Search

The motion of an object or camera platform makes the acquired image blurred. This degradation is a major reason to obtain a poor-quality image from an imaging sensor. Therefore, developing an efficient deep-learning-based image processing method to remove the blur artifact is desirable. Deep learning has recently demonstrated significant efficacy in image deblurring, primarily through convolutional neural networks (CNNs) and Transformers. However, the limited receptive fields of CNNs restrict their ability to capture long-range structural dependencies. In contrast, Transformers excel at modeling these dependencies, but they are computationally expensive for high-resolution inputs and lack the appropriate inductive bias. To overcome these challenges, we propose an Efficient Hybrid Network (EHNet) that employs CNN encoders for local feature extraction and Transformer decoders with a dual-attention module to capture spatial and channel-wise dependencies. This synergy facilitates the acquisition of rich contextual information for high-quality image deblurring. Additionally, we introduce the Simple Feature-Embedding Module (SFEM) to replace the pointwise and depthwise convolutions to generate simplified embedding features in the self-attention mechanism. This innovation substantially reduces computational complexity and memory usage while maintaining overall performance. Finally, through comprehensive experiments, our compact model yields promising quantitative and qualitative results for image deblurring on various benchmark datasets. [ABSTRACT FROM AUTHOR]

Advances in deep learning and computer vision techniques have made impacts in the field of remote sensing, enabling efficient data analysis for applications such as land cover classification and change detection. Convolutional neural networks (CNNs) and transformer architectures have been utilized in visual perception algorithms due to their effectiveness in analyzing local features and global context. In this paper, we propose a hybrid transformer architecture that consists of a CNN-based encoder and transformer-based decoder. We propose a feature adjustment module that refines the multiscale feature maps extracted from an EfficientNet backbone network. The adjusted feature maps are integrated into the transformer-based decoder to perform the semantic segmentation of the remote sensing images. This paper refers to the proposed encoder–decoder architecture as a semantic feature adjustment network (SFA-Net). To demonstrate the effectiveness of the SFA-Net, experiments were thoroughly conducted with four public benchmark datasets, including the UAVid, ISPRS Potsdam, ISPRS Vaihingen, and LoveDA datasets. The proposed model achieved state-of-the-art accuracy on the UAVid, ISPRS Vaihingen, and LoveDA datasets for the segmentation of the remote sensing images. On the ISPRS Potsdam dataset, our method achieved comparable accuracy to the latest model while reducing the number of trainable parameters from 113.8 M to 10.7 M. [ABSTRACT FROM AUTHOR]

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

AbstractAs known, ceramics possess desirable strength, hardness and low density, which have been extensively applied in engineering fields. However, the inherent brittleness makes the ceramics present poor toughness and consequently impact resistance as well. In recent decades, some ingenious architectures of biomaterials employed to improve the mechanical properties of brittle bulk materials, such as laminate and brick-mud, have become increasingly popular. Here, in order to further enhance the impact resistance of ceramics, a kind of hybrid ceramic-based composite structure is designed according to the gradient feature in dactyl club of mantis shrimp. The impact resistance performances of bulk, laminate, brick-mud and hybrid structures are investigated by finite element simulation. The results show that the hybrid structure can effectively avoid catastrophic failure, thereby having large scope of deformation area to store energy. Moreover, the damage mass of impactor for the hybrid plate is the largest due to the beginning hard collision between impactor and target plate and long cumulative damaging time of impactor. As a result, with the highest internal energy and eroding kinetic energy of impactor simultaneously, the hybrid structure dissipates the most energy of impactor. Further, under extreme working conditions of oblique and high velocity impact, the hybrid structure still exhibits optical impact resistance performance. [ABSTRACT FROM AUTHOR]

Digital coding and programmable metasurfaces stand out as promising candidates in future wireless communications due to their low costs and fast‐action reprogrammable capability. However, 2 significant obstacles, namely integration and power consumption, must be addressed before large‐scale engineering application of the programmable metasurface. This work proposes an easy‐integration and energy‐saving meta‐system and demonstrates its application in wireless communications. The meta‐system features a hybrid architecture and consists of a programmable feed array and a metasurface lens. The feed array comprises 3 subarrays, in which active elements are doped with passive elements and interconnected configurations are used to efficiently reduce the number of active elements by one‐third. The meta‐lens acts as a passive phase shifter to enhance the beamforming performance, further reducing the number of active elements. Moreover, the meta‐system can switch between different modes to achieve various functions. In particular, the maximum power consumption of the meta‐system is only 54 mW, which can be considered nearly passive. A wireless communication experiment is presented, where the meta‐system simultaneously serves as the direct digital modulator and the transmitting antenna. Owing to the low cost, high integration, and nearly passive and programmable features, the meta‐system has remarkable potential for future applications in wireless communications. [ABSTRACT FROM AUTHOR]

Passive RFID has the advantages of rapid identification of multi-target objects and low implementation cost. It is the most critical technology in the Industrial Internet of Things information-gathering layer and is extensively applied in various industries, such as smart production, asset management, and monitoring. The signal collision caused by the communication between the reader/writer and tags sharing the same wireless channel has caused a series of problems, such as the reduction of the identification efficiency of the reader/writer and the increment of the missed reading rate, thus restricting the further development of RFID. At present, many hybrid anti-collision algorithms integrate the advantages of Aloha and TS algorithms to optimize RFID system performance, but these solutions also suffer from performance bottlenecks. In order to break through such performance bottleneck, based on the ISE-BS algorithm, we combined the sub-frame observation mechanism and the Q value adjustment strategy and proposed two hybrid anti-collision algorithms. The experimental results show that the two algorithms proposed in this paper have obvious advantages in system throughput, time efficiency and other metrics, surpassing existing UHF RFID anti-collision algorithms. [ABSTRACT FROM AUTHOR]

Current unsupervised industrial product surface anomaly detection methods suffer from poor reconstructed image quality and difficulty in detecting low-contrast anomalies, resulting in low anomaly detection accuracy. To address the above problems, we propose an unsupervised masked hybrid convolutional Transformer anomaly detection model, which forces the model to predict missing or edited regions based on unmasked information by introducing a mask reconstruction strategy, and utilises convolutional blocks and Transformer self-attention mechanism to extract the local features and global context of the image at different resolutions, enhancing the model’s ability to understand the interrelationships among image parts and the overall structure. information to enhance the model’s ability to understand the interrelationships between image parts and the overall structure, and to improve the reconstruction ability of the model; then a method based on Gaussian difference significance is proposed, which is combined with gradient magnitude similarity and colour difference to compare the differences between reconstructed and original images from multiple perspectives, and to improve the anomaly localisation performance of the model. We conducted extensive experiments on the industrial datasets MVTec AD and MTD to validate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Autonomous driving navigation relies on diverse approaches, each with advantages and limitations depending on various factors. For HD maps, modular systems excel, while end-to-end methods dominate mapless scenarios. However, few leverage the strengths of both. This paper innovates by proposing a hybrid architecture that seamlessly integrates modular perception and control modules with data-driven path planning. This innovative design leverages the strengths of both approaches, enabling a clear understanding and debugging of individual components while simultaneously harnessing the learning power of end-to-end approaches. Our proposed architecture achieved first and second place in the 2023 CARLA Autonomous Driving Challenge's SENSORS and MAP tracks, respectively. These results demonstrate the architecture's effectiveness in both map-based and mapless navigation. We achieved a driving score of 41.56 and the highest route completion of 86.03 in the MAP track of the CARLA Challenge leaderboard 1, and driving scores of 35.36 and 1.23 in the CARLA Challenge SENSOR track with route completions of 85.01 and 9.55 , for, respectively, leaderboard 1 and 2. The results of leaderboard 2 raised the hybrid architecture to the first position, winning the edition of the 2023 CARLA Autonomous Driving Competition. [ABSTRACT FROM AUTHOR]

Electricity demand in residential areas is generally met by the local low-voltage grid or, alternatively, the national grid, which produces electricity using thermal power stations based on conventional sources. These generators are holding back the revolution and the transition to a green planet, being unable to cope with climatic constraints. In the residential context, to ensure a smooth transition to an ecological green city, the idea of using alternative sources will offer the solution. These alternatives must be renewable and naturally available on the planet. This requires a generation that is very responsive to the constraints of the 21st century. However, these sources are intermittent and require a hybrid solution known as Hybrid Renewable Energy Systems (HRESs). To this end, we have designed a hybrid system based on PV-, wind-turbine- and grid-supported battery storage and an electric vehicle connected to a residential building. We proposed an energy management system based on nonlinear programming. This optimization was solved using sequential quadrature programming. The data were then processed using a long short-term memory (LSTM) model to predict, with the contribution and cooperation of each source, how to meet the energy needs of each home. The prediction was ensured with an accuracy of around 95%. These prediction results have been injected into K-nearest neighbors (KNN), random forest (RF) and gradient boost (GRU) repressors to predict the storage collaboration rates handled by the local battery and the electric vehicle. Results have shown an R2_score of 0.6953, 0.8381, and 0.739, respectively. This combination permitted an efficient prediction of the potential consumption from the grid with a value of an R²-score of around 0.9834 using LSTM. This methodology is effective in allowing us to know in advance the amount of energy of each source, storage, and excess grid injection and to propose the switching control of the hybrid architecture. [ABSTRACT FROM AUTHOR]

In order to improve the efficiency of the educational administration system in colleges and universities, a technology based on the hybrid architecture of C/S and B/S is proposed. Based on the development experience of previous information management systems, this method analyzes and compares the commonly used C/S architecture and B/S architecture. From the perspective of software engineering, the design idea of genetic algorithm course scheduling management system based on C/S and B/S cross-parallel mode is proposed. The experimental results show that the course goodness analysis method in the class schedule arrangement scheme using genetic algorithm shows that the average course goodness value is 3.2, which is not greater than the weighted total goodness value of 0.4. The feasibility of the design idea of the genetic algorithm class scheduling management system based on the C/S and B/S cross-parallel mode is effectively proved. It is concluded that it is proved that the technology research based on the hybrid architecture of C/S and B/S can effectively improve the effect of the educational administration system in colleges and universities.

In the realm of positive building design, local energy production systems are gaining prominence. Strategies aimed at advancing this concept often involve optimizing hybrid renewable systems through various means, including sizing, maximizing power extraction, or energy management. Among these, energy management holds particular significance, especially in systems incorporating dual energy storage sources, as addressed in our work. Here, we explore the optimization of hybrid renewable systems, focusing on photovoltaic, wind, pumped storage, and battery storage as energy sources in a proposed hybrid local energy generation system. Managed by a multi-source controller, driven by an optimal energy management system, our approach aims to better fulfill the thermal needs of buildings in semi-arid climates. The system’s modeling and calculation processes are conducted within the TRNSYS environment. Utilizing TRNSYS embedded computers, the controller is programmed according to a tailored energy management algorithm. Simulation results demonstrate the system’s capability to effectively meet energy demands by leveraging renewable energy sources. Notably, our system showcases a remarkable 39% improvement in energy self-sufficiency compared to conventional approaches. Future research endeavors will explore the integration of high-resistivity phase change materials to further enhance positive building design.

Abstract Background Identification of difficult laryngoscopy is a frequent demand in cervical spondylosis clinical surgery. This work aims to develop a hybrid architecture for identifying difficult laryngoscopy based on new indexes. Methods Initially, two new indexes for identifying difficult laryngoscopy are proposed, and their efficacy for predicting difficult laryngoscopy is compared to that of two conventional indexes. Second, a hybrid adaptive architecture with convolutional layers, spatial extraction, and a vision transformer is proposed for predicting difficult laryngoscopy. The proposed adaptive hybrid architecture is then optimized by determining the optimal location for extracting spatial information. Results The test accuracy of four indexes using simple model is 0.8320. The test accuracy of optimized hybrid architecture using four indexes is 0.8482. Conclusion The newly proposed two indexes, the angle between the lower margins of the second and sixth cervical spines and the vertical direction, are validated to be effective for recognizing difficult laryngoscopy. In addition, the optimized hybrid architecture employing four indexes demonstrates improved efficacy in detecting difficult laryngoscopy. Trial registration Ethics permission for this research was obtained from the Medical Scientific Research Ethics Committee of Peking University Third Hospital (IRB00006761-2015021) on 30 March 2015. A well-informed agreement has been received from all participants. Patients were enrolled in this research at the Chinese Clinical Trial Registry ( http://www.chictr.org.cn , identifier: ChiCTR-ROC-16008598) on 6 June 2016.

Abstract As the first point of contact for patients, General Practitioners (GPs) play a crucial role in the National Health Service (NHS). An accurate primary diagnosis from the GP can alleviate the burden on specialists and reduce the time needed to re-confirm the patient’s condition, allowing for more efficient further examinations. However, GPs have broad but less specialized knowledge, which limits the accuracy of their diagnosis. Therefore, it is imperative to introduce an intelligent system to assist GPs in making decisions. This paper introduces two data augmentation methods, the Complaint Symptoms Integration Method and Symptom Dot Separating Method, to integrate essential information into the Integration dataset. Additionally, it proposes a hybrid architecture that fuses the features of words from different representation spaces. Experiments demonstrate that, compared to commonly used pre-trained attention-based models, our hybrid architecture delivers the best classification performance for four common neurological diseases on the enhanced Integration dataset. For example, the classification accuracy of the BERT+CNN hybrid architecture is 0.897, which is a 5.1% improvement over both BERT and CNN with 0.846. Finally, this paper develops an AI diagnosis assistant web application that leverages the superior performance of this architecture to help GPs complete primary diagnosis efficiently and accurately.

The motion of an object or camera platform makes the acquired image blurred. This degradation is a major reason to obtain a poor-quality image from an imaging sensor. Therefore, developing an efficient deep-learning-based image processing method to remove the blur artifact is desirable. Deep learning has recently demonstrated significant efficacy in image deblurring, primarily through convolutional neural networks (CNNs) and Transformers. However, the limited receptive fields of CNNs restrict their ability to capture long-range structural dependencies. In contrast, Transformers excel at modeling these dependencies, but they are computationally expensive for high-resolution inputs and lack the appropriate inductive bias. To overcome these challenges, we propose an Efficient Hybrid Network (EHNet) that employs CNN encoders for local feature extraction and Transformer decoders with a dual-attention module to capture spatial and channel-wise dependencies. This synergy facilitates the acquisition of rich contextual information for high-quality image deblurring. Additionally, we introduce the Simple Feature-Embedding Module (SFEM) to replace the pointwise and depthwise convolutions to generate simplified embedding features in the self-attention mechanism. This innovation substantially reduces computational complexity and memory usage while maintaining overall performance. Finally, through comprehensive experiments, our compact model yields promising quantitative and qualitative results for image deblurring on various benchmark datasets.

Advances in deep learning and computer vision techniques have made impacts in the field of remote sensing, enabling efficient data analysis for applications such as land cover classification and change detection. Convolutional neural networks (CNNs) and transformer architectures have been utilized in visual perception algorithms due to their effectiveness in analyzing local features and global context. In this paper, we propose a hybrid transformer architecture that consists of a CNN-based encoder and transformer-based decoder. We propose a feature adjustment module that refines the multiscale feature maps extracted from an EfficientNet backbone network. The adjusted feature maps are integrated into the transformer-based decoder to perform the semantic segmentation of the remote sensing images. This paper refers to the proposed encoder–decoder architecture as a semantic feature adjustment network (SFA-Net). To demonstrate the effectiveness of the SFA-Net, experiments were thoroughly conducted with four public benchmark datasets, including the UAVid, ISPRS Potsdam, ISPRS Vaihingen, and LoveDA datasets. The proposed model achieved state-of-the-art accuracy on the UAVid, ISPRS Vaihingen, and LoveDA datasets for the segmentation of the remote sensing images. On the ISPRS Potsdam dataset, our method achieved comparable accuracy to the latest model while reducing the number of trainable parameters from 113.8 M to 10.7 M.

Abstract-Achieving high solution efficiency on conventional sequential computation architecture is a challenging task due to penetration of multiple renewable energy sources (RESs). This challenge has become the bottleneck for the application in real-time grid operation, grid planning and analysis of the large-scale and complicated modern power system. Therefore, this paper proposes a parallel multi-GPU and multi-process Fast Decoupled (FD) method to accelerate the power flow calculation, reducing the system responsive time and guaranteeing real-time performance on a large-scale modern power system. In this paper, two hierarchy architecture, task parallelism and data parallelism, are designed to optimize the FD solver parallelization. Moreover, the GPUDirect technology is employed to enhance efficiency of data transmission and drastically reduce copy overhead. The proposed method in this paper achieves a speedup of 9 × ∼ 33 × , compared to the single GPU on a sample large-scale power system. [ABSTRACT FROM AUTHOR]

Machine learning and deep neural networks have shown satisfactory performance in the supervised classification of Polarimetric Synthetic Aperture Radar (PolSAR) images. However, the PolSAR image classification task still faces some challenges. First, the current form of model input used for this task inevitably involves tedious preprocessing. In addition, issues such as insufficient labels and the design of the model also affect classification performance. To address these issues, this study proposes an augmentation method to better utilize the labeled data and improve the input format of the model, and an end-to-end PolSAR image global classification is implemented on our proposed hybrid network, PolSARMixer. Experimental results demonstrate that, compared to existing methods, our proposed method reduces the steps for the classification of PolSAR images, thus eliminating repetitive data preprocessing procedures and significantly improving classification performance. [ABSTRACT FROM AUTHOR]

As the first point of contact for patients, General Practitioners (GPs) play a crucial role in the National Health Service (NHS). An accurate primary diagnosis from the GP can alleviate the burden on specialists and reduce the time needed to re-confirm the patient's condition, allowing for more efficient further examinations. However, GPs have broad but less specialized knowledge, which limits the accuracy of their diagnosis. Therefore, it is imperative to introduce an intelligent system to assist GPs in making decisions. This paper introduces two data augmentation methods, the Complaint Symptoms Integration Method and Symptom Dot Separating Method, to integrate essential information into the Integration dataset. Additionally, it proposes a hybrid architecture that fuses the features of words from different representation spaces. Experiments demonstrate that, compared to commonly used pre-trained attention-based models, our hybrid architecture delivers the best classification performance for four common neurological diseases on the enhanced Integration dataset. For example, the classification accuracy of the BERT+CNN hybrid architecture is 0.897, which is a 5.1% improvement over both BERT and CNN with 0.846. Finally, this paper develops an AI diagnosis assistant web application that leverages the superior performance of this architecture to help GPs complete primary diagnosis efficiently and accurately. [ABSTRACT FROM AUTHOR]

Background: Identification of difficult laryngoscopy is a frequent demand in cervical spondylosis clinical surgery. This work aims to develop a hybrid architecture for identifying difficult laryngoscopy based on new indexes. Methods: Initially, two new indexes for identifying difficult laryngoscopy are proposed, and their efficacy for predicting difficult laryngoscopy is compared to that of two conventional indexes. Second, a hybrid adaptive architecture with convolutional layers, spatial extraction, and a vision transformer is proposed for predicting difficult laryngoscopy. The proposed adaptive hybrid architecture is then optimized by determining the optimal location for extracting spatial information. Results: The test accuracy of four indexes using simple model is 0.8320. The test accuracy of optimized hybrid architecture using four indexes is 0.8482. Conclusion: The newly proposed two indexes, the angle between the lower margins of the second and sixth cervical spines and the vertical direction, are validated to be effective for recognizing difficult laryngoscopy. In addition, the optimized hybrid architecture employing four indexes demonstrates improved efficacy in detecting difficult laryngoscopy. Trial registration: Ethics permission for this research was obtained from the Medical Scientific Research Ethics Committee of Peking University Third Hospital (IRB00006761-2015021) on 30 March 2015. A well-informed agreement has been received from all participants. Patients were enrolled in this research at the Chinese Clinical Trial Registry (http://www.chictr.org.cn, identifier: ChiCTR-ROC-16008598) on 6 June 2016. [ABSTRACT FROM AUTHOR]

The healthcare sector is witnessing rapid transformation with the rise of the Internet of Medical Things (IoMT). This presents unpar-alleled opportunities for continuous, personalized health surveillance. To truly tap into the IoMT's capabilities, it's essential to employ a flexible and robust data processing framework. In this article, we introduce a comprehensive four-tiered architecture tailored for the IoMT. This model, which we foresee as a benchmark for similar platforms, spans from interconnected devices to an edge computing layer, extends through a fog computing level, and culminates in the cloud. To bolster the system's resilience and features, two cross-sectional layers - one centered on security, and the other on artificial intelligence (AI) - are integrated across the four tiers. Additionally, we outline strategies for efficient load balancing, enhancing overall system performance. This initiative marks a pivotal advancement in IoMT architectural standardization, setting the stage for broader, more effective deployment in healthcare. [ABSTRACT FROM AUTHOR]

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

In order to solve the problem of insufficient analytical capabilities, application depth, and data innovation arising from the large amount of business data accumulated in the organization of heavy-haul railway transportation, and promote the application of big data technology in the field of heavy haul railway transportation, this paper proposes a design solution for a big data analytics platform in the field of heavy-haul railway transportation. Based on an analysis of the characteristics of heavy-haul railways, a framework for the platform is presented, which utilizes a hybrid architecture combining Hadoop and MPP (Massively Parallel Processing). The study also designs the application and technical architecture of the big data platform, and discusses the key technologies and business applications of the platform. Currently, the construction and operation of the heavy-haul railway big data analytics platform based on this design solution have been implemented [ABSTRACT FROM AUTHOR]

Plants are widely grown around the world and have high economic benefits. plant leaf diseases not only negatively affect the healthy growth and development of plants, but also have a negative impact on the environment. While traditional manual methods of identifying plant pests and diseases are costly, inefficient and inaccurate, computer vision technologies can avoid these drawbacks and also achieve shorter control times and associated cost reductions. The focusing mechanism of Transformer-based models(such as Visual Transformer) improves image interpretability and enhances the achievements of convolutional neural network (CNN) in image recognition, but Visual Transformer(ViT) performs poorly on small and medium-sized datasets. Therefore, in this paper, we propose a new hybrid architecture named FOTCA, which uses Transformer architecture based on adaptive Fourier Neural Operators(AFNO) to extract the global features in advance, and further down sampling by convolutional kernel to extract local features in a hybrid manner. To avoid the poor performance of Transformer-based architecture on small datasets, we adopt the idea of migration learning to make the model have good scientific generalization on OOD (Out-of-Distribution) samples to improve the model's overall understanding of images. In further experiments, Focal loss and hybrid architecture can greatly improve the convergence speed and recognition accuracy of the model in ablation experiments compared with traditional models. The model proposed in this paper has the best performance with an average recognition accuracy of 99.8% and an F1-score of 0.9931. It is sufficient for deployment in plant leaf disease image recognition. [ABSTRACT FROM AUTHOR]

Autonomous driving navigation relies on diverse approaches, each with advantages and limitations depending on various factors. For HD maps, modular systems excel, while end-to-end methods dominate mapless scenarios. However, few leverage the strengths of both. This paper innovates by proposing a hybrid architecture that seamlessly integrates modular perception and control modules with data-driven path planning. This innovative design leverages the strengths of both approaches, enabling a clear understanding and debugging of individual components while simultaneously harnessing the learning power of end-to-end approaches. Our proposed architecture achieved first and second place in the 2023 CARLA Autonomous Driving Challenge’s SENSORS and MAP tracks, respectively. These results demonstrate the architecture’s effectiveness in both map-based and mapless navigation. We achieved a driving score of 41.56 and the highest route completion of 86.03 in the MAP track of the CARLA Challenge leaderboard 1, and driving scores of 35.36 and 1.23 in the CARLA Challenge SENSOR track with route completions of 85.01 and 9.55, for, respectively, leaderboard 1 and 2. The results of leaderboard 2 raised the hybrid architecture to the first position, winning the edition of the 2023 CARLA Autonomous Driving Competition.

Blurring is one of the main degradation factors in image degradation, so image deblurring is of great interest as a fundamental problem in low-level computer vision. Because of the limited receptive field, traditional CNNs lack global fuzzy region modeling, and do not make full use of rich context information between features. Recently, a transformer-based neural network structure has performed well in natural language tasks, inspiring rapid development in the field of defuzzification. Therefore, in this paper, a hybrid architecture based on CNN and transformers is used for image deblurring. Specifically, we first extract the shallow features of the blurred images using a cross-layer feature fusion block that emphasizes the contextual information of each feature extraction layer. Secondly, an efficient transformer module for extracting deep features is designed, which fully aggregates feature information at medium and long distances using vertical and horizontal intra- and inter-strip attention layers, and a dual gating mechanism is used as a feedforward neural network, which effectively reduces redundant features. Finally, the cross-layer feature fusion block is used to complement the feature information to obtain the deblurred image. Extensive experimental results on publicly available benchmark datasets GoPro, HIDE, and the real dataset RealBlur show that the proposed method outperforms the current mainstream deblurring algorithms and recovers the edge contours and texture details of the images more clearly. [ABSTRACT FROM AUTHOR]

Rice is a staple food for over half of the global population, but it faces significant yield losses: up to 52% due to leaf blast disease and brown spot diseases, respectively. This study aimed at proposing a hybrid architecture, namely ResViT-Rice, by taking advantage of both CNN and transformer for accurate detection of leaf blast and brown spot diseases. We employed ResNet as the backbone network to establish a detection model and introduced the encoder component from the transformer architecture. The convolutional block attention module was also integrated to ResViT-Rice to further enhance the feature-extraction ability. We processed 1648 training and 104 testing images for two diseases and the healthy class. To verify the effectiveness of the proposed ResViT-Rice, we conducted comparative evaluation with popular deep learning models. The experimental result suggested that ResViT-Rice achieved promising results in the rice disease-detection task, with the highest accuracy reaching 0.9904. The corresponding precision, recall, and F1-score were all over 0.96, with an AUC of up to 0.9987, and the corresponding loss rate was 0.0042. In conclusion, the proposed ResViT-Rice can better extract features of different rice diseases, thereby providing a more accurate and robust classification output. [ABSTRACT FROM AUTHOR]

Synchrotron radiation sources are widely used in interdisciplinary research, generating an enormous amount of data while posing serious challenges to the storage, processing, and analysis capabilities of the large-scale scientific facilities worldwide. A flexible and scalable computing architecture, suitable for complex application scenarios, combined with efficient and intelligent scheduling strategies, plays a key role in addressing these issues. In this work, we present a novel cloud–edge hybrid intelligent system (CEHIS), which was architected, developed, and deployed by the Big Data Science Center (BDSC) at the Shanghai Synchrotron Radiation Facility (SSRF) and meets the computational needs of the large-scale scientific facilities. Our methodical simulations demonstrate that the CEHIS is more efficient and performs better than the cloud-based model. Here, we have applied a deep reinforcement learning approach to the task scheduling system, finding that it effectively reduces the total time required for the task completion. Our findings prove that the cloud–edge hybrid intelligent architectures are a viable solution to address the requirements and conditions of the modern synchrotron radiation facilities, further enhancing their data processing and analysis capabilities. [ABSTRACT FROM AUTHOR]

The article discusses HPC system computing resources distribution management during execution of interactive and batch jobs. A set of queues for interactive and batch jobs is proposed, and an algorithm for the dynamic resources allocation between the proposed job queues is described. [ABSTRACT FROM AUTHOR]

Diabetic retinopathy (DR) is the most severe ocular complication of diabetes. It leads to serious eye complications such as vision impairment and blindness. A computer-aided diagnosis may help in the early detection of this disease, which increases the chances of treating it efficiently. This paper carried out an empirical evaluation of the performances of 28 deep hybrid architectures for an automatic binary classification of the referable diabetic retinopathy, and compared them to seven end-to-end deep learning (DL) architectures. For the hybrid architectures, we combined seven DL techniques for feature extraction (DenseNet201, VGG16, VGG19, MobileNet_V2, Inception_V3, Inception_ResNet_V2 and ResNet50) and four classifiers (SVM, MLP, DT and KNN). For the end-to-end DL architectures, we used the same techniques used for the feature extraction in the hybrid architectures. The architectures were compared in terms of accuracy, sensitivity, precision and F1-score using the Scott Knott test and the Borda count voting method. All the empirical evaluations were over three datasets: APTOS, Kaggle DR and Messidor-2, using a k-fold cross validation method. The results showed the potential of combining deep learning techniques for feature extraction and classical machine learning techniques to classify referable diabetic retinopathy. The hybrid architecture using the SVM classifier and MobileNet_V2 for feature extraction was the top performing architecture and it was classified with the best performing end-to-end architectures in the best clusters of APTOS, Kaggle DR and Messidor-2 datasets with an accuracy equal to 88.80%, 84.01% and 84.05% respectively. Note that the two end-to-end architectures DenseNet201 and MobileNet_V2 outperformed all the hybrid architectures over the three datasets. However, we recommend the use of the hybrid architecture designed with SVM and MobileNet_V2 since it is promising and less time consuming, and requires less parameter tuning compared to the end-to-end techniques. [ABSTRACT FROM AUTHOR]

Machine learning and deep neural networks have shown satisfactory performance in the supervised classification of Polarimetric Synthetic Aperture Radar (PolSAR) images. However, the PolSAR image classification task still faces some challenges. First, the current form of model input used for this task inevitably involves tedious preprocessing. In addition, issues such as insufficient labels and the design of the model also affect classification performance. To address these issues, this study proposes an augmentation method to better utilize the labeled data and improve the input format of the model, and an end-to-end PolSAR image global classification is implemented on our proposed hybrid network, PolSARMixer. Experimental results demonstrate that, compared to existing methods, our proposed method reduces the steps for the classification of PolSAR images, thus eliminating repetitive data preprocessing procedures and significantly improving classification performance.

The increasing frequency of skin tumour across the globe and their timely diagnosis is one of the most promising research directions in the healthcare domain. The most important cause behind the skin cancer mortalities is delayed detection. Early detection followed by adequate treatment may enhance the chances of human survival to a great extent. However, extracting the features from the tumour images for the possible detection of skin cancer is not a trivial task. Numerous deep learning models are extensively employed for the efficient features' extraction for skin cancer detection but literature demonstrate further scope of improvements in various performance measures. In this paper, we propose a hybrid deep Convolutional Neural Network architecture inspired from pretrained architectures for the skin cancer detection by incorporating three major heuristics viz. usage of multiple smaller sized convolutional filters instead of using a single larger filter homogeneously and consistently across the entire model, utilization of skip or residual connections to mitigate the vanishing gradient problem in the deeper model, and learning rate annealing by introducing cyclic learning rate. Experimental results performed on HAM10000 dataset observed an improvement in various performance measures and faster model convergence to a significant extent in comparison with the state-of-the-arts. [ABSTRACT FROM AUTHOR]

• This study contributes a novel approach to autoimmune disorder detection by integrating relational graph representations with sequential patient health records. • Leveraging Relational Neural Networks (RNNs) and Long Short-Term Memory (LSTM) networks, the method captures intricate interactions within heterogeneous data. • By learning node embeddings from graph structures and fusing them with patient health time-series, the approach enhances diagnostic capabilities. • This innovative framework promises advancements in clinical decision support and personalized healthcare interventions, improving diagnostic accuracy and patient outcomes. Autoimmune diseases, a diverse group of disorders, involve the immune system mistakenly attacking the body's own tissues, leading to inflammation and potential damage across various organs and systems. Conditions like Rheumatoid Arthritis and Type 1 Diabetes collectively impact millions globally, presenting diverse symptoms and complexities. Treatment primarily focuses on alleviating symptoms, managing inflammation, and mitigating complications through medication, while lifestyle adjustments play a supplementary role in enhancing overall well-being. Rheumatoid Arthritis (RA) exemplifies the systemic nature of autoimmune disorders, characterized by chronic inflammation primarily affecting the joints. The immune systems misguided attack on the synovium the membranes surrounding joints results in inflammation, pain, stiffness, and swelling, potentially causing damage to cartilage and bone. RA extends its impact beyond the joints, affecting organs such as the skin, eyes, heart, lungs, and blood vessels, manifesting systemic symptoms like fatigue, fever, and weight loss. While the exact cause of RA remains elusive, a combination of genetic, environmental, and hormonal factors is implicated. Genetic predispositions, particularly genes related to the immune system, may contribute to susceptibility. Environmental triggers such as smoking, infections, and exposure to pollutants are also believed to play a role in disease onset. In the realm of autoimmune disorder management, accurate and timely detection is paramount for effective treatment and intervention. To address this challenge, study proposes a novel approach leveraging Relational Neural Networks (RNNs) and Long Short-Term Memory (LSTM) networks. By integrating the relational structure of autoimmune disorder data, captured through graph representations, with sequential patient health records, method offers a comprehensive framework for detection. Construct a heterogeneous graph where nodes represent patients, medical conditions, and clinical attributes, with edges encoding relationships like symptom co-occurrence or shared diagnoses. Utilizing Relational Neural Networks, learn node embedding that encapsulates intricate interactions and dependencies within the graph. These embedding are then fused with sequential patient data time-series sequences of clinical events using LSTM networks. The hybrid architecture facilitates effective capture of both the relational context of autoimmune disorders and the temporal evolution of patient health states. Evaluation on a real-world dataset comprising electronic health records of autoimmune disorder patients demonstrates the efficacy of approach, outperforming baseline methods in accurately identifying autoimmune disorders. The accuracy of the proposed RNN-LSTM system is 99.54%, Sensitivity of 92.34% and Precision value of 94.5 %. The potential of integrating relational and sequential information to enhance autoimmune disorder detection, promising advancements in clinical decision support systems and personalized healthcare interventions. [ABSTRACT FROM AUTHOR]

Vision transformers have been successfully applied to image recognition tasks due to their ability to capture long-range dependencies within an image. However, they still suffer from weak local feature extraction, easy loss of channel interaction information in one-dimensional multi-head self-attention modeling, and large number of parameters. This paper proposes a lightweight image classification hybrid architecture named EPSViTs (Efficient Parameter Shared Transformer, EPSViTs). Firstly, a new local feature extraction module is designed to effectively enhance the expression of local features. Secondly, using the parameter sharing approach, a lightweight multi-head self-attention module based on information interaction is designed, which can globally model the image from both spatial and channel dimensions, and mine the potential correlation of the image in space and channel. Extensive experiments are conducted on three public datasets, a subset of ImageNet, Cifar100 and APTOS2019, a private dataset Mushroom66, and the results show that the hybrid architecture EPSViTs proposed in this paper based on parameter sharing for multi-head self-attentive image classification has obvious advantages, especially on a subset of ImageNet to reach 89.18%, which is a 3.8% improvement compared to Edgevits_xxs, verifying the effectiveness of the model. • This paper designs a fine-grained local feature extraction module LFE. • This paper designs a lightweight parameter sharing attention mechanism EPSA. • A new lightweight hybrid architecture EPSViTs is built based on the LFE and EPSA. • The reliability and generalization of our model were validated on four datasets. [ABSTRACT FROM AUTHOR]

Convolution is widely adapted in deep learning models on image classification tasks for extracting hidden spatial-domain representations. However, as convolution is channel-specific, the potential cross-channel correlations in images are often neglected. This paper proposes a novel model, namely I-CNet, which leverages involution and convolution for improving the accuracy of image classification tasks, by extracting feature representations on both channel domain and spatial domain. The proposed I-CNet has been experimented on three image classification datasets. The experimental results show that involution component employed in I-CNet model can effectively represent cross-channel features in images and I-CNet is superior to other comparatives with higher classification accuracy achieved on all the three datasets.

The bottom-up assembly utilizing nanocrystals (NCs) as building blocks is a highly anticipated approach to access novel metamaterials. Limited by the library of binary spherical assembly, colloidal components of building blocks are necessarily extended for targeting the desired assembled-superstructure with emergent properties. Here, a generalized approach is presented that carbon nanotubes (CNTs) capped with long-chain hydrocarbon ligands, as creative colloidal components, can be compatibly co-assembled with various NCs into binary supraspheres based on a confinement microemulsion environment. Such cross-dimensional assemblies possess the structural merits of excellent conductivity, abundant voids and high stability, thus synergistically boosting its electrochemical performance. As a proof of concept, the co-assembled CoFe 2 O 4 /CNTs supraspheres exhibit extraordinary capacity (890 mAh g−1 at 0.5 A g−1), rate capability (318 mAh g−1 at 20 A g−1), and structural stability (458 mAh g−1 at 5 A g−1 after 1000 cycles) relative to pure CoFe 2 O 4 when evaluated in Li-ion battery. Beyond that, its derived phosphides CoFeP/CNTs assemblies outperform most of the state-of-art electrodes when acting as electrocatalysts for hydrogen evolution reactions (HER). This work provides a new perspective for constructing superstructures of NCs for potential applications in energy storage devices. [Display omitted] [ABSTRACT FROM AUTHOR]

Evaluating the severity of ulcerative colitis (UC) through the Mayo endoscopic subscore (MES) is crucial for understanding patient conditions and providing effective treatment. However, UC lesions present different characteristics in endoscopic images, exacerbating interclass similarities and intraclass differences in MES classification. In addition, inexperience and review fatigue in endoscopists introduces nontrivial challenges to the reliability and repeatability of MES evaluations. In this paper, we propose a pyramid hybrid feature fusion framework ( PHF 3 ) as an auxiliary diagnostic tool for clinical UC severity classification. Specifically, the PHF 3 model has a dual-branch hybrid architecture with ResNet50 and a pyramid vision Transformer (PvT), where the local features extracted by ResNet50 represent the relationship between the intestinal wall at the near-shot point and its depth, and the global representations modeled by the PvT capture similar information in the cross-section of the intestinal cavity. Furthermore, a feature fusion module (FFM) is designed to combine local features with global representations, while second-order pooling (SOP) is applied to enhance discriminative information in the classification process. The experimental results show that, compared with existing methods, the proposed PHF 3 model has competitive performance. The area under the receiver operating characteristic curve (AUC) of MES 0, MES 1, MES 2, and MES 3 reached 0.996, 0.972, 0.967, and 0.990, respectively, and the overall accuracy reached 88.91%. Thus, our proposed method is valuable for developing an auxiliary assessment system for UC severity. [ABSTRACT FROM AUTHOR]

Blurring is one of the main degradation factors in image degradation, so image deblurring is of great interest as a fundamental problem in low-level computer vision. Because of the limited receptive field, traditional CNNs lack global fuzzy region modeling, and do not make full use of rich context information between features. Recently, a transformer-based neural network structure has performed well in natural language tasks, inspiring rapid development in the field of defuzzification. Therefore, in this paper, a hybrid architecture based on CNN and transformers is used for image deblurring. Specifically, we first extract the shallow features of the blurred images using a cross-layer feature fusion block that emphasizes the contextual information of each feature extraction layer. Secondly, an efficient transformer module for extracting deep features is designed, which fully aggregates feature information at medium and long distances using vertical and horizontal intra- and inter-strip attention layers, and a dual gating mechanism is used as a feedforward neural network, which effectively reduces redundant features. Finally, the cross-layer feature fusion block is used to complement the feature information to obtain the deblurred image. Extensive experimental results on publicly available benchmark datasets GoPro, HIDE, and the real dataset RealBlur show that the proposed method outperforms the current mainstream deblurring algorithms and recovers the edge contours and texture details of the images more clearly.

The article discusses HPC system computing resources distribution management during execution of interactive and batch jobs. A set of queues for interactive and batch jobs is proposed, and an algorithm for the dynamic resources allocation between the proposed job queues is described.

Rice is a staple food for over half of the global population, but it faces significant yield losses: up to 52% due to leaf blast disease and brown spot diseases, respectively. This study aimed at proposing a hybrid architecture, namely ResViT-Rice, by taking advantage of both CNN and transformer for accurate detection of leaf blast and brown spot diseases. We employed ResNet as the backbone network to establish a detection model and introduced the encoder component from the transformer architecture. The convolutional block attention module was also integrated to ResViT-Rice to further enhance the feature-extraction ability. We processed 1648 training and 104 testing images for two diseases and the healthy class. To verify the effectiveness of the proposed ResViT-Rice, we conducted comparative evaluation with popular deep learning models. The experimental result suggested that ResViT-Rice achieved promising results in the rice disease-detection task, with the highest accuracy reaching 0.9904. The corresponding precision, recall, and F1-score were all over 0.96, with an AUC of up to 0.9987, and the corresponding loss rate was 0.0042. In conclusion, the proposed ResViT-Rice can better extract features of different rice diseases, thereby providing a more accurate and robust classification output.

The reputation-based scheme is a promising solution to prevent malicious behaviors in Vehicular Ad-hoc Networks (VANETs). However, traditional centralized reputation schemes are not suited for distributed networks, while decentralized reputation schemes are vulnerable to malicious vehicles spreading false messages. Most of these schemes assume that the behavior of vehicles can be accurately measured as reputation from the communication, ignoring that malicious vehicles may behave intelligently to avoid being detected. In this paper, we propose a hybrid reputation system (HDRS) which allows vehicles and roadside units (RSU) to complete reputation evaluations separately and provide references to each other. HDRS utilizes a reliability evaluation module to filter out unreliable calculation results and reference records. Furthermore, HDRS includes a dynamic adjustment mechanism for the reputation update interval, employing Analytic Hierarchy Process (AHP) and reliability evaluation results to resist intelligent attacks. Simulation results illustrate that HDRS can maintain a high detection rate and low false-positive rate for detecting malicious vehicles in different environments. Compared with existing schemes, HDRS increases the detection rates of collusion and intelligent attacks by 30% and 16%, respectively. [ABSTRACT FROM AUTHOR]

A cognitive agent performing in the real world needs to learn relevant concepts about its environment (e.g., objects, color, and shapes) and react accordingly. In addition to learning the concepts, it needs to learn relations between the concepts, in particular spatial relations between objects. In this paper, we propose three approaches that allow a cognitive agent to learn spatial relations. First, using an embodied model, the agent learns to reach toward an object based on simple instructions involving left-right relations. Since the level of realism and its complexity does not permit large-scale and diverse experiences in this approach, we devise as a second approach a simple visual dataset for geometric feature learning and show that recent reasoning models can learn directional relations in different frames of reference. Yet, embodied and simple simulation approaches together still do not provide sufficient experiences. To close this gap, we thirdly propose utilizing knowledge bases for disembodied spatial relation reasoning. Since the three approaches (i.e., embodied learning, learning from simple visual data, and use of knowledge bases) are complementary, we conceptualise a cognitive architecture that combines these approaches in the context of spatial relation learning [ABSTRACT FROM AUTHOR]

Context provides additional information to determine the actual emotional state of a person as part of a scene. Existing works on emotion recognition in context focused only on the features extracted from the entire image and the target's body. In this work, we propose a comprehensive multi-cue based emotion recognition framework that incorporates the context, using a hybrid architecture comprised of four separate deep convolutional neural networks and a novel feature fusion mechanism. Each deep network presented in our proposed approach effectively learns the emotion-related features from the facial, body pose, non-target subject and entire image, individually. Experiments on Emotic, an in-painted and a newly constructed EmoSec datasets show that our proposed emotion recognition framework is promising when compared to existing methods in terms of accurately classifying emotions. Comparison with the state-of-the-art deep networks and off-the-shelf fusion techniques demonstrates that our network showed an improved performance. Furthermore, the performance evaluation of our proposed approach on all three datasets confirms that the contextual information influences the emotional state of a human. [ABSTRACT FROM AUTHOR]

• Impact resistance of architected composites is amplified by positive hybrid design. • Lamellar bottom enhances toughness; Brick-mud top guarantees flexural stiffness. • Declining volume ratio of cohesive and adhesive phases promotes energy dissipation. • Increasing gradient in material property contributes rigidity-toughness coupling. In recent decades, architected composite materials have received increasing attention due to its promising potential in improving the impact resistance performance compared with bulk materials. Some ingenious microstructures of biomaterials can enlighten the design of novel architected composite materials, such as nacre, conch shell and dactyl club of mantis shrimp, which have been extensively applied. Here, in order to further enhance the impact resistance of architected composite materials, hybrid architectures are proposed based on nacre-like brick-mud and sponge-like lamellar architectures. Then, their impact resistance performances and mechanism are investigated by combining experiment with finite element simulation. The results show that through positive hybrid design inspired of the dactyl club, the lamellar bottom makes the hybrid architecture retain excellent toughness, and the brick-mud top guarantees large flexural stiffness, thereby achieving rigidity-toughness coupling and boosting the energy absorption dramatically. Moreover, as the decline of volume ratio of soft phases within and between brick-mud layers, the energy absorption of brick-mud architecture gradually increases owing to the rising damage resistance. Further, the rigidity-toughness coupled effect can be optimized by positively hybridizing multiple architectures with larger variation in flexural stiffness and toughness, even achieving 375 % improvement in energy absorption compared with the basic brick-mud architecture. [Display omitted] [ABSTRACT FROM AUTHOR]

Synchrotron radiation sources are widely used in interdisciplinary research, generating an enormous amount of data while posing serious challenges to the storage, processing, and analysis capabilities of the large-scale scientific facilities worldwide. A flexible and scalable computing architecture, suitable for complex application scenarios, combined with efficient and intelligent scheduling strategies, plays a key role in addressing these issues. In this work, we present a novel cloud–edge hybrid intelligent system (CEHIS), which was architected, developed, and deployed by the Big Data Science Center (BDSC) at the Shanghai Synchrotron Radiation Facility (SSRF) and meets the computational needs of the large-scale scientific facilities. Our methodical simulations demonstrate that the CEHIS is more efficient and performs better than the cloud-based model. Here, we have applied a deep reinforcement learning approach to the task scheduling system, finding that it effectively reduces the total time required for the task completion. Our findings prove that the cloud–edge hybrid intelligent architectures are a viable solution to address the requirements and conditions of the modern synchrotron radiation facilities, further enhancing their data processing and analysis capabilities.

The design and evaluation of a new hybrid architecture for digital pulse-width modulation (DPWM) are presented in this brief. This scheme uses a counter-based module to obtain the coarse adjustment of the duty cycle and a pulse former with a delay-line approach for fine-tuning. The latter employs fewer delay stages than other schemes by using the propagation delay of the multi-bit tunable delay elements as part of the desired pulse width variation. Our delay element provides very good linearity with the capability to compensate for deviations in the digital-to-time conversion. The prototype fabricated in a standard 130-nm CMOS process has 18 bits resolution (an 8-bit counter combined with a 10-bit delay-line-based pulse former), and low-power consumption when it operates with switching frequencies from 100 kHz to 1 MHz and a clock frequency of 17 MHz. Comparison with other recent alternatives in the literature is also provided. [ABSTRACT FROM AUTHOR]

The forthcoming communication networks for public safety authorities rely on the fifth generation (5G) of mobile networking technologies. Police officers, paramedics, border guards, as well as fire and rescue personnel, will connect through commercial operator’s access network and rapidly deployable tactical bubbles. This transition from closed and dedicated infrastructure to hybrid architecture will expand the threat surface and expose mission-critical applications and sensitive information to cyber and physical adversaries. We explore and survey security architecture and enablers for prioritized public safety communication in 5G networks. We identify security threat scenarios and analyze enabling vulnerabilities, threat actors, attacks vectors, as well as risk levels. Security enablers are surveyed for tactical access and core networks, commercial infrastructure, and mission-critical applications, starting from push-to-talk and group video communication and leading to situational-awareness and remote-controlled systems. Two solutions are trialed and described in more detail: remote attestation enhanced access control for constrained devices, and securing of satellite backhauls. We also discuss future research directions highlighting the need for enablers to automate security of rapid deployments, for military-grade cost-effective customizations of commercial network services to ensure robustness, and for hardening of various types of public safety equipment.

Cell-free massive MIMO systems consist of many distributed access points with simple components that jointly serve the users. In millimeter wave bands, only a limited set of predetermined beams can be supported. In a network that consolidates these technologies, downlink analog beam selection stands as a challenging task for the network sum-rate maximization. Low-cost digital filters can improve the network sum-rate further. In this work, we propose low-cost joint designs of analog beam selection and digital filters. The proposed joint designs achieve significantly higher sum-rates than the disjoint design benchmark. Supervised machine learning (ML) algorithms can efficiently approximate the input-output mapping functions of the beam selection decisions of the joint designs with low computational complexities. Since the training of ML algorithms is performed off-line, we propose a well-constructed joint design that combines multiple initializations, iterations, and selection features, as well as beam conflict control, i.e., the same beam cannot be used for multiple users. The numerical results indicate that ML algorithms can retain 99–100% of the original sum-rate results achieved by the proposed well-constructed designs.

Facial expression is an important aspect to recognize emotions between humans. However, this task remains difficult for machines. Several approaches have been developed aiming at strengthening the machine and endowing it, with the ability to decipher and read people's emotions from their faces in order to interact more intelligently. In this context, many deep learning (DL) approaches have been applied due to their outstanding recognition Accuracy. Aiming to gain better performance for facial expression recognition (FER) systems, we propose a hybrid DL architecture based on convolutional neural network and Stacked AutoEncoder. The main idea of this work is to combine the two feature vectors generated by each of these architectures and feed the resulting vector to a meaningful neural network. The architecture of the latter leads to learn each feature by dedicating a set of neurons for each component of the vector before combining them all together in the last layer. The publicly available four facial expression Datasets: Japanse Female Facial Expression (JAFFE), Extended Chon-Kanade (CK+), Facial Expression Recognition 2013 (FER2013) and AffectNet, were used during this research for both training and testing. The experimental results of our proposed architecture are comparable to or better than the relevant state-of-the-art methods in term of Accuracy, Recall, Precision and F-measure. We noted that our proposed approach obtains the best accuracy of 98.65% on the CK+, 95.78% on the JAFFE, 63.14% on the AffectNet and 80.02% on the FER2013 Datasets. [ABSTRACT FROM AUTHOR]

Synchronization is a fundamental procedure in cellular systems whereby user equipment (UE) acquires the time and frequency information required to decode the data transmitted by a base station (BS). Due to the small antenna aperture at millimeter wave (mmWave), large antenna arrays are used to compensate for the low signal-to-noise ratio (SNR) at the receiver. For this reason, synchronization is often performed jointly with beam training as in 5G New Radio (NR) to boost the receive SNR. However, if the mmWave multiple-input multiple-output (MIMO) channel is to be estimated, then synchronization must be performed in the low SNR regime owing to the lack of directional beamforming during compressive channel estimation. Current mmWave channel estimation algorithms implicitly take synchronization for granted, which is a not practical assumption in the low SNR regime. To solve this problem, this work proposes the first synchronization framework for hybrid mmWave MIMO that is robust to carrier frequency offset (CFO) and phase noise (PHN) synchronization errors. We propose two novel multi-stage algorithms to jointly estimate the various unknown parameters, based on the expectation-maximization (EM) algorithm. By using the QuaDRiGa 5G NR channel simulator, we show that compressive channel estimation without prior synchronization is possible, and the proposed approaches outperform current solutions for joint beam training and synchronization currently considered in 5G NR. [ABSTRACT FROM AUTHOR]