Dear Editor,The letter proposes a tensor low-rank orthogonal compression(TLOC)model for a convolutional neural network(CNN),which facilitates its efficient and highly-accurate low-rank representation.Model compression...Dear Editor,The letter proposes a tensor low-rank orthogonal compression(TLOC)model for a convolutional neural network(CNN),which facilitates its efficient and highly-accurate low-rank representation.Model compression is crucial for deploying deep neural network(DNN)models on resource-constrained embedded devices.展开更多
Convolutional neural networks(CNNs)have shown remarkable success across numerous tasks such as image classification,yet the theoretical understanding of their convergence remains underdeveloped compared to their empir...Convolutional neural networks(CNNs)have shown remarkable success across numerous tasks such as image classification,yet the theoretical understanding of their convergence remains underdeveloped compared to their empirical achievements.In this paper,the first filter learning framework with convergence-guaranteed learning laws for end-to-end learning of deep CNNs is proposed.Novel update laws with convergence analysis are formulated based on the mathematical representation of each layer in convolutional neural networks.The proposed learning laws enable concurrent updates of weights across all layers of the deep convolutional neural network and the analysis shows that the training errors converge to certain bounds which are dependent on the approximation errors.Case studies are conducted on benchmark datasets and the results show that the proposed concurrent filter learning framework guarantees the convergence and offers more consistent and reliable results during training with a trade-off in performance compared to stochastic gradient descent methods.This framework represents a significant step towards enhancing the reliability and effectiveness of deep convolutional neural network by developing a theoretical analysis which allows practical implementation of the learning laws with automatic tuning of the learning rate to guarantee the convergence during training.展开更多
Landslide susceptibility mapping(LSM)plays a crucial role in assessing geological risks.The current LSM techniques face a significant challenge in achieving accurate results due to uncertainties associated with region...Landslide susceptibility mapping(LSM)plays a crucial role in assessing geological risks.The current LSM techniques face a significant challenge in achieving accurate results due to uncertainties associated with regional-scale geotechnical parameters.To explore rainfall-induced LSM,this study proposes a hybrid model that combines the physically-based probabilistic model(PPM)with convolutional neural network(CNN).The PPM is capable of effectively capturing the spatial distribution of landslides by incorporating the probability of failure(POF)considering the slope stability mechanism under rainfall conditions.This significantly characterizes the variation of POF caused by parameter uncertainties.CNN was used as a binary classifier to capture the spatial and channel correlation between landslide conditioning factors and the probability of landslide occurrence.OpenCV image enhancement technique was utilized to extract non-landslide points based on the POF of landslides.The proposed model comprehensively considers physical mechanics when selecting non-landslide samples,effectively filtering out samples that do not adhere to physical principles and reduce the risk of overfitting.The results indicate that the proposed PPM-CNN hybrid model presents a higher prediction accuracy,with an area under the curve(AUC)value of 0.85 based on the landslide case of the Niangniangba area of Gansu Province,China compared with the individual CNN model(AUC=0.61)and the PPM(AUC=0.74).This model can also consider the statistical correlation and non-normal probability distributions of model parameters.These results offer practical guidance for future research on rainfall-induced LSM at the regional scale.展开更多
The integration of image analysis through deep learning(DL)into rock classification represents a significant leap forward in geological research.While traditional methods remain invaluable for their expertise and hist...The integration of image analysis through deep learning(DL)into rock classification represents a significant leap forward in geological research.While traditional methods remain invaluable for their expertise and historical context,DL offers a powerful complement by enhancing the speed,objectivity,and precision of the classification process.This research explores the significance of image data augmentation techniques in optimizing the performance of convolutional neural networks(CNNs)for geological image analysis,particularly in the classification of igneous,metamorphic,and sedimentary rock types from rock thin section(RTS)images.This study primarily focuses on classic image augmentation techniques and evaluates their impact on model accuracy and precision.Results demonstrate that augmentation techniques like Equalize significantly enhance the model's classification capabilities,achieving an F1-Score of 0.9869 for igneous rocks,0.9884 for metamorphic rocks,and 0.9929 for sedimentary rocks,representing improvements compared to the baseline original results.Moreover,the weighted average F1-Score across all classes and techniques is 0.9886,indicating an enhancement.Conversely,methods like Distort lead to decreased accuracy and F1-Score,with an F1-Score of 0.949 for igneous rocks,0.954 for metamorphic rocks,and 0.9416 for sedimentary rocks,exacerbating the performance compared to the baseline.The study underscores the practicality of image data augmentation in geological image classification and advocates for the adoption of DL methods in this domain for automation and improved results.The findings of this study can benefit various fields,including remote sensing,mineral exploration,and environmental monitoring,by enhancing the accuracy of geological image analysis both for scientific research and industrial applications.展开更多
Severe acute respiratory syndrome coronavirus 2(SARS-CoV-2)mutations are influenced by random and uncontrollable factors,and the risk of the next widespread epidemic remains.Dual-target drugs that synergistically act ...Severe acute respiratory syndrome coronavirus 2(SARS-CoV-2)mutations are influenced by random and uncontrollable factors,and the risk of the next widespread epidemic remains.Dual-target drugs that synergistically act on two targets exhibit strong therapeutic effects and advantages against mutations.In this study,a novel computational workflow was developed to design dual-target SARS-CoV-2 candidate inhibitors with the Envelope protein and Main protease selected as the two target proteins.The drug-like molecules of our self-constructed 3D scaffold database were used as high-throughput molecular docking probes for feature extraction of two target protein pockets.A multi-layer perceptron(MLP)was employed to embed the binding affinities into a latent space as conditional vectors to control conditional distribution.Utilizing a conditional generative neural network,cG-SchNet,with 3D Euclidean group(E3)symmetries,the conditional probability distributions of molecular 3D structures were acquired and a set of novel SARS-CoV-2 dual-target candidate inhibitors were generated.The 1D probability,2D joint probability,and 2D cumulative probability distribution results indicate that the generated sets are significantly enhanced compared to the training set in the high binding affinity area.Among the 201 generated molecules,42 molecules exhibited a sum binding affinity exceeding 17.0 kcal/mol while 9 of them having a sum binding affinity exceeding 19.0 kcal/mol,demonstrating structure diversity along with strong dual-target affinities,good absorption,distribution,metabolism,excretion,and toxicity(ADMET)properties,and ease of synthesis.Dual-target drugs are rare and difficult to find,and our“high-throughput docking-multi-conditional generation”workflow offers a wide range of options for designing or optimizing potent dual-target SARS-CoV-2 inhibitors.展开更多
The migration of tasks aided by machine learning(ML)predictions IN(DPM)is a system-level design technique that is used to reduce energy by enhancing the overall performance of the processor.In this paper,we address th...The migration of tasks aided by machine learning(ML)predictions IN(DPM)is a system-level design technique that is used to reduce energy by enhancing the overall performance of the processor.In this paper,we address the issue of system-level higher task dissipation during the execution of parallel workloads with common deadlines by introducing a machine learning-based framework that includes task migration using energy-efficient earliest deadline first scheduling(EA-EDF).ML-based EA-EDF enhances the overall throughput and optimizes the energy to avoid delay and performance degradation in a multiprocessor system.The proposed system model allocates processors to the ready task set in such a way that their deadlines are guaranteed.A full task migration policy is also integrated to ensure proper task mapping that ensures inter-process linkage among the arrived tasks with the same deadlines.The execution of a task can halt on one CPU and reschedule the execution on a different processor to avoid delay and ensure meeting the deadline.Our approach shows promising potential for machine-learning-based schedulability analysis enables a comparison between different ML models and shows a promising reduction in energy as compared with other ML-aware task migration techniques for SoC like Multi-Layer Feed-Forward Neural Networks(MLFNN)based on convolutional neural network(CNN),Random Forest(RF)and Deep learning(DL)algorithm.The Simulations are conducted using super pipelined microarchitecture of advanced micro devices(AMD)XScale PXA270 using instruction and data cache per core 32 Kbyte I-cache and 32 Kbyte D-cache on various utilization factors(u_(i))12%,31%and 50%.The proposed approach consumes 5.3%less energy when almost half of the CPU is running and on a lower workload consumes 1.04%less energy.The proposed design accumulatively gives significant improvements by reducing the energy dissipation on three clock rates by 4.41%,on 624 MHz by 5.4%and 5.9%on applications operating on 416 and 312 MHz standard operating frequencies.展开更多
Hemoglobin is a vital protein in red blood cells responsible for transporting oxygen throughout the body.Its accurate measurement is crucial for diagnosing and managing conditions such as anemia and diabetes,where abn...Hemoglobin is a vital protein in red blood cells responsible for transporting oxygen throughout the body.Its accurate measurement is crucial for diagnosing and managing conditions such as anemia and diabetes,where abnormal hemoglobin levels can indicate significant health issues.Traditional methods for hemoglobin measurement are invasive,causing pain,risk of infection,and are less convenient for frequent monitoring.PPG is a transformative technology in wearable healthcare for noninvasive monitoring and widely explored for blood pressure,sleep,blood glucose,and stress analysis.In this work,we propose a hemoglobin estimation method using an adaptive lightweight convolutional neural network(HMALCNN)from PPG.The HMALCNN is designed to capture both fine-grained local waveform characteristics and global contextual patterns,ensuring robust performance across acquisition settings.We validated our approach on two multi-regional datasets containing 152 and 68 subjects,respectively,employing a subjectindependent 5-fold cross-validation strategy.The proposed method achieved root mean square errors(RMSE)of 0.90 and 1.20 g/dL for the two datasets,with strong Pearson correlations of 0.82 and 0.72.We conducted extensive posthoc analyses to assess clinical utility and interpretability.A±1 g/dL clinical error tolerance evaluation revealed that 91.3%and 86.7%of predictions for the two datasets fell within the acceptable clinical range.Hemoglobin range-wise analysis demonstrated consistently high accuracy in the normal and low hemoglobin categories.Statistical significance testing using the Wilcoxon signed-rank test confirmed the stability of performance across validation folds(p>0.05 for both RMSE and correlation).Furthermore,model interpretability was enhanced using Gradient-weighted Class Activation Mapping(Grad-CAM),supporting the model’s clinical trustworthiness.The proposed HMALCNN offers a computationally efficient,clinically interpretable,and generalizable framework for noninvasive hemoglobin monitoring,with strong potential for integration into wearable healthcare systems as a practical alternative to invasive measurement techniques.展开更多
To enhance the inference efficiency of convolutional neural network(CNN),tensor parallelism is employed to improve the parallelism within operators.However,existing methods are customized to specific networks and hard...To enhance the inference efficiency of convolutional neural network(CNN),tensor parallelism is employed to improve the parallelism within operators.However,existing methods are customized to specific networks and hardware,limiting their generalizability.This paper proposes an approach called resource-adaptive tensor decomposition(RATD)for CNN operators,which aims to achieve an optimal match between computational resources and parallel computing tasks.Firstly,CNN is represented with fine-grained tensors at the lower graph level,thereby decoupling tensors that can be computed in parallel within operators.Secondly,the convolution and pooling operators are fused,and the decoupled tensor blocks are scheduled in parallel.Finally,a cost model is constructed,based on runtime and resource utilization,to iteratively refine the process of tensor block decomposition and automatically determine the optimal tensor decomposition.Experimental results demonstrate that the proposed RATD improves the accuracy of the model by 11%.Compared with CUDA(compute unified device architecture)deep neural network library(cuDNN),RATD achieves an average speedup ratio of 1.21 times in inference time across various convolution kernels,along with a 12%increase in computational resource utilization.展开更多
Deep convolutional neural networks(CNNs)have demonstrated remarkable performance in video super-resolution(VSR).However,the ability of most existing methods to recover fine details in complex scenes is often hindered ...Deep convolutional neural networks(CNNs)have demonstrated remarkable performance in video super-resolution(VSR).However,the ability of most existing methods to recover fine details in complex scenes is often hindered by the loss of shallow texture information during feature extraction.To address this limitation,we propose a 3D Convolutional Enhanced Residual Video Super-Resolution Network(3D-ERVSNet).This network employs a forward and backward bidirectional propagation module(FBBPM)that aligns features across frames using explicit optical flow through lightweight SPyNet.By incorporating an enhanced residual structure(ERS)with skip connections,shallow and deep features are effectively integrated,enhancing texture restoration capabilities.Furthermore,3D convolution module(3DCM)is applied after the backward propagation module to implicitly capture spatio-temporal dependencies.The architecture synergizes these components where FBBPM extracts aligned features,ERS fuses hierarchical representations,and 3DCM refines temporal coherence.Finally,a deep feature aggregation module(DFAM)fuses the processed features,and a pixel-upsampling module(PUM)reconstructs the high-resolution(HR)video frames.Comprehensive evaluations on REDS,Vid4,UDM10,and Vim4 benchmarks demonstrate well performance including 30.95 dB PSNR/0.8822 SSIM on REDS and 32.78 dB/0.8987 on Vim4.3D-ERVSNet achieves significant gains over baselines while maintaining high efficiency with only 6.3M parameters and 77ms/frame runtime(i.e.,20×faster than RBPN).The network’s effectiveness stems from its task-specific asymmetric design that balances explicit alignment and implicit fusion.展开更多
The brain exhibits complex physiology characterized by unique features such as a brain-specific extracellular matrix, compartmentalized structure (white and grey matter), and an aligned axonal network. These physiolog...The brain exhibits complex physiology characterized by unique features such as a brain-specific extracellular matrix, compartmentalized structure (white and grey matter), and an aligned axonal network. These physiological characteristics underpin brain function and facilitate signal transduction similar to that in an electrical circuit. Therefore, investigating these features in vitro is crucial for understanding the interactions between neuronal signal transduction processes and the pathology of neurological diseases. Compared to neurons on patterned substrates, three-dimensional (3D) bioprinting-based neural models provide significant advantages in replicating axonal kinetics without physical limitations. This study proposes the development of a 3D bioprinted engineered neural network (BENN) model to replicate the physiological features of the brain, suggesting its application as a tool for studying neurodegenerative diseases. We employed 3D bioprinting to reconstruct the compartmentalized structure of the brain, and controlled the directionality of axonal growth by applying electrical stimuli to the printed neural structure for overcoming spatial constraints. The reconstructed axonal network demonstrated reliability as a neural analog, including the visualization of mature neuronal features and spontaneous calcium reactions. Furthermore, these brain-like neural network models have demonstrated usefulness for studying neurodegeneration by enabling the visualization of degenerative pathophysiology in alcohol-exposed neurons. The BENN facilitates the visualization of region-specific pathological markers in soma or axon populations, including amyloid-beta formation and axonal deformation. Overall, the BENN closely mimics brain physiology, offers insights into the dynamics of axonal networks, and can be applied to studying neurological diseases.展开更多
3D medical image reconstruction has significantly enhanced diagnostic accuracy,yet the reliance on densely sampled projection data remains a major limitation in clinical practice.Sparse-angle X-ray imaging,though safe...3D medical image reconstruction has significantly enhanced diagnostic accuracy,yet the reliance on densely sampled projection data remains a major limitation in clinical practice.Sparse-angle X-ray imaging,though safer and faster,poses challenges for accurate volumetric reconstruction due to limited spatial information.This study proposes a 3D reconstruction neural network based on adaptive weight fusion(AdapFusionNet)to achieve high-quality 3D medical image reconstruction from sparse-angle X-ray images.To address the issue of spatial inconsistency in multi-angle image reconstruction,an innovative adaptive fusion module was designed to score initial reconstruction results during the inference stage and perform weighted fusion,thereby improving the final reconstruction quality.The reconstruction network is built on an autoencoder(AE)framework and uses orthogonal-angle X-ray images(frontal and lateral projections)as inputs.The encoder extracts 2D features,which the decoder maps into 3D space.This study utilizes a lung CT dataset to obtain complete three-dimensional volumetric data,from which digitally reconstructed radiographs(DRR)are generated at various angles to simulate X-ray images.Since real-world clinical X-ray images rarely come with perfectly corresponding 3D“ground truth,”using CT scans as the three-dimensional reference effectively supports the training and evaluation of deep networks for sparse-angle X-ray 3D reconstruction.Experiments conducted on the LIDC-IDRI dataset with simulated X-ray images(DRR images)as training data demonstrate the superior performance of AdapFusionNet compared to other fusion methods.Quantitative results show that AdapFusionNet achieves SSIM,PSNR,and MAE values of 0.332,13.404,and 0.163,respectively,outperforming other methods(SingleViewNet:0.289,12.363,0.182;AvgFusionNet:0.306,13.384,0.159).Qualitative analysis further confirms that AdapFusionNet significantly enhances the reconstruction of lung and chest contours while effectively reducing noise during the reconstruction process.The findings demonstrate that AdapFusionNet offers significant advantages in 3D reconstruction of sparse-angle X-ray images.展开更多
Variable material screw-based material extrusion(S-MEX)3D printing technology provides a novel approach for fabricating composites with continuous material gradients.Nevertheless,achieving precise alignment between th...Variable material screw-based material extrusion(S-MEX)3D printing technology provides a novel approach for fabricating composites with continuous material gradients.Nevertheless,achieving precise alignment between the process parameters and material compositions is challenging because of fluctuations in the melt rheological state caused by material variations.In this study,an invertible extrusion prediction model for 0-40 wt% short carbon fiber reinforced polyether-ether-ketone(SCF/PEEK)in the S-MEX process was established using an invertible neural network(INN)that demonstrated the capabilities of forward flow rate prediction and inverse process optimization with accuracies of 0.852 and 0.877,respectively.Moreover,a strategy for adjusting the screw speeds using process parameters obtained from the INN was developed to maintain a consistent flow rate during the variable material printing process.Benefiting from uniform flow,the linewidth accuracy was improved by 77%,and the surface roughness was reduced by 51%.Adjusting the process parameters by using an INN offers significant potential for flow rate control and the enhancement of the overall performance of variable material 3D printing.展开更多
Palmprint recognition and palm vein recognition are two emerging biometrics technologies.In the past two decades,many traditional methods have been proposed for palmprint recognition and palm vein recognition,and have...Palmprint recognition and palm vein recognition are two emerging biometrics technologies.In the past two decades,many traditional methods have been proposed for palmprint recognition and palm vein recognition,and have achieved impressive results.However,the research on deep learningbased palmprint recognition and palm vein recognition is still very preliminary.In this paper,in order to investigate the problem of deep learning based 2D and 3D palmprint recognition and palm vein recognition indepth,we conduct performance evaluation of seventeen representative and classic convolutional neural networks(CNNs)on one 3D palmprint database,five 2D palmprint databases and two palm vein databases.A lot of experiments have been carried out in the conditions of different network structures,different learning rates,and different numbers of network layers.We have also conducted experiments on both separate data mode and mixed data mode.Experimental results show that these classic CNNs can achieve promising recognition results,and the recognition performance of recently proposed CNNs is better.Particularly,among classic CNNs,one of the recently proposed classic CNNs,i.e.,EfficientNet achieves the best recognition accuracy.However,the recognition performance of classic CNNs is still slightly worse than that of some traditional recognition methods.展开更多
Lung cancer is the leading cause of cancer deaths worldwide. Accurate early diagnosis is critical in increasing the 5-year survival rate of lung cancer, so the efficient and accurate detection of lung nodules,the pote...Lung cancer is the leading cause of cancer deaths worldwide. Accurate early diagnosis is critical in increasing the 5-year survival rate of lung cancer, so the efficient and accurate detection of lung nodules,the potential precursors to lung cancer, is paramount. In this paper, a computer-aided lung nodule detection system using 3D deep convolutional neural networks(CNNs) is developed. The first multi-scale 11-layer 3D fully convolutional neural network(FCN) is used for screening all lung nodule candidates. Considering relative small sizes of lung nodules and limited memory, the input of the FCN consists of 3D image patches rather than of whole images. The candidates are further classified in the second CNN to get the final result. The proposed method achieves high performance in the LUNA16 challenge and demonstrates the effectiveness of using 3D deep CNNs for lung nodule detection.展开更多
To overcome the computational burden of processing three-dimensional(3 D)medical scans and the lack of spatial information in two-dimensional(2 D)medical scans,a novel segmentation method was proposed that integrates ...To overcome the computational burden of processing three-dimensional(3 D)medical scans and the lack of spatial information in two-dimensional(2 D)medical scans,a novel segmentation method was proposed that integrates the segmentation results of three densely connected 2 D convolutional neural networks(2 D-CNNs).In order to combine the lowlevel features and high-level features,we added densely connected blocks in the network structure design so that the low-level features will not be missed as the network layer increases during the learning process.Further,in order to resolve the problems of the blurred boundary of the glioma edema area,we superimposed and fused the T2-weighted fluid-attenuated inversion recovery(FLAIR)modal image and the T2-weighted(T2)modal image to enhance the edema section.For the loss function of network training,we improved the cross-entropy loss function to effectively avoid network over-fitting.On the Multimodal Brain Tumor Image Segmentation Challenge(BraTS)datasets,our method achieves dice similarity coefficient values of 0.84,0.82,and 0.83 on the BraTS2018 training;0.82,0.85,and 0.83 on the BraTS2018 validation;and 0.81,0.78,and 0.83 on the BraTS2013 testing in terms of whole tumors,tumor cores,and enhancing cores,respectively.Experimental results showed that the proposed method achieved promising accuracy and fast processing,demonstrating good potential for clinical medicine.展开更多
The recent surge of mobile subscribers and user data traffic has accelerated the telecommunication sector towards the adoption of the fifth-generation (5G) mobile networks. Cloud radio access network (CRAN) is a promi...The recent surge of mobile subscribers and user data traffic has accelerated the telecommunication sector towards the adoption of the fifth-generation (5G) mobile networks. Cloud radio access network (CRAN) is a prominent framework in the 5G mobile network to meet the above requirements by deploying low-cost and intelligent multiple distributed antennas known as remote radio heads (RRHs). However, achieving the optimal resource allocation (RA) in CRAN using the traditional approach is still challenging due to the complex structure. In this paper, we introduce the convolutional neural network-based deep Q-network (CNN-DQN) to balance the energy consumption and guarantee the user quality of service (QoS) demand in downlink CRAN. We first formulate the Markov decision process (MDP) for energy efficiency (EE) and build up a 3-layer CNN to capture the environment feature as an input state space. We then use DQN to turn on/off the RRHs dynamically based on the user QoS demand and energy consumption in the CRAN. Finally, we solve the RA problem based on the user constraint and transmit power to guarantee the user QoS demand and maximize the EE with a minimum number of active RRHs. In the end, we conduct the simulation to compare our proposed scheme with nature DQN and the traditional approach.展开更多
In this paper, the complete process of constructing 3D digital core by fullconvolutional neural network is described carefully. A large number of sandstone computedtomography (CT) images are used as training input for...In this paper, the complete process of constructing 3D digital core by fullconvolutional neural network is described carefully. A large number of sandstone computedtomography (CT) images are used as training input for a fully convolutional neural networkmodel. This model is used to reconstruct the three-dimensional (3D) digital core of Bereasandstone based on a small number of CT images. The Hamming distance together with theMinkowski functions for porosity, average volume specifi c surface area, average curvature,and connectivity of both the real core and the digital reconstruction are used to evaluate theaccuracy of the proposed method. The results show that the reconstruction achieved relativeerrors of 6.26%, 1.40%, 6.06%, and 4.91% for the four Minkowski functions and a Hammingdistance of 0.04479. This demonstrates that the proposed method can not only reconstructthe physical properties of real sandstone but can also restore the real characteristics of poredistribution in sandstone, is the ability to which is a new way to characterize the internalmicrostructure of rocks.展开更多
This study introduces an innovative“Big Model”strategy to enhance Bridge Structural Health Monitoring(SHM)using a Convolutional Neural Network(CNN),time-frequency analysis,and fine element analysis.Leveraging ensemb...This study introduces an innovative“Big Model”strategy to enhance Bridge Structural Health Monitoring(SHM)using a Convolutional Neural Network(CNN),time-frequency analysis,and fine element analysis.Leveraging ensemble methods,collaborative learning,and distributed computing,the approach effectively manages the complexity and scale of large-scale bridge data.The CNN employs transfer learning,fine-tuning,and continuous monitoring to optimize models for adaptive and accurate structural health assessments,focusing on extracting meaningful features through time-frequency analysis.By integrating Finite Element Analysis,time-frequency analysis,and CNNs,the strategy provides a comprehensive understanding of bridge health.Utilizing diverse sensor data,sophisticated feature extraction,and advanced CNN architecture,the model is optimized through rigorous preprocessing and hyperparameter tuning.This approach significantly enhances the ability to make accurate predictions,monitor structural health,and support proactive maintenance practices,thereby ensuring the safety and longevity of critical infrastructure.展开更多
基金supported by the Science and Technology Innovation Key R&D Program of Chongqing(CSTB2025TIAD-STX0032)National Key Research and Development Program of China(2024YFF0908200)+1 种基金the Chongqing Technology Innovation and Application Development Special Key Project(CSTB2024TIAD-KPX0018)the Southwest University Graduate Student Research Innovation(SWUB24051)。
文摘Dear Editor,The letter proposes a tensor low-rank orthogonal compression(TLOC)model for a convolutional neural network(CNN),which facilitates its efficient and highly-accurate low-rank representation.Model compression is crucial for deploying deep neural network(DNN)models on resource-constrained embedded devices.
基金supported by the Ministry of Education(MOE)Singapore,Academic Research Fund(AcRF)Tier 1(RG65/22)。
文摘Convolutional neural networks(CNNs)have shown remarkable success across numerous tasks such as image classification,yet the theoretical understanding of their convergence remains underdeveloped compared to their empirical achievements.In this paper,the first filter learning framework with convergence-guaranteed learning laws for end-to-end learning of deep CNNs is proposed.Novel update laws with convergence analysis are formulated based on the mathematical representation of each layer in convolutional neural networks.The proposed learning laws enable concurrent updates of weights across all layers of the deep convolutional neural network and the analysis shows that the training errors converge to certain bounds which are dependent on the approximation errors.Case studies are conducted on benchmark datasets and the results show that the proposed concurrent filter learning framework guarantees the convergence and offers more consistent and reliable results during training with a trade-off in performance compared to stochastic gradient descent methods.This framework represents a significant step towards enhancing the reliability and effectiveness of deep convolutional neural network by developing a theoretical analysis which allows practical implementation of the learning laws with automatic tuning of the learning rate to guarantee the convergence during training.
基金funding support from the National Natural Science Foundation of China(Grant Nos.U22A20594,52079045)Hong-Zhi Cui acknowledges the financial support of the China Scholarship Council(Grant No.CSC:202206710014)for his research at Universitat Politecnica de Catalunya,Barcelona.
文摘Landslide susceptibility mapping(LSM)plays a crucial role in assessing geological risks.The current LSM techniques face a significant challenge in achieving accurate results due to uncertainties associated with regional-scale geotechnical parameters.To explore rainfall-induced LSM,this study proposes a hybrid model that combines the physically-based probabilistic model(PPM)with convolutional neural network(CNN).The PPM is capable of effectively capturing the spatial distribution of landslides by incorporating the probability of failure(POF)considering the slope stability mechanism under rainfall conditions.This significantly characterizes the variation of POF caused by parameter uncertainties.CNN was used as a binary classifier to capture the spatial and channel correlation between landslide conditioning factors and the probability of landslide occurrence.OpenCV image enhancement technique was utilized to extract non-landslide points based on the POF of landslides.The proposed model comprehensively considers physical mechanics when selecting non-landslide samples,effectively filtering out samples that do not adhere to physical principles and reduce the risk of overfitting.The results indicate that the proposed PPM-CNN hybrid model presents a higher prediction accuracy,with an area under the curve(AUC)value of 0.85 based on the landslide case of the Niangniangba area of Gansu Province,China compared with the individual CNN model(AUC=0.61)and the PPM(AUC=0.74).This model can also consider the statistical correlation and non-normal probability distributions of model parameters.These results offer practical guidance for future research on rainfall-induced LSM at the regional scale.
文摘The integration of image analysis through deep learning(DL)into rock classification represents a significant leap forward in geological research.While traditional methods remain invaluable for their expertise and historical context,DL offers a powerful complement by enhancing the speed,objectivity,and precision of the classification process.This research explores the significance of image data augmentation techniques in optimizing the performance of convolutional neural networks(CNNs)for geological image analysis,particularly in the classification of igneous,metamorphic,and sedimentary rock types from rock thin section(RTS)images.This study primarily focuses on classic image augmentation techniques and evaluates their impact on model accuracy and precision.Results demonstrate that augmentation techniques like Equalize significantly enhance the model's classification capabilities,achieving an F1-Score of 0.9869 for igneous rocks,0.9884 for metamorphic rocks,and 0.9929 for sedimentary rocks,representing improvements compared to the baseline original results.Moreover,the weighted average F1-Score across all classes and techniques is 0.9886,indicating an enhancement.Conversely,methods like Distort lead to decreased accuracy and F1-Score,with an F1-Score of 0.949 for igneous rocks,0.954 for metamorphic rocks,and 0.9416 for sedimentary rocks,exacerbating the performance compared to the baseline.The study underscores the practicality of image data augmentation in geological image classification and advocates for the adoption of DL methods in this domain for automation and improved results.The findings of this study can benefit various fields,including remote sensing,mineral exploration,and environmental monitoring,by enhancing the accuracy of geological image analysis both for scientific research and industrial applications.
基金supported by Interdisciplinary Innova-tion Project of“Bioarchaeology Laboratory”of Jilin University,China,and“MedicineþX”Interdisciplinary Innovation Team of Norman Bethune Health Science Center of Jilin University,China(Grant No.:2022JBGS05).
文摘Severe acute respiratory syndrome coronavirus 2(SARS-CoV-2)mutations are influenced by random and uncontrollable factors,and the risk of the next widespread epidemic remains.Dual-target drugs that synergistically act on two targets exhibit strong therapeutic effects and advantages against mutations.In this study,a novel computational workflow was developed to design dual-target SARS-CoV-2 candidate inhibitors with the Envelope protein and Main protease selected as the two target proteins.The drug-like molecules of our self-constructed 3D scaffold database were used as high-throughput molecular docking probes for feature extraction of two target protein pockets.A multi-layer perceptron(MLP)was employed to embed the binding affinities into a latent space as conditional vectors to control conditional distribution.Utilizing a conditional generative neural network,cG-SchNet,with 3D Euclidean group(E3)symmetries,the conditional probability distributions of molecular 3D structures were acquired and a set of novel SARS-CoV-2 dual-target candidate inhibitors were generated.The 1D probability,2D joint probability,and 2D cumulative probability distribution results indicate that the generated sets are significantly enhanced compared to the training set in the high binding affinity area.Among the 201 generated molecules,42 molecules exhibited a sum binding affinity exceeding 17.0 kcal/mol while 9 of them having a sum binding affinity exceeding 19.0 kcal/mol,demonstrating structure diversity along with strong dual-target affinities,good absorption,distribution,metabolism,excretion,and toxicity(ADMET)properties,and ease of synthesis.Dual-target drugs are rare and difficult to find,and our“high-throughput docking-multi-conditional generation”workflow offers a wide range of options for designing or optimizing potent dual-target SARS-CoV-2 inhibitors.
文摘The migration of tasks aided by machine learning(ML)predictions IN(DPM)is a system-level design technique that is used to reduce energy by enhancing the overall performance of the processor.In this paper,we address the issue of system-level higher task dissipation during the execution of parallel workloads with common deadlines by introducing a machine learning-based framework that includes task migration using energy-efficient earliest deadline first scheduling(EA-EDF).ML-based EA-EDF enhances the overall throughput and optimizes the energy to avoid delay and performance degradation in a multiprocessor system.The proposed system model allocates processors to the ready task set in such a way that their deadlines are guaranteed.A full task migration policy is also integrated to ensure proper task mapping that ensures inter-process linkage among the arrived tasks with the same deadlines.The execution of a task can halt on one CPU and reschedule the execution on a different processor to avoid delay and ensure meeting the deadline.Our approach shows promising potential for machine-learning-based schedulability analysis enables a comparison between different ML models and shows a promising reduction in energy as compared with other ML-aware task migration techniques for SoC like Multi-Layer Feed-Forward Neural Networks(MLFNN)based on convolutional neural network(CNN),Random Forest(RF)and Deep learning(DL)algorithm.The Simulations are conducted using super pipelined microarchitecture of advanced micro devices(AMD)XScale PXA270 using instruction and data cache per core 32 Kbyte I-cache and 32 Kbyte D-cache on various utilization factors(u_(i))12%,31%and 50%.The proposed approach consumes 5.3%less energy when almost half of the CPU is running and on a lower workload consumes 1.04%less energy.The proposed design accumulatively gives significant improvements by reducing the energy dissipation on three clock rates by 4.41%,on 624 MHz by 5.4%and 5.9%on applications operating on 416 and 312 MHz standard operating frequencies.
基金funded by the Deanship of Graduate Studies and Scientific Research at Qassim University for financial support(QU-APC-2025).
文摘Hemoglobin is a vital protein in red blood cells responsible for transporting oxygen throughout the body.Its accurate measurement is crucial for diagnosing and managing conditions such as anemia and diabetes,where abnormal hemoglobin levels can indicate significant health issues.Traditional methods for hemoglobin measurement are invasive,causing pain,risk of infection,and are less convenient for frequent monitoring.PPG is a transformative technology in wearable healthcare for noninvasive monitoring and widely explored for blood pressure,sleep,blood glucose,and stress analysis.In this work,we propose a hemoglobin estimation method using an adaptive lightweight convolutional neural network(HMALCNN)from PPG.The HMALCNN is designed to capture both fine-grained local waveform characteristics and global contextual patterns,ensuring robust performance across acquisition settings.We validated our approach on two multi-regional datasets containing 152 and 68 subjects,respectively,employing a subjectindependent 5-fold cross-validation strategy.The proposed method achieved root mean square errors(RMSE)of 0.90 and 1.20 g/dL for the two datasets,with strong Pearson correlations of 0.82 and 0.72.We conducted extensive posthoc analyses to assess clinical utility and interpretability.A±1 g/dL clinical error tolerance evaluation revealed that 91.3%and 86.7%of predictions for the two datasets fell within the acceptable clinical range.Hemoglobin range-wise analysis demonstrated consistently high accuracy in the normal and low hemoglobin categories.Statistical significance testing using the Wilcoxon signed-rank test confirmed the stability of performance across validation folds(p>0.05 for both RMSE and correlation).Furthermore,model interpretability was enhanced using Gradient-weighted Class Activation Mapping(Grad-CAM),supporting the model’s clinical trustworthiness.The proposed HMALCNN offers a computationally efficient,clinically interpretable,and generalizable framework for noninvasive hemoglobin monitoring,with strong potential for integration into wearable healthcare systems as a practical alternative to invasive measurement techniques.
基金Supported by the National Science and Technology Major Project of China(No.2022ZD0119003)the National Natural Science Foundation of China(No.61834005).
文摘To enhance the inference efficiency of convolutional neural network(CNN),tensor parallelism is employed to improve the parallelism within operators.However,existing methods are customized to specific networks and hardware,limiting their generalizability.This paper proposes an approach called resource-adaptive tensor decomposition(RATD)for CNN operators,which aims to achieve an optimal match between computational resources and parallel computing tasks.Firstly,CNN is represented with fine-grained tensors at the lower graph level,thereby decoupling tensors that can be computed in parallel within operators.Secondly,the convolution and pooling operators are fused,and the decoupled tensor blocks are scheduled in parallel.Finally,a cost model is constructed,based on runtime and resource utilization,to iteratively refine the process of tensor block decomposition and automatically determine the optimal tensor decomposition.Experimental results demonstrate that the proposed RATD improves the accuracy of the model by 11%.Compared with CUDA(compute unified device architecture)deep neural network library(cuDNN),RATD achieves an average speedup ratio of 1.21 times in inference time across various convolution kernels,along with a 12%increase in computational resource utilization.
基金supported in part by the Basic and Applied Basic Research Foundation of Guangdong Province[2025A1515011566]in part by the State Key Laboratory for Novel Software Technology,Nanjing University[KFKT2024B08]+1 种基金in part by Leading Talents in Gusu Innovation and Entrepreneurship[ZXL2023170]in part by the Basic Research Programs of Taicang 2024,[TC2024JC32].
文摘Deep convolutional neural networks(CNNs)have demonstrated remarkable performance in video super-resolution(VSR).However,the ability of most existing methods to recover fine details in complex scenes is often hindered by the loss of shallow texture information during feature extraction.To address this limitation,we propose a 3D Convolutional Enhanced Residual Video Super-Resolution Network(3D-ERVSNet).This network employs a forward and backward bidirectional propagation module(FBBPM)that aligns features across frames using explicit optical flow through lightweight SPyNet.By incorporating an enhanced residual structure(ERS)with skip connections,shallow and deep features are effectively integrated,enhancing texture restoration capabilities.Furthermore,3D convolution module(3DCM)is applied after the backward propagation module to implicitly capture spatio-temporal dependencies.The architecture synergizes these components where FBBPM extracts aligned features,ERS fuses hierarchical representations,and 3DCM refines temporal coherence.Finally,a deep feature aggregation module(DFAM)fuses the processed features,and a pixel-upsampling module(PUM)reconstructs the high-resolution(HR)video frames.Comprehensive evaluations on REDS,Vid4,UDM10,and Vim4 benchmarks demonstrate well performance including 30.95 dB PSNR/0.8822 SSIM on REDS and 32.78 dB/0.8987 on Vim4.3D-ERVSNet achieves significant gains over baselines while maintaining high efficiency with only 6.3M parameters and 77ms/frame runtime(i.e.,20×faster than RBPN).The network’s effectiveness stems from its task-specific asymmetric design that balances explicit alignment and implicit fusion.
基金supported by Korean Fund for Regenerative Medicine funded by Ministry of Science and ICT,and Ministry of Health and Welfare(22A0106L1,Republic of Korea)the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT)(No.2022M3C1A3081359).
文摘The brain exhibits complex physiology characterized by unique features such as a brain-specific extracellular matrix, compartmentalized structure (white and grey matter), and an aligned axonal network. These physiological characteristics underpin brain function and facilitate signal transduction similar to that in an electrical circuit. Therefore, investigating these features in vitro is crucial for understanding the interactions between neuronal signal transduction processes and the pathology of neurological diseases. Compared to neurons on patterned substrates, three-dimensional (3D) bioprinting-based neural models provide significant advantages in replicating axonal kinetics without physical limitations. This study proposes the development of a 3D bioprinted engineered neural network (BENN) model to replicate the physiological features of the brain, suggesting its application as a tool for studying neurodegenerative diseases. We employed 3D bioprinting to reconstruct the compartmentalized structure of the brain, and controlled the directionality of axonal growth by applying electrical stimuli to the printed neural structure for overcoming spatial constraints. The reconstructed axonal network demonstrated reliability as a neural analog, including the visualization of mature neuronal features and spontaneous calcium reactions. Furthermore, these brain-like neural network models have demonstrated usefulness for studying neurodegeneration by enabling the visualization of degenerative pathophysiology in alcohol-exposed neurons. The BENN facilitates the visualization of region-specific pathological markers in soma or axon populations, including amyloid-beta formation and axonal deformation. Overall, the BENN closely mimics brain physiology, offers insights into the dynamics of axonal networks, and can be applied to studying neurological diseases.
基金Supported by Sichuan Science and Technology Program(2023YFSY0026,2023YFH0004).
文摘3D medical image reconstruction has significantly enhanced diagnostic accuracy,yet the reliance on densely sampled projection data remains a major limitation in clinical practice.Sparse-angle X-ray imaging,though safer and faster,poses challenges for accurate volumetric reconstruction due to limited spatial information.This study proposes a 3D reconstruction neural network based on adaptive weight fusion(AdapFusionNet)to achieve high-quality 3D medical image reconstruction from sparse-angle X-ray images.To address the issue of spatial inconsistency in multi-angle image reconstruction,an innovative adaptive fusion module was designed to score initial reconstruction results during the inference stage and perform weighted fusion,thereby improving the final reconstruction quality.The reconstruction network is built on an autoencoder(AE)framework and uses orthogonal-angle X-ray images(frontal and lateral projections)as inputs.The encoder extracts 2D features,which the decoder maps into 3D space.This study utilizes a lung CT dataset to obtain complete three-dimensional volumetric data,from which digitally reconstructed radiographs(DRR)are generated at various angles to simulate X-ray images.Since real-world clinical X-ray images rarely come with perfectly corresponding 3D“ground truth,”using CT scans as the three-dimensional reference effectively supports the training and evaluation of deep networks for sparse-angle X-ray 3D reconstruction.Experiments conducted on the LIDC-IDRI dataset with simulated X-ray images(DRR images)as training data demonstrate the superior performance of AdapFusionNet compared to other fusion methods.Quantitative results show that AdapFusionNet achieves SSIM,PSNR,and MAE values of 0.332,13.404,and 0.163,respectively,outperforming other methods(SingleViewNet:0.289,12.363,0.182;AvgFusionNet:0.306,13.384,0.159).Qualitative analysis further confirms that AdapFusionNet significantly enhances the reconstruction of lung and chest contours while effectively reducing noise during the reconstruction process.The findings demonstrate that AdapFusionNet offers significant advantages in 3D reconstruction of sparse-angle X-ray images.
基金supported by National Natural Science Foundation of China(Grant Nos.12202547,62461160259)Shaanxi Province Qingchuangyuan“Scientist and Engineering”Team Construction Project(Grant Nos.2022KXJ-102,2022KXJ-106)+1 种基金Fundamental Research Funds for the Central UniversitiesProgram for Innovation Team of Shaanxi Province(Grant No.2023-CX-TD-17).
文摘Variable material screw-based material extrusion(S-MEX)3D printing technology provides a novel approach for fabricating composites with continuous material gradients.Nevertheless,achieving precise alignment between the process parameters and material compositions is challenging because of fluctuations in the melt rheological state caused by material variations.In this study,an invertible extrusion prediction model for 0-40 wt% short carbon fiber reinforced polyether-ether-ketone(SCF/PEEK)in the S-MEX process was established using an invertible neural network(INN)that demonstrated the capabilities of forward flow rate prediction and inverse process optimization with accuracies of 0.852 and 0.877,respectively.Moreover,a strategy for adjusting the screw speeds using process parameters obtained from the INN was developed to maintain a consistent flow rate during the variable material printing process.Benefiting from uniform flow,the linewidth accuracy was improved by 77%,and the surface roughness was reduced by 51%.Adjusting the process parameters by using an INN offers significant potential for flow rate control and the enhancement of the overall performance of variable material 3D printing.
基金National Science Foundation of China(Nos.61673157,62076086,61972129 and 61702154)Key Research and Development Program in Anhui Province(Nos.202004d07020008 and 201904d07020010).
文摘Palmprint recognition and palm vein recognition are two emerging biometrics technologies.In the past two decades,many traditional methods have been proposed for palmprint recognition and palm vein recognition,and have achieved impressive results.However,the research on deep learningbased palmprint recognition and palm vein recognition is still very preliminary.In this paper,in order to investigate the problem of deep learning based 2D and 3D palmprint recognition and palm vein recognition indepth,we conduct performance evaluation of seventeen representative and classic convolutional neural networks(CNNs)on one 3D palmprint database,five 2D palmprint databases and two palm vein databases.A lot of experiments have been carried out in the conditions of different network structures,different learning rates,and different numbers of network layers.We have also conducted experiments on both separate data mode and mixed data mode.Experimental results show that these classic CNNs can achieve promising recognition results,and the recognition performance of recently proposed CNNs is better.Particularly,among classic CNNs,one of the recently proposed classic CNNs,i.e.,EfficientNet achieves the best recognition accuracy.However,the recognition performance of classic CNNs is still slightly worse than that of some traditional recognition methods.
基金the National Natural Science Foundation of China(No.81371624)the National Key Research and Development Program of China(No.2016YFC0104608)+1 种基金the National Basic Research Program of China(No.2010CB834302)the Shanghai Jiao Tong University Medical Engineering Cross Research Funds(Nos.YG2013MS30 and YG2014ZD05)
文摘Lung cancer is the leading cause of cancer deaths worldwide. Accurate early diagnosis is critical in increasing the 5-year survival rate of lung cancer, so the efficient and accurate detection of lung nodules,the potential precursors to lung cancer, is paramount. In this paper, a computer-aided lung nodule detection system using 3D deep convolutional neural networks(CNNs) is developed. The first multi-scale 11-layer 3D fully convolutional neural network(FCN) is used for screening all lung nodule candidates. Considering relative small sizes of lung nodules and limited memory, the input of the FCN consists of 3D image patches rather than of whole images. The candidates are further classified in the second CNN to get the final result. The proposed method achieves high performance in the LUNA16 challenge and demonstrates the effectiveness of using 3D deep CNNs for lung nodule detection.
基金the National Natural Science Foundation of China(No.81830052)the Shanghai Natural Science Foundation of China(No.20ZR1438300)the Shanghai Science and Technology Support Project(No.18441900500),China。
文摘To overcome the computational burden of processing three-dimensional(3 D)medical scans and the lack of spatial information in two-dimensional(2 D)medical scans,a novel segmentation method was proposed that integrates the segmentation results of three densely connected 2 D convolutional neural networks(2 D-CNNs).In order to combine the lowlevel features and high-level features,we added densely connected blocks in the network structure design so that the low-level features will not be missed as the network layer increases during the learning process.Further,in order to resolve the problems of the blurred boundary of the glioma edema area,we superimposed and fused the T2-weighted fluid-attenuated inversion recovery(FLAIR)modal image and the T2-weighted(T2)modal image to enhance the edema section.For the loss function of network training,we improved the cross-entropy loss function to effectively avoid network over-fitting.On the Multimodal Brain Tumor Image Segmentation Challenge(BraTS)datasets,our method achieves dice similarity coefficient values of 0.84,0.82,and 0.83 on the BraTS2018 training;0.82,0.85,and 0.83 on the BraTS2018 validation;and 0.81,0.78,and 0.83 on the BraTS2013 testing in terms of whole tumors,tumor cores,and enhancing cores,respectively.Experimental results showed that the proposed method achieved promising accuracy and fast processing,demonstrating good potential for clinical medicine.
基金supported by the Universiti Tunku Abdul Rahman (UTAR) Malaysia under UTARRF (IPSR/RMC/UTARRF/2021-C1/T05)
文摘The recent surge of mobile subscribers and user data traffic has accelerated the telecommunication sector towards the adoption of the fifth-generation (5G) mobile networks. Cloud radio access network (CRAN) is a prominent framework in the 5G mobile network to meet the above requirements by deploying low-cost and intelligent multiple distributed antennas known as remote radio heads (RRHs). However, achieving the optimal resource allocation (RA) in CRAN using the traditional approach is still challenging due to the complex structure. In this paper, we introduce the convolutional neural network-based deep Q-network (CNN-DQN) to balance the energy consumption and guarantee the user quality of service (QoS) demand in downlink CRAN. We first formulate the Markov decision process (MDP) for energy efficiency (EE) and build up a 3-layer CNN to capture the environment feature as an input state space. We then use DQN to turn on/off the RRHs dynamically based on the user QoS demand and energy consumption in the CRAN. Finally, we solve the RA problem based on the user constraint and transmit power to guarantee the user QoS demand and maximize the EE with a minimum number of active RRHs. In the end, we conduct the simulation to compare our proposed scheme with nature DQN and the traditional approach.
基金the National Natural Science Foundation of China(No.41274129)Chuan Qing Drilling Engineering Company's Scientific Research Project:Seismic detection technology and application of complex carbonate reservoir in Sulige Majiagou Formation and the 2018 Central Supporting Local Co-construction Fund(No.80000-18Z0140504)the Construction and Development of Universities in 2019-Joint Support for Geophysics(Double First-Class center,80000-19Z0204)。
文摘In this paper, the complete process of constructing 3D digital core by fullconvolutional neural network is described carefully. A large number of sandstone computedtomography (CT) images are used as training input for a fully convolutional neural networkmodel. This model is used to reconstruct the three-dimensional (3D) digital core of Bereasandstone based on a small number of CT images. The Hamming distance together with theMinkowski functions for porosity, average volume specifi c surface area, average curvature,and connectivity of both the real core and the digital reconstruction are used to evaluate theaccuracy of the proposed method. The results show that the reconstruction achieved relativeerrors of 6.26%, 1.40%, 6.06%, and 4.91% for the four Minkowski functions and a Hammingdistance of 0.04479. This demonstrates that the proposed method can not only reconstructthe physical properties of real sandstone but can also restore the real characteristics of poredistribution in sandstone, is the ability to which is a new way to characterize the internalmicrostructure of rocks.
文摘This study introduces an innovative“Big Model”strategy to enhance Bridge Structural Health Monitoring(SHM)using a Convolutional Neural Network(CNN),time-frequency analysis,and fine element analysis.Leveraging ensemble methods,collaborative learning,and distributed computing,the approach effectively manages the complexity and scale of large-scale bridge data.The CNN employs transfer learning,fine-tuning,and continuous monitoring to optimize models for adaptive and accurate structural health assessments,focusing on extracting meaningful features through time-frequency analysis.By integrating Finite Element Analysis,time-frequency analysis,and CNNs,the strategy provides a comprehensive understanding of bridge health.Utilizing diverse sensor data,sophisticated feature extraction,and advanced CNN architecture,the model is optimized through rigorous preprocessing and hyperparameter tuning.This approach significantly enhances the ability to make accurate predictions,monitor structural health,and support proactive maintenance practices,thereby ensuring the safety and longevity of critical infrastructure.
