RESUMEN
The current study aims to improve the efficiency of automatic identification of pavement distress and improve the status quo of difficult identification and detection of pavement distress. First, the identification method of pavement distress and the types of pavement distress are analysed. Then, the design concept of deep learning in pavement distress recognition is described. Finally, the mask region-based convolutional neural network (Mask R-CNN) model is designed and applied in the recognition of road crack distress. The results show that in the evaluation of the model's comprehensive recognition performance, the highest accuracy is 99%, and the lowest accuracy is 95% after the test and evaluation of the designed model in different datasets. In the evaluation of different crack identification and detection methods, the highest accuracy of transverse crack detection is 98% and the lowest accuracy is 95%. In longitudinal crack detection, the highest accuracy is 98% and the lowest accuracy is 92%. In mesh crack detection, the highest accuracy is 98% and the lowest accuracy is 92%. This work not only provides an in-depth reference for the application of deep CNNs in pavement distress recognition but also promotes the improvement of road traffic conditions, thus contributing to the progression of smart cities in the future. This article is part of the theme issue 'Artificial intelligence in failure analysis of transportation infrastructure and materials'.
RESUMEN
With the rapid development of the Internet of Things (IoT), the frequency of attackers using botnets to control IoT devices in order to perform distributed denial-of-service attacks (DDoS) and other cyber attacks on the internet has significantly increased. In the actual attack process, the small percentage of attack packets in IoT leads to low accuracy of intrusion detection. Based on this problem, the paper proposes an oversampling algorithm, KG-SMOTE, based on Gaussian distribution and K-means clustering, which inserts synthetic samples through Gaussian probability distribution, extends the clustering nodes in minority class samples in the same proportion, increases the density of minority class samples, and improves the amount of minority class sample data in order to provide data support for IoT-based DDoS attack detection. Experiments show that the balanced dataset generated by this method effectively improves the intrusion detection accuracy in each category and effectively solves the data imbalance problem.
RESUMEN
Realizing a densely packed waveguide antenna array is of great importance in light detection and ranging (LIDAR), owing to its suppressed grating lobes. In this work, a low-cross-talk half-wavelength pitch silicon waveguide array is proposed and experimentally demonstrated. It has a periodic arrangement of silicon strip nanophotonic waveguides, between which deep-subwavelength silicon strips are placed. Our experimental results show that this array's cross talk suppression is nearly 20 dB and has a bandwidth covering a wavelength range from 1500 nm to 1560 nm. Our realization of a half-wavelength pitch waveguide array may offer a promising platform for studying integrated optical phased arrays for solid-state LIDAR with a very low grating lobe and thus potentially a large field of view.
RESUMEN
The purposes are to explore the effect of Digital Twins (DTs) in Unmanned Aerial Vehicles (UAVs) on providing medical resources quickly and accurately during COVID-19 prevention and control. The feasibility of UAV DTs during COVID-19 prevention and control is analyzed. Deep Learning (DL) algorithms are introduced. A UAV DTs information forecasting model is constructed based on improved AlexNet, whose performance is analyzed through simulation experiments. As end-users and task proportion increase, the proposed model can provide smaller transmission delays, lesser energy consumption in throughput demand, shorter task completion time, and higher resource utilization rate under reduced transmission power than other state-of-art models. Regarding forecasting accuracy, the proposed model can provide smaller errors and better accuracy in Signal-to-Noise Ratio (SNR), bit quantizer, number of pilots, pilot pollution coefficient, and number of different antennas. Specifically, its forecasting accuracy reaches 95.58% and forecasting velocity stabilizes at about 35 Frames-Per-Second (FPS). Hence, the proposed model has stronger robustness, making more accurate forecasts while minimizing the data transmission errors. The research results can reference the precise input of medical resources for COVID-19 prevention and control.
RESUMEN
A novel, to the best of our knowledge, method to extract optical microring resonators' loss characteristics is proposed and demonstrated using optical frequency domain reflectometry (OFDR). Compared with the traditional optical transmission measurement method, the spatial-resolved backscattering optical signals obtained from the OFDR can clearly show the resonance mode's increased optical path length due to its circulation inside the resonator. By further processing the backscattered optical signals, loaded $Q$-factors of several resonators can be accurately determined. A calculation model is proposed to derive the resonance mode's intrinsic $Q$-factor from OFDR measurements of a series of loaded resonators.
RESUMEN
An ultra-compact half-wavelength pitch silicon waveguide array with very low crosstalk is proposed and analyzed in this work. We first show the design of a pair of low-crosstalk silicon waveguides with only half-wavelength spacing, where the placement of two thin silicon strips asymmetrically in between the waveguides is key to having very low crosstalk. We next extend this nano-structured two-waveguide design to form a low-crosstalk half-wavelength pitch silicon waveguide array. Coupled-mode theory shows that, for an array length of 1 mm, the insertion loss of the input waveguide is as low as -0.13 dB for the TE-like mode at 1550 nm, and the crosstalk in all other waveguides remains below about -18 dB. This half-wavelength pitch waveguide array also exhibits a favorable fabrication error tolerance when taking into account the waveguide width variations in practice. It offers a promising platform for realization of integrated optical phased arrays for solid-state lidars with a large field of view.
RESUMEN
A machine learning assisted modal power analyzing scheme designed for optical modes in integrated multi-mode waveguides is proposed and studied in this work. Convolutional neural networks (CNNs) are successfully trained to correlate the far-field diffraction intensity patterns of a superposition of multiple waveguide modes with its modal power distribution. In particular, a specialized CNN is trained to analyze thin optical waveguides, which are single-moded along one axis and multi-moded along the other axis. A full-scale CNN is also trained to cross-validate the results obtained from this specialized CNN model. Prediction accuracy for modal power is benchmarked statistically with square error and absolute error distribution. It is found that the overall accuracy of our trained specialized CNN is very satisfactory for thin optical waveguides while that of our trained full-scale CNN remains nearly unchanged but the training time doubles. This approach is further generalized and applied to a waveguide that is multi-moded along both horizontal and vertical axes and the influence of noise on our trained network is studied. Overall, we find that the performance in this general condition keeps nearly unchanged. This new concept of analyzing modal power may open the door for high fidelity information recovery in far field and holds great promise for potential applications in both integrated and fiber-based spatial-division demultiplexing.
RESUMEN
Waveguide crossing is an important integrated photonic component that will be routinely used for high-density and large-scale photonic integrated circuits, such as optical switches and routers. Several techniques have been reported in achieving high performance waveguide crossings on a silicon-on-insulator photonic platform, i.e., low-loss and low-crosstalk waveguide crossings based on multimode interference, bi-layer tapering, optical transformation, metamaterials, and subwavelength gratings. Until recently, not much attention has been given to the reduction of the footprint of waveguide crossings. Here we experimentally demonstrate an ultra-compact waveguide crossing on silicon photonic platform with a footprint only ~1 × 1 µm2. Our simulations show that it has a low insertion loss (< 0.175 dB) and low crosstalk (< -37dB) across the whole C-band, while the fabricated one has an insertion loss < 0.28 dB and crosstalk around -30 dB for the C-band.
RESUMEN
In this paper, a spectral model by incorporating SRS effect is proposed and established, which is feasible for analyzing the SRS effect both in high-power fiber oscillator and master oscillator power amplifier (MOPA) system. The theoretical results show that the SRS effect is tightly related to the bandwidths of the fiber Bragg gratings (FBGs) and it can be efficiently suppressed by optimizing the bandwidth of the FBGs. Besides, the established theoretical model is also feasible for analyzing the influence of seed power on the SRS effect. The theoretical predictions agree well with the previous experimental results.
RESUMEN
In this paper the stimulated Raman scattering (SRS) effect in high-power fiber amplifiers seeded by the narrow-band filtered superfluorescent source (SFS) is firstly analyzed both theoretically and experimentally. Spectral models for the formation of the SFS and the spectral evolution in high-power fiber amplifiers seeded by filtered SFS are proposed. It is found that the SRS effect in high-power fiber amplifiers depends on the spectral width of the filtered SFS seed. The theoretical predictions are in qualitative agreements with the experimental results.
RESUMEN
We present a new method of SBS suppression in fiber amplifier system by employing simultaneously phase and intensity modulation. In this way, a GHz narrow-linewidth polarization-maintaining (PM) all-fiber pulsed laser is obtained based on a master oscillator power amplifier (MOPA) configuration. The pulsed seed is generated from a single-frequency continuous wave (CW) laser at 1064 nm by simultaneous modulation using an electro-optic intensity modulator (EOIM) and an electro-optic phase modulator (EOPM). Theoretical model is built and simulation framework has been established to estimate the SBS threshold of the pulsed amplifier system before and after modulation. In experiment, in order to suppress SBS effectively, the pulse width is set to be 4 ns and the phase modulation voltage is set to be 5 V. After amplifying by the amplifier chain, a ~3.5 ns pulsed laser with average/peak power of 293 W/3.9 kW is obtained at intensity repetition rate of 20 MHz and phase repetition rate of 100MHz, showing good agreement with simulation results. The linewidth of the output laser is ~4.5 GHz, the M(2) factor at maximal output power is measured to be ~1.1 and the slope efficiency is ~86%.This method provides some references to suppress the SBS in narrow linewidth pulsed amplifier systems.
RESUMEN
We demonstrate a direct diode-pumped all-fiber-integrated fiber laser based on master oscillator power amplifier configuration at 1080 nm, producing maximum output power of 3.15 kW with corresponding optical to optical efficiency of 75.1%. Further power scaling is pump-limited and theoretical analysis demonstrates that 4 kW output power can be further achieved without stimulated Raman scattering. Near diffraction-limited beam quality (M(2) ~ 1.6 in the x and y directions) is also achieved at the maximum output power. This compact prototype laser has excellent stability and reliability, which could benefit many practical applications, such as industrial processing.
RESUMEN
With the rapid development of information technology, great changes have taken place in the way of managing, analyzing, and using data in all walks of life. Using deep learning algorithm for data analysis in the field of medicine can improve the accuracy of disease recognition. The purpose is to realize the intelligent medical service mode of sharing medical resources among many people under the dilemma of limited medical resources. Firstly, the Digital Twins module in the Deep Learning algorithm is used to establish the medical care and disease auxiliary diagnosis model. With the help of the digital visualization model of Internet of Things technology, data is collected at the client and server. Based on the improved Random Forest algorithm, the demand analysis and target function design of the medical and health care system are carried out. Based on data analysis, the medical and health care system is designed using the improved algorithm. The results show that the intelligent medical service platform can collect and analyze the clinical trial data of patients. The accuracy of improved ReliefF & Wrapper Random Forest (RW-RF) for sepsis disease recognition can reach about 98%, and the accuracy of algorithm for disease recognition is also more than 80%, which can provide better technical support for disease recognition and medical care services. It provides a solution and experimental reference for the practical problem of scarce medical resources.
RESUMEN
OBJECTIVE: it aims to adopt deep transfer learning combined with Digital Twins (DTs) in Magnetic Resonance Imaging (MRI) medical image enhancement. METHODS: MRI image enhancement method based on metamaterial composite technology is proposed by analyzing the application status of DTs in medical direction and the principle of MRI imaging. On the basis of deep transfer learning, MRI super-resolution deep neural network structure is established. To address the problem that different medical imaging methods have advantages and disadvantages, a multi-mode medical image fusion algorithm based on adaptive decomposition is proposed and verified by experiments. RESULTS: the optimal Peak Signal to Noise Ratio (PSNR) of 34.11dB can be obtained by introducing modified linear element and loss function of deep transfer learning neural network structure. The Structural Similarity Coefficient (SSIM) is 85.24%. It indicates that the MRI truthfulness and sharpness obtained by adding composite metasurface are improved greatly. The proposed medical image fusion algorithm has the highest overall score in the subjective evaluation of the six groups of fusion image results. Group III had the highest score in Magnetic Resonance Imaging- Positron Emission Computed Tomography (MRI-PET) image fusion, with a score of 4.67, close to the full score of 5. As for the objective evaluation in group I of Magnetic Resonance Imaging- Single Photon Emission Computed Tomography (MRI-SPECT) images, the Root Mean Square Error (RMSE), Relative Average Spectral Error (RASE) and Spectral Angle Mapper (SAM) are the highest, which are 39.2075, 116.688, and 0.594, respectively. Mutual Information (MI) is 5.8822. CONCLUSION: the proposed algorithm has better performance than other algorithms in preserving spatial details of MRI images and color information direction of SPECT images, and the other five groups have achieved similar results.
RESUMEN
Lignocellulose is an abundant xylose-containing biomass found in agricultural wastes, and has arisen as a suitable alternative to fossil fuels for the production of bioethanol. Although Saccharomyces cerevisiae has been thoroughly used for the production of bioethanol, its potential to utilize lignocellulose remains poorly understood. In this work, xylose-metabolic genes of Pichia stipitis and Candida tropicalis, under the control of different promoters, were introduced into S. cerevisiae. RNA-seq analysis was use to examine the response of S. cerevisiae metabolism to the introduction of xylose-metabolic genes. The use of the PGK1 promoter to drive xylitol dehydrogenase (XDH) expression, instead of the TEF1 promoter, improved xylose utilization in "XR-pXDH" strain by overexpressing xylose reductase (XR) and XDH form C. tropicalis, enhancing the production of xylitol (13.66 ± 0.54 g/L after 6 days fermentation). Overexpression of xylulokinase and XR/XDH from P. stipitis remarkably decreased xylitol accumulation (1.13 ± 0.06 g/L and 0.89 ± 0.04 g/L xylitol, respectively) and increased ethanol production (196.14 % and 148.50 % increases during the xylose utilization stage, respectively), in comparison with the results of XR-pXDH. This result may be produced due to the enhanced xylose transport, Embden-Meyerhof and pentose phosphate pathways, as well as alleviated oxidative stress. The low xylose consumption rate in these recombinant as well as alleviated strains comparing with P. stipitis and C. tropicalis may be explained by the insufficient supplementation of NADPH and NAD +. The results obtained in this work provide new insights on the potential utilization of xylose using bioengineered S. cerevisiae strains.
Asunto(s)
Saccharomyces cerevisiae , Xilosa , Aldehído Reductasa/genética , Aldehído Reductasa/metabolismo , Candida/genética , Candida/metabolismo , D-Xilulosa Reductasa/genética , D-Xilulosa Reductasa/metabolismo , Fermentación , Pichia/genética , Pichia/metabolismo , Saccharomyces cerevisiae/genética , Saccharomyces cerevisiae/metabolismo , Xilitol/metabolismo , Xilosa/metabolismoRESUMEN
The purpose is to explore the feature recognition, diagnosis, and forecasting performances of Semi-Supervised Support Vector Machines (S3VMs) for brain image fusion Digital Twins (DTs). Both unlabeled and labeled data are used regarding many unlabeled data in brain images, and semi supervised support vector machine (SVM) is proposed. Meantime, the AlexNet model is improved, and the brain images in real space are mapped to virtual space by using digital twins. Moreover, a diagnosis and prediction model of brain image fusion digital twins based on semi supervised SVM and improved AlexNet is constructed. Magnetic Resonance Imaging (MRI) data from the Brain Tumor Department of a Hospital are collected to test the performance of the constructed model through simulation experiments. Some state-of-art models are included for performance comparison: Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), AlexNet, and Multi-Layer Perceptron (MLP). Results demonstrate that the proposed model can provide a feature recognition and extraction accuracy of 92.52%, at least an improvement of 2.76% compared to other models. Its training lasts for about 100 s, and the test takes about 0.68 s. The Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) of the proposed model are 4.91 and 5.59%, respectively. Regarding the assessment indicators of brain image segmentation and fusion, the proposed model can provide a 79.55% Jaccard coefficient, a 90.43% Positive Predictive Value (PPV), a 73.09% Sensitivity, and a 75.58% Dice Similarity Coefficient (DSC), remarkably better than other models. Acceleration efficiency analysis suggests that the improved AlexNet model is suitable for processing massive brain image data with a higher speedup indicator. To sum up, the constructed model can provide high accuracy, good acceleration efficiency, and excellent segmentation and recognition performances while ensuring low errors, which can provide an experimental basis for brain image feature recognition and digital diagnosis.
RESUMEN
The present work aims to explore the performance of fuzzy system-based medical image processing for predicting the brain disease. The imaging mechanism of NMR (Nuclear Magnetic Resonance) and the complexity of human brain tissues cause the brain MRI (Magnetic Resonance Imaging) images to present varying degrees of noise, weak boundaries, and artifacts. Hence, improvements are made over the fuzzy clustering algorithm. A brain image processing and brain disease diagnosis prediction model is designed based on improved fuzzy clustering and HPU-Net (Hybrid Pyramid U-Net Model for Brain Tumor Segmentation) to ensure the model safety performance. Brain MRI images collected from a Hospital, are employed in simulation experiments to validate the performance of the proposed algorithm. Moreover, CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), FCM (Fuzzy C-Means), LDCFCM (Local Density Clustering Fuzzy C-Means), and AFCM (Adaptive Fuzzy C-Means) are included in simulation experiments for performance comparison. Results demonstrate that the proposed algorithm has more nodes, lower energy consumption, and more stable changes than other models under the same conditions. Regarding the overall network performance, the proposed algorithm can complete the data transmission tasks the fastest, basically maintaining at about 4.5 s on average, which performs remarkably better than other models. A further prediction performance analysis reveals that the proposed algorithm provides the highest prediction accuracy for the Whole Tumor under DSC (Dice Similarity Coefficient), reaching 0.936. Besides, its Jaccard coefficient is 0.845, proving its superior segmentation accuracy over other models. In a word, the proposed algorithm can provide higher accuracy, a more apparent denoising effect, and the best segmentation and recognition effect than other models while ensuring energy consumption. The results can provide an experimental basis for the feature recognition and predictive diagnosis of brain images.