Advanced Network Architectures Research Articles

A Survey an Optimized Dense CNN Model for Recognizing Deepfake Images

Deepfake detection relies on a deep learning model. Deepfake content is generated using artificial intelligence and machine learning techniques to swap one person’s face with another’s in images. These altered images are certain to have a significant impact on society. Deepfakes utilize advanced technologies such as machine learning (ML) and deep learning (DL) to develop automated techniques for generating deceptive content. The model is trained using the DFDC (Deepfake Detection Challenge) dataset, which includes 1, 00,000 videos comprising both real and fake content. This survey examines recent progress in deep learning techniques for detecting deepfake images, highlighting their growing significance in today’s digital landscape. The paper outlines the issues arising from deepfakes, explores the techniques for generating deceptive content, and analyzes various deep learning models employed in deepfake detection, such as convolutional neural networks (CNN’s), MTCNN (Multi-task CNN), and Facial Landmark. It emphasizes the ongoing battle between the progression of deepfakes and the evolution of detection methods, underscoring the need for advanced and flexible neural network architectures to effectively curb the dissemination of deceptive information.

TinyML for smart agriculture: Comparative analysis of TinyML platforms and practical deployment for maize leaf disease identification

Maize, a vital global food crop, is susceptible to various diseases impacting agricultural yield and quality, necessitating effective monitoring strategies. This study presents a comprehensive approach utilizing advanced Convolutional Neural Network (CNN) architecture for efficient maize leaf disease detection and classification. Leveraging datasets from Kaggle, Harvard Dataverse, and Mendeley, the proposed CNN maize disease detection (MDD) system integrates optimization techniques, data augmentation, batch size adjustments, and fine-tuning strategies. Evaluation through data splitting reveals significant correlations among maize leaf classifications and the influence of data augmentation on image quality. Following training, the CNN-MDD model is deployed on a Microcontroller Unit (MCU) simulation, achieving promising optimization parameters: latency of 7.60 ms, RAM size of 726.60 KB, classifier flash size of 344.70 KB, and an accuracy of 94.60 %. These optimized parameters align well with MCU specifications, highlighting the potential for practical deployment on low-cost edge devices like Arduino Ble 33 Sense. Additionally, this research contributes to the creation of a Ghanaian maize crop pest and disease dataset and the development of an optimal and cost-effective CNN-MDD MCU edge device, aiming to provide local farmers in Ghana and Sub-Saharan Africa with a swift, offline solution for maize disease detection, thus fostering sustainable agricultural practices. We conduct a comparative analysis of Edge Impulse and TensorFlow, two prominent machine-learning platforms for TinyML model development and deployment in various applications, including smart crop disease detection. Utilizing a dataset comprising 12,344 raw images and 7,454 augmented images of healthy and diseased maize leaves, we evaluate the workflow, features, performance, and usability of each platform. Our findings indicate that Edge Impulse holds a slight advantage over TensorFlow in terms of ease of use, reliability, and memory footprint, whereas TensorFlow offers greater flexibility, customization options, and higher accuracy. We conclude that while both platforms are suitable for crop disease detection, a combined approach allows for leveraging TensorFlow's robust training capabilities for complex neural networks alongside the simplicity of Edge Impulse for deploying models on low-powered edge devices. This synergistic approach enhances the effectiveness and efficiency of crop disease detection systems, ultimately benefiting agricultural practices and food security.

Deep Learning for Exploring Landslides with Remote Sensing and Geo-Environmental Data: Frameworks, Progress, Challenges, and Opportunities

This article offers a comprehensive AI-centric review of deep learning in exploring landslides with remote-sensing techniques, breaking new ground beyond traditional methodologies. We categorize deep learning tasks into five key frameworks—classification, detection, segmentation, sequence, and the hybrid framework—and analyze their specific applications in landslide-related tasks. Following the presented frameworks, we review state-or-art studies and provide clear insights into the powerful capability of deep learning models for landslide detection, mapping, susceptibility mapping, and displacement prediction. We then discuss current challenges and future research directions, emphasizing areas like model generalizability and advanced network architectures. Aimed at serving both newcomers and experts on remote sensing and engineering geology, this review highlights the potential of deep learning in advancing landslide risk management and preservation.

Development and characterization of novel ultra-stable high internal phase Pickering emulsions gel: Interface structure, stabilization mechanism, and applications

While both protein-polysaccharide complexes and continuous phase gelation showed great potential for stabilizing high internal phase Pickering emulsions (HIPPEs), it is not known which strategies are more effective and adaptive. In this work, HIPPEs gel was prepared by Ca2+-induced continuous phase gelation of ovalbumin (OVA)-low methoxyl pectin (LMP) electrostatic complexes particles (OLECP)-stabilized HIPPEs. The microstructure, rheological properties, stability, and application of HIPPEs and HIPPEs gel were systematically investigated and compared at different LMP concentrations. The results showed that, compared to HIPPEs, continuous phase gelation effectively improved the viscoelasticity, water binding capacity, and hardness of HIPPEs gel due to its robust and developed interface architecture. CLSM and cryo-SEM observations revealed that OLECP interconnected with each other in the continuous phase to form an advanced network architecture in HIPPEs. Whereas, a firm gel network structure and a thick interfacial layer were found in HIPPEs gel. More importantly, with increasing pectin concentration, the cross-linking phenomenon of OLECP mainly at the interface, transformed into the gelation behavior of excess pectin molecules dominating in the continuous phase. LMP contributed to the formation of a finer 3D gel network structure and a firmer interfacial layer, providing a smaller oil droplet size and superior stability for HIPPEs gel. The photochemical stability of curcumin was increased from 16.1 ± 1.5 % to 93.7 ± 2.2 % after encapsulation in HIPPEs gel. In vitro digestion results testified that HIPPEs gel (10:15, w/w) can promote fat hydrolysis (from 32.35 ± 3.06 % to 45.52 ± 2.75 %) and enhance the bioaccessibility of curcumin (from 20.1 ± 1.5 % to 32.8 ± 2.6 %) compared to bulk oil. Besides, compared with HIPPEs, HIPPEs gel showed poor printability because of improper printing adaptability and liquidity. In a nutshell, our work provides valuable information for developing novel HIPPEs gels for various food applications.

General Hyperspectral Image Super-Resolution via Meta-Transfer Learning.

Recent advances in deep learning-based methods have led to significant progress in the hyperspectral super-resolution (SR). However, the scarcity and the high dimension of data have hindered further development since deep models require sufficient data to learn stable patterns. Moreover, the huge domain differences between hyperspectral image (HSI) datasets pose a significant challenge in generalizability. To address these problems, we present a general hyperspectral SR framework via meta-transfer learning (MTL). We randomly sample various spectral ranges for SR tasks during MTL, allowing the model to accumulate diverse task experiences. Additionally, we implement a task schedule to gradually expand the number of bands, bridging the significant domain differences between datasets. By leveraging multiple datasets, we are able to achieve better performance and greater generalizability, making it applicable under various circumstances. Meanwhile, as a general framework, our scheme can be applied to existing methods to obtain performance improvements. In addition, we design an advanced network architecture based on the multifusion features to further improve the performance. Experiments demonstrate that our method not only achieves superior performance in both qualitative and quantitative terms but also can adapt robustly to a new and difficult sample, where few epochs can yield quite considerable results.

Improving a Deep Learning Model to Accurately Diagnose LVNC.

Accurate diagnosis of Left Ventricular Noncompaction Cardiomyopathy (LVNC) is critical for proper patient treatment but remains challenging. This work improves LVNC detection by improving left ventricle segmentation in cardiac MR images. Trabeculated left ventricle indicates LVNC, but automatic segmentation is difficult. We present techniques to improve segmentation and evaluate their impact on LVNC diagnosis. Three main methods are introduced: (1) using full 800 × 800 MR images rather than 512 × 512; (2) a clustering algorithm to eliminate neural network hallucinations; (3) advanced network architectures including Attention U-Net, MSA-UNet, and U-Net++.Experiments utilize cardiac MR datasets from three different hospitals. U-Net++ achieves the best segmentation performance using 800 × 800 images, and it improves the mean segmentation Dice score by 0.02 over the baseline U-Net, the clustering algorithm improves the mean Dice score by 0.06 on the images it affected, and the U-Net++ provides an additional 0.02 mean Dice score over the baseline U-Net. For LVNC diagnosis, U-Net++ achieves 0.896 accuracy, 0.907 precision, and 0.912 F1-score outperforming the baseline U-Net. Proposed techniques enhance LVNC detection, but differences between hospitals reveal problems in improving generalization. This work provides validated methods for precise LVNC diagnosis.

Detecting train driveshaft damages using accelerometer signals and Differential Convolutional Neural Networks

Maintaining railway axles is crucial to prevent catastrophic failures and enormous human and economic costs. In recent years, there has been a growing interest in the railway industry to adopt condition monitoring techniques to enhance the safety and efficiency of the rail transport system, which maintenance is currently based on periodic inspections. In this context, this work presents a technique for real-time crack diagnosis on railway axles, based on advanced 2D-Convolutional Neural Network (CNN) architectures applied to time–frequency representations of vibration signals. One of the critical novelties is introducing a differential CNN structure that captures the system’s statistical properties, enabling generalisation between different mechanical sets and conditions. The proposed system has been trained with data corresponding to a unique wheelset assembly, showing that the model is able to diagnose cracks on the three different wheelset tested in operation under 32 different combinations of conditions, such as load, speed, sense of rotation and vibration direction. Four different crack levels have been introduced, representing the maximum one a 0.08% of the axle diameter, and the method proposed achieved Area Under the Curve (AUC) score of 0.85, significantly outperforming results obtained with other architectures proposed in the state-of-the-art, the score of the next below is 0.76. The results demonstrate the effectiveness and practicality of this approach to accurately classify the four crack levels tested within a condition monitoring system for non-stationary conditions, that would enable reliable real-time diagnosis, thus paving the way towards a more robust and efficient railway axle maintenance system.

An Optical 1×4 Power Splitter Based on Silicon-Nitride MMI Using Strip Waveguide Structures.

This paper presents a new design for a 1 × 4 optical power splitter using multimode interference (MMI) coupler in silicon nitride (Si3N4) strip waveguide structures. The main functionality of the proposed design is to use Si3N4 for dealing with the back reflection (BR) effect that usually happens in silicon (Si) MMI devices due to the self-imaging effect and the higher index contrast between Si and silicon dioxide (SiO2). The optimal device parameters were determined through numerical optimizations using the beam propagation method (BPM) and finite difference time domain (FDTD). Results demonstrate that the power splitter with a length of 34.6 μm can reach equal distribution power in each output port up to 24.3% of the total power across the O-band spectrum with 0.13 dB insertion loss and good tolerance MMI coupler parameters with a shift of ±250 nm. Additionally, the back reflection range over the O-band was found to be 40.25-42.44 dB. This demonstrates the effectiveness of the incorporation using Si3N4 MMI and adiabatic input and output tapers in mitigating unwanted BR to ensure that a good signal is received from the laser. This design showcases the significant potential for data-center networks, offering a promising solution for efficient signal distribution and facilitating high-performance and reliable optical signal routing within the O-band range. By leveraging the advantages of Si3N4 and the MMI coupler, this design opens possibilities for advanced optical network architectures and enables efficient transmission of optical signals in the O-band range.

A Dominant Interferer Plus Mean Field-Based Approximation for SINR Meta Distribution in Wireless Networks

This paper proposes a novel approach for computing the meta distribution of the signal-to-interference-plus-noise ratio (SINR) for the downlink transmission in a wireless network with Rayleigh fading. The novel approach relies on an approximation mix of exact and mean-field analysis of interference (dominant interferer-based approximation) to reduce the complexity of analysis and enhance tractability. In particular, the proposed approximation omits the need to compute the first or the second moment of the SINR that is used in the beta approximation typically adopted in the literature but requires of computing the joint distance distributions. We first derive the proposed approximation based on a Poisson point process (PPP) network with a standard path-loss and Rayleigh fading and then illustrate its accuracy and operability in another four widely used point processes: Poisson bipolar network, Matérn cluster process (MCP), <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">K</i> -tier PPP and Poisson line Cox process (PLCP). Specifically, we obtain the SINR meta distribution for PLCP networks for the first time. Even though the proposed approximation looks simple but it shows good matching in comparison to the popular beta approximation as well as the Monte-Carlo simulations, which opens the door to adopting this approximation in more advanced network architectures.

Egypt Monuments Dataset version 1: A Scalable Benchmark for Image Classification and Monument Recognition

The success of machine learning (ML) as well as deep learning (DL) depends largely on data availability and quality. The system’s performance is frequently more affected by the amount and quality of its training data than by its architecture and training specifics. Consequently, demand exists for challenging datasets that both precisely measure performance and present unique challenges with real-world applications. The Egypt Monuments Dataset v1 (EGYPT-v1) is introduced as a new scalable benchmark for fine-image classification (IC) and object recognition (OR) in the domain of ancient Egyptian monuments. EGYPT-v1 dataset is by far the world’s first large specified such dataset to date, with over seven thousand images and 40 distinct instance labels. The dataset composes different categories of monuments such as pyramids, temples, mummies, statues, head statues, bust statues, heritage sites, palaces and shrines. Several advanced deep network architectures were tested to appraise the classification difficulty in the EGYPT-v1 dataset, namely ResNet50, Inception V3, and LeNet5 models. The models achieved accuracy rates as follows: 99.13%, 90.90%, and 92.64%, respectively. The dataset was predominantly created by manually collecting images from the popular global online video-sharing and social media platform, Youtube, as well as WATCHiT, Egypt’s top streaming entertainment service. Additionally, Wikimedia Commons, the largest crowdsourced media repository in the world, was used as a secondary source of images. The images that comprise the dataset can be accessed on the GitHub repository https://github.com/mennatallahhassan/egypt-monuments-dataset.

Neural operator search

Existing neural architecture search (NAS) methods usually explore a limited feature-transformation-only search space, ignoring other advanced feature operations such as feature self-calibration by attention and dynamic convolutions. This disables the NAS algorithms to discover more advanced network architectures. We address this limitation by additionally exploiting feature self-calibration operations, resulting in a heterogeneous search space. To solve the challenges of operation heterogeneity and significantly larger search space, we formulate a neural operator search (NOS) method. NOS presents a novel heterogeneous residual block for integrating the heterogeneous operations in a unified structure, and an attention guided search strategy for facilitating the search process over a vast space. Extensive experiments show that NOS can search novel cell architectures with highly competitive performance on the CIFAR and ImageNet benchmarks.

MR-based synthetic CT image for intensity-modulated proton treatment planning of nasopharyngeal carcinoma patients

Purpose To develop an advanced deep convolutional neural network (DCNN) architecture to generate synthetic CT (SCT) images from MR images for intensity-modulated proton therapy (IMPT) treatment planning of nasopharyngeal cancer (NPC) patients. Methods T1-weighted MR images and paired CT (PCT) images were obtained from 206 NPC patients. For each patient, deformable image registration was performed between MR and PCT images to create an MR-CT image pair. Thirty pairs were randomly chosen as the independent test set and the remaining 176 pairs (14 for validation and 162 for training) were used to build two conditional generative adversarial networks (GANs): 1) GAN3D: using a 3D U-net enhanced with residual connections and attentional mechanism as the generator and 2) GAN2D: using a 2D U-net as the generator. For each test patient, SCT images were generated using the generators with the MR images as input and were compared with respect to the corresponding PCT image. A clinical IMPT plan was created and optimized on the PCT image. The dose was recalculated on the SCT images and compared with the one calculated on the PCT image. Results The mean absolute errors (MAEs) between the PCT and SCT, within the body, were (64.89 ± 5.31) HU and (64.31 ± 4.61) HU for the GAN2D and GAN3D. Within the high-density bone (HU > 600), the GAN3D achieved a smaller MAE compared with the GAN2D (p < 0.001). Within the body, the absolute point dose deviation was reduced from (0.58 ± 1.61) Gy for the GAN2D to (0.47 ± 0.94) Gy for the GAN3D. The (3 mm/3%) gamma passing rates were above 97.32% for all SCT images. Conclusions The SCT images generated using GANs achieved clinical acceptable dosimetric accuracy for IMPT of NPC patients. Using advanced DCNN architecture design, such as residual connections and attention mechanism, SCT image quality was further improved and resulted in a small dosimetric improvement.

Electronically Steerable Antennas for Future Heterogeneous Communication Networks: Review and Perspectives

Satellite, fifth generation (5G), and sixth generation (6G) mobile communication systems operating at micro- and millimeter wave frequencies are an essential pillar in advanced network architectures for high-throughput low latency services. The decisive factors for this current development were the significant technological advances accomplished over the last two decades, which meanwhile enable highly integrated, feature-rich, and cost-effective realizations of complete phased array transceiver topologies. Motivated through the today's trend for heterogeneous constellation types to provide truly global coverage, this contribution reviews the current state-of-the-art of electronically steerable antennas for terrestrial and non-terrestrial communication systems up to 100 GHz. First, the potential benefits and limitations of the most relevant technologies are contrasted and put into context with recent system architectures for adaptive beamforming. Their operating principles along with various experimental implementation and achievements providing advanced capabilities such as multi-band/multi-beam operation, polarization agility, and wide-angle scanning are thoroughly presented. Particular emphasis is laid on the review of direct radiating arrays, quasi-optical antenna configurations, and metasurface-based antennas.

Deep neural network based UAV deployment and dynamic power control for 6G-Envisioned intelligent warehouse logistics system

The intelligent warehouse logistics system (IWLS) is an essential component in the emerging industry 5.0. To well assist the IWLS, advanced network architecture and control policies with the adaptive capability to the variation of traffic should be specially designed. In this paper, we consider an air-and-ground cooperative wireless network that enables dynamic coverage to support the flexible scheduling of the automatic guided vehicles (AGVs) in the IWLS. Jointly considering the dynamic deployment of unmanned aerial vehicles (UAVs) as well as the adaptive power control of AGVs, we formulate a two time-scale network control problem to minimize the transmission power consumption of all AGVs under their individual rate requirement. On large time scales, we first propose a particle swarm optimization based algorithm (PSOA) to obtain the deployment position of the ABS. Then, using the results of the PSOA as training data, we design a deep neural network (DNN) framework aimed at reducing the computational time of the PSOA. On small time scales, we devise an online power control algorithm (OPCA) by using some of stochastic network optimization methods. With current channel conditions, the OPCA can generate the real-time power control policy and ensure the long-term rate requirement. Numerical simulations indicate that the DNN framework enhances the coverage performance of the network only consuming a few milliseconds of computation time. Incorporated with the OPCA, the total transmission power of the AGVs is significantly reduced.

Guest Editorial Machine Learning in Antenna Design, Modeling, and Measurements

Machine learning (ML) is the study of computational methods for improving performance by mechanizing the acquisition of knowledge from experience. As a modern data-driven optimization and applied regression methodology, ML aims to provide increasing levels of automation in the knowledge engineering process, replacing much time-consuming human activity with automatic techniques that improve accuracy and/or efficiency by discovering and exploiting regularities in training data. Indeed, many ML and data-driven methods, such as conventional artificial neural networks (ANNs), were introduced and studied within electromagnetics a few decades ago. However, these past studies did not benefit from the most recent advances in ML, which have been driven by the present confluence of improved hardware performance at lower cost, advanced network algorithms and architectures, data science, and considerable efforts dedicated to advancing the computational electromagnetics (CEM) benchmark. Today, a broader family of ML techniques based on ANNs has been developed. Examples include deep neural network (DNN), convolutional neural network (CNN), recurrent neural network, generative adversarial network, and deep reinforcement learning, which have been successfully applied to different engineering and science problems, ranging from image and video recognition, social media services, virtual personal assistant to autonomous vehicles, to name a few. This naturally suggests that applying ML to real-world electromagnetic problems could be one of the emerging trends in ML and artificial intelligence (AI) <xref ref-type="bibr" rid="ref1" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[1]</xref> – <xref ref-type="bibr" rid="ref2" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"/> <xref ref-type="bibr" rid="ref3" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"/> <xref ref-type="bibr" rid="ref4" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[4]</xref> . Indeed, ML has been becoming an important complement to existing experimental, computational, and theoretical aspects of electromagnetics.

Real-Time Multi-Class Infection Classification for Respiratory Diseases

Real-time disease prediction has emerged as the main focus of study in the field of computerized medicine. Intelligent disease identification framework can assist medical practitioners in diagnosing disease in a way that is reliable, consistent, and timely, successfully lowering mortality rates, particularly during endemics and pandemics. To prevent this pandemic’s rapid and widespread, it is vital to quickly identify, confine, and treat affected individuals. The need for auxiliary computer-aided diagnostic (CAD) systems has grown. Numerous recent studies have indicated that radiological pictures contained critical information regarding the COVID-19 virus. Utilizing advanced convolutional neural network (CNN) architectures in conjunction with radiological imaging makes it possible to provide rapid, accurate, and extremely useful susceptible classifications. This research work proposes a methodology for real-time detection of COVID-19 infections caused by the Corona Virus. The purpose of this study is to offer a two-way COVID-19 (2WCD) diagnosis prediction deep learning system that is built on Transfer Learning Methodologies (TLM) and features customized fine-tuning on top of fully connected layered pre-trained CNN architectures. 2WCD has applied modifications to pre-trained models for better performance. It is designed and implemented to improve the generalization ability of the classifier for binary and multi-class models. Along with the ability to differentiate COVID-19 and No-Patient in the binary class model and COVID-19, No-Patient, and Pneumonia in the multi-class model, our framework is augmented with a critical add-on for visually demonstrating infection in any tested radiological image by highlighting the affected region in the patient’s lung in a recognizable color pattern. The proposed system is shown to be extremely robust and reliable for real-time COVID-19 diagnostic prediction. It can also be used to forecast other lung-related disorders. As the system can assist medical practitioners in diagnosing the greatest number of patients in the shortest amount of time, radiologists can also be used or published online to assist any less-experienced individual in obtaining an accurate immediate screening for their radiological images.

DisOptNet: Distilling Semantic Knowledge From Optical Images for Weather-Independent Building Segmentation

Synthetic aperture radar (SAR) images provide all-weather and all-time capabilities for Earth observation, which becomes highly beneficial in the field of intelligent remote sensing (RS) image interpretation. Due to these advantages, SAR images have been widely exploited in automatic building segmentation tasks under poor weather conditions, especially when disasters happen. However, compared to optical images, the semantics inherent to SAR images are less rich and interpretable due to factors such as speckle noise and imaging geometry. In this scenario, most state-of-the-art methods are focused on designing advanced network architectures or loss functions for building footprint extraction. However, few works have been oriented toward improving segmentation performance through knowledge transfer from optical images. In this article, we propose a novel method based on the <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DisOptNet</monospace> network, which can distill the useful semantic knowledge from optical images into a network only trained with SAR data. Specifically, we first analyze the multilevel feature discrepancies between multiple stages of the networks pretrained on the two image modalities. We observe that feature discrepancies start to increase as the encoding stage gradually changes from low level to high level. Based on such observation, we reuse the early stage features and construct parallel convolutional neural network (CNN) branches that are responsible for capturing high-level domain-specific knowledge for each image modality. The optical branch is aimed at mimicking feature generation at the optical pretrained network given the input SAR images. Then, an aggregation module is introduced to calibrate and fuse the features from different modalities while generating the building segments. Extensive experiments were conducted on a large-scale multisensor all-weather building segmentation dataset with state-of-the-art methods used for comparison. Our experimental results validate the effectiveness of <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DisOptNet</monospace> , which demonstrates great potential in the task of weather-independent building footprint generation under real scenarios. The codes of this article will be made publicly available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/jiankang1991/TGRS_DisOptNet</uri> .

An ultra-fast bi-phase advanced network for segmenting crop plants from dense weeds

Fast and accurate perception of crop plants is critical for robotic weeding. However, due to the similar appearance, it remains challenging to precisely and efficiently segment crop plants from weeds, especially when the weed density is high. To tackle this problem, we devise an ultra-fast bi-phase advanced network (BA-Net) architecture that advances both the feature extraction phase and the model training phase. A bi-path fusion tree structure and a bi-polar distributed loss are introduced for the two phases, respectively. The bi-path fusion tree merges features from two adjacent convolutional stages at each step and efficiently models the scene through a tree-like structure. The bi-polar distributed loss strengthens the training at ambiguous areas, which can greatly enhance the distinctness at detailed structures like the crop plant boundaries. Experiments demonstrate the superiority of the proposed BA-Net to other state-of-the-art methods in terms of both accuracy and efficiency. BA-Net can accurately segment the crop plants from the weedy background, achieving an intersection over union score of 0.963, a F-measure score of 0.981, and a mean absolute error of 0.00376. Also, BA-Net is lightweight with only 0.55 M trainable parameters and achieves ultra-fast speed, i.e. , over 340 FPS on a desktop and over 26 FPS on a Jetson TX2 embedded computer. The code of BA-Net is available at https://github.com/ZhangXG001/BA-Net . • An ultra-fast deep network is proposed for segmenting crop plants from dense weeds. • The bi-path fusion tree structure can perform efficient multi-scale feature fusion. • The bi-polar distributed loss can improve the segmentation at the obscure areas. • The network achieves superior accuracy and efficiency over state-of-the-arts. • The network runs over 26 FPS on a Jetson TX2 embedded computer.

МЕТОД ПІДВИЩЕННЯ ЕФЕКТИВНОСТІ РОЗПОДІЛЕНОЇ ОБРОБКИ ДАНИХ У КОМП’ЮТЕРНИХ СИСТЕМАХ ОПЕРАТОРІВ СТІЛЬНИКОВОГО ЗВ’ЯЗКУ

За аналізом якості обслуговування абонентів у реалізованих проектах мереж 5G в світі встановлено, що заявлені вимоги до мереж п’ятого покоління не досягаються в жодній із реалізованих мереж, а фактична якість обслуговування абонентів перебуває на не досить високому рівні, що свідчить про низьку ефективність наявних методів керування мережами стільникового оператора та розподіленої обробки даних у них. Тому було розроблено метод оптимізації розміщення масштабованих послуг на розподілених обчислювальних ресурсах мережі стільникового оператора. Також було проведено імітаційне моделювання для оцінювання ефективності розробленого у роботі методу, результати якого підтвердили його ефективність.

Explainable COVID-19 Detection Using Chest CT Scans and Deep Learning.

This paper explores how well deep learning models trained on chest CT images can diagnose COVID-19 infected people in a fast and automated process. To this end, we adopted advanced deep network architectures and proposed a transfer learning strategy using custom-sized input tailored for each deep architecture to achieve the best performance. We conducted extensive sets of experiments on two CT image datasets, namely, the SARS-CoV-2 CT-scan and the COVID19-CT. The results show superior performances for our models compared with previous studies. Our best models achieved average accuracy, precision, sensitivity, specificity, and F1-score values of , , , , and on the SARS-CoV-2 dataset, and , , , , and on the COVID19-CT dataset, respectively. For better interpretability of the results, we applied visualization techniques to provide visual explanations for the models’ predictions. Feature visualizations of the learned features show well-separated clusters representing CT images of COVID-19 and non-COVID-19 cases. Moreover, the visualizations indicate that our models are not only capable of identifying COVID-19 cases but also provide accurate localization of the COVID-19-associated regions, as indicated by well-trained radiologists.