addition, the proposed model uses the VGG-16 network for semantic segmentation, which uses a pixel classification layer to predict the categorical label for every pixel in an input image. The proposed model is applied to several datasets, including TORONTO, MIT300, MIT1003, and DUT-OMRON, to illustrate its efficiency. The results of the proposed model are quantitatively and qualitatively compared to classic and state-of-the-art deep learning models. Using the proposed deep learning model, a global accuracy of up to 96.22% is achieved for the prediction of visual saliency. [ABSTRACT FROM AUTHOR]

In recent years, endless link prediction algorithms based on network representation learning have emerged. Network representation learning mainly constructs feature vectors by capturing the neighborhood structure information of network nodes for link prediction. However, this type of algorithm only focuses on learning topology information from the simple neighbor network node. For example, DeepWalk takes a random walk path as the neighborhood of nodes. In addition, such algorithms only take advantage of the potential features of nodes, but the explicit features of nodes play a good role in link prediction. In this paper, a link prediction method based on deep convolutional neural network is proposed. It constructs a model of the residual attention network to capture the link structure features from the sub-graph. Further study finds that the information flow transmission efficiency of the residual attention mechanism was not high, so a densely convolutional neural network model was proposed for link prediction. We evaluate our proposed method on four published data sets. The results show that our method is better than several other benchmark algorithms on link prediction. [ABSTRACT FROM AUTHOR]

Domain generation algorithms (DGAs) represent a class of malware used to generate large numbers of new domain names to achieve command-and-control (C2) communication between the malware program and its C2 server to avoid detection by cybersecurity measures. Deep learning has proven successful in serving as a mechanism to implement real-time DGA detection, specifically through the use of recurrent neural networks (RNNs) and convolutional neural networks (CNNs). This paper compares several state-of-the-art deep-learning implementations of DGA detection found in the literature with two novel models: a deeper CNN model and a one-dimensional (1D) Capsule Networks (CapsNet) model. The comparison shows that the 1D CapsNet model performs as well as the best-performing model from the literature. [ABSTRACT FROM AUTHOR]

This survey paper describes a literature review of deep learning (DL) methods for cyber security applications. A short tutorial-style description of each DL method is provided, including deep autoencoders, restricted Boltzmann machines, recurrent neural networks, generative adversarial networks, and several others. Then we discuss how each of the DL methods is used for security applications. We cover a broad array of attack types including malware, spam, insider threats, network intrusions, false data injection, and malicious domain names used by botnets. [ABSTRACT FROM AUTHOR]