Aiming at the problems of low detection accuracy and large model size of existing object detection algorithms applied to complex road scenes,an improved you only look once version 8(YOLOv8)object detection algorithm f...Aiming at the problems of low detection accuracy and large model size of existing object detection algorithms applied to complex road scenes,an improved you only look once version 8(YOLOv8)object detection algorithm for infrared images,F-YOLOv8,is proposed.First,a spatial-to-depth network replaces the traditional backbone network's strided convolution or pooling layer.At the same time,it combines with the channel attention mechanism so that the neural network focuses on the channels with large weight values to better extract low-resolution image feature information;then an improved feature pyramid network of lightweight bidirectional feature pyramid network(L-BiFPN)is proposed,which can efficiently fuse features of different scales.In addition,a loss function of insertion of union based on the minimum point distance(MPDIoU)is introduced for bounding box regression,which obtains faster convergence speed and more accurate regression results.Experimental results on the FLIR dataset show that the improved algorithm can accurately detect infrared road targets in real time with 3%and 2.2%enhancement in mean average precision at 50%IoU(mAP50)and mean average precision at 50%—95%IoU(mAP50-95),respectively,and 38.1%,37.3%and 16.9%reduction in the number of model parameters,the model weight,and floating-point operations per second(FLOPs),respectively.To further demonstrate the detection capability of the improved algorithm,it is tested on the public dataset PASCAL VOC,and the results show that F-YOLO has excellent generalized detection performance.展开更多
This paper proposes a lightweight traffic sign detection system based on you only look once(YOLO).Firstly,the classification to fusion(C2f)structure is integrated into the backbone network,employing deformable convolu...This paper proposes a lightweight traffic sign detection system based on you only look once(YOLO).Firstly,the classification to fusion(C2f)structure is integrated into the backbone network,employing deformable convolution and bi-directional feature pyramid network(BiFPN)_Concat to improve the adaptability of the network.Secondly,the simple attention module(SimAm)is embedded to prioritize key features and reduce the complexity of the model after the C2f layer at the end of the backbone network.Next,the focal efficient intersection over union(EloU)is introduced to adjust the weights of challenging samples.Finally,we accomplish the design and deployment for the mobile app.The results demonstrate improvements,with the F1 score of 0.8987,mean average precision(mAP)@0.5 of 98.8%,mAP@0.5:0.95 of 75.6%,and the detection speed of 50 frames per second(FPS).展开更多
UAV-based object detection is rapidly expanding in both civilian and military applications,including security surveillance,disaster assessment,and border patrol.However,challenges such as small objects,occlusions,comp...UAV-based object detection is rapidly expanding in both civilian and military applications,including security surveillance,disaster assessment,and border patrol.However,challenges such as small objects,occlusions,complex backgrounds,and variable lighting persist due to the unique perspective of UAV imagery.To address these issues,this paper introduces DAFPN-YOLO,an innovative model based on YOLOv8s(You Only Look Once version 8s).Themodel strikes a balance between detection accuracy and speed while reducing parameters,making itwell-suited for multi-object detection tasks from drone perspectives.A key feature of DAFPN-YOLO is the enhanced Drone-AFPN(Adaptive Feature Pyramid Network),which adaptively fuses multi-scale features to optimize feature extraction and enhance spatial and small-object information.To leverage Drone-AFPN’smulti-scale capabilities fully,a dedicated 160×160 small-object detection head was added,significantly boosting detection accuracy for small targets.In the backbone,the C2f_Dual(Cross Stage Partial with Cross-Stage Feature Fusion Dual)module and SPPELAN(Spatial Pyramid Pooling with Enhanced LocalAttentionNetwork)modulewere integrated.These components improve feature extraction and information aggregationwhile reducing parameters and computational complexity,enhancing inference efficiency.Additionally,Shape-IoU(Shape Intersection over Union)is used as the loss function for bounding box regression,enabling more precise shape-based object matching.Experimental results on the VisDrone 2019 dataset demonstrate the effectiveness ofDAFPN-YOLO.Compared to YOLOv8s,the proposedmodel achieves a 5.4 percentage point increase inmAP@0.5,a 3.8 percentage point improvement in mAP@0.5:0.95,and a 17.2%reduction in parameter count.These results highlight DAFPN-YOLO’s advantages in UAV-based object detection,offering valuable insights for applying deep learning to UAV-specific multi-object detection tasks.展开更多
With the rapid urbanization and exponential population growth in China,two-wheeled vehicles have become a popular mode of transportation,particularly for short-distance travel.However,due to a lack of safety awareness...With the rapid urbanization and exponential population growth in China,two-wheeled vehicles have become a popular mode of transportation,particularly for short-distance travel.However,due to a lack of safety awareness,traffic violations by two-wheeled vehicle riders have become a widespread concern,contributing to urban traffic risks.Currently,significant human and material resources are being allocated to monitor and intercept non-compliant riders to ensure safe driving behavior.To enhance the safety,efficiency,and cost-effectiveness of traffic monitoring,automated detection systems based on image processing algorithms can be employed to identify traffic violations from eye-level video footage.In this study,we propose a robust detection algorithm specifically designed for two-wheeled vehicles,which serves as a fundamental step toward intelligent traffic monitoring.Our approach integrates a novel convolutional and attention mechanism to improve detection accuracy and efficiency.Additionally,we introduce a semi-supervised training strategy that leverages a large number of unlabeled images to enhance the model’s learning capability by extracting valuable background information.This method enables the model to generalize effectively to diverse urban environments and varying lighting conditions.We evaluate our proposed algorithm on a custom-built dataset,and experimental results demonstrate its superior performance,achieving an average precision(AP)of 95%and a recall(R)of 90.6%.Furthermore,the model maintains a computational efficiency of only 25.7 GFLOPs while achieving a high processing speed of 249 FPS,making it highly suitable for deployment on edge devices.Compared to existing detection methods,our approach significantly enhances the accuracy and robustness of two-wheeled vehicle identification while ensuring real-time performance.展开更多
Brain tumors are neoplastic diseases caused by the proliferation of abnormal cells in brain tissues,and their appearance may lead to a series of complex symptoms.However,current methods struggle to capture deeper brai...Brain tumors are neoplastic diseases caused by the proliferation of abnormal cells in brain tissues,and their appearance may lead to a series of complex symptoms.However,current methods struggle to capture deeper brain tumor image feature information due to the variations in brain tumor morphology,size,and complex background,resulting in low detection accuracy,high rate of misdiagnosis and underdiagnosis,and challenges in meeting clinical needs.Therefore,this paper proposes the CMS-YOLO network model for multi-category brain tumor detection,which is based on the You Only Look Once version 10(YOLOv10s)algorithm.This model innovatively integrates the Convolutional Medical UNet extended block(CMUNeXt Block)to design a brand-new CSP Bottleneck with 2 convolutions(C2f)structure,which significantly enhances the ability to extract features of the lesion area.Meanwhile,to address the challenge of complex backgrounds in brain tumor detection,a Multi-Scale Attention Aggregation(MSAA)module is introduced.The module integrates features of lesions at different scales,enabling the model to effectively capture multi-scale contextual information and enhance detection accuracy in complex scenarios.Finally,during the model training process,the Shape-IoU loss function is employed to replace the Complete-IoU(CIoU)loss function for optimizing bounding box regression.This ensures that the predicted bounding boxes generated by the model closely match the actual tumor contours,thereby further enhancing the detection precision.The experimental results show that the improved method achieves 94.80%precision,93.60%recall,96.20%score,and 79.60%on the MRI for Brain Tumor with Bounding Boxes dataset.Compared to the YOLOv10s model,this represents improvements of 1.0%,1.1%,1.0%,and 1.1%,respectively.The method can achieve automatic detection and localization of three distinct categories of brain tumors—glioma,meningioma,and pituitary tumor,which can accurately detect and identify brain tumors,assist doctors in early diagnosis,and promote the development of early treatment.展开更多
Pulmonary nodules represent an early manifestation of lung cancer.However,pulmonary nodules only constitute a small portion of the overall image,posing challenges for physicians in image interpretation and potentially...Pulmonary nodules represent an early manifestation of lung cancer.However,pulmonary nodules only constitute a small portion of the overall image,posing challenges for physicians in image interpretation and potentially leading to false positives or missed detections.To solve these problems,the YOLOv8 network is enhanced by adding deformable convolution and atrous spatial pyramid pooling(ASPP),along with the integration of a coordinate attention(CA)mechanism.This allows the network to focus on small targets while expanding the receptive field without losing resolution.At the same time,context information on the target is gathered and feature expression is enhanced by attention modules in different directions.It effectively improves the positioning accuracy and achieves good results on the LUNA16 dataset.Compared with other detection algorithms,it improves the accuracy of pulmonary nodule detection to a certain extent.展开更多
Load time series analysis is critical for resource management and optimization decisions,especially automated analysis techniques.Existing research has insufficiently interpreted the overall characteristics of samples...Load time series analysis is critical for resource management and optimization decisions,especially automated analysis techniques.Existing research has insufficiently interpreted the overall characteristics of samples,leading to significant differences in load level detection conclusions for samples with different characteristics(trend,seasonality,cyclicality).Achieving automated,feature-adaptive,and quantifiable analysis methods remains a challenge.This paper proposes a Threshold Recognition-based Load Level Detection Algorithm(TRLLD),which effectively identifies different load level regions in samples of arbitrary size and distribution type based on sample characteristics.By utilizing distribution density uniformity,the algorithm classifies data points and ultimately obtains normalized load values.In the feature recognition step,the algorithm employs the Density Uniformity Index Based on Differences(DUID),High Load Level Concentration(HLLC),and Low Load Level Concentration(LLLC)to assess sample characteristics,which are independent of specific load values,providing a standardized perspective on features,ensuring high efficiency and strong interpretability.Compared to traditional methods,the proposed approach demonstrates better adaptive and real-time analysis capabilities.Experimental results indicate that it can effectively identify high load and low load regions in 16 groups of time series samples with different load characteristics,yielding highly interpretable results.The correlation between the DUID and sample density distribution uniformity reaches 98.08%.When introducing 10% MAD intensity noise,the maximum relative error is 4.72%,showcasing high robustness.Notably,it exhibits significant advantages in general and low sample scenarios.展开更多
Combining the principle of antibody concentration with the idea of biological evolution, this paper proposes an adaptive target detection algorithm for cloud service security based on Bio-Inspired Performance Evaluati...Combining the principle of antibody concentration with the idea of biological evolution, this paper proposes an adaptive target detection algorithm for cloud service security based on Bio-Inspired Performance Evaluation Process Algebra(Bio-PEPA). The formal modelling of cloud services is formally modded by Bio-PEPA and the modules are transformed between cloud service internal structures and various components. Then, the security adaptive target detection algorithm of cloud service is divided into two processes, the short-term optimal action selection process which selects the current optimal detective action through the iterative operation of the expected function and the adaptive function, and the long-term detective strategy realized through the updates and eliminations of action planning table. The combination of the two processes reflects the self-adaptability of cloud service system to target detection. The simulating test detects three different kinds of security risks and then analyzes the relationship between the numbers of components with time in the service process. The performance of this method is compared with random detection method and three anomaly detection methods by the cloud service detection experiment. The detection time of this method is 50.1% of three kinds of detection methods and 86.3% of the random detection method. The service success rate is about 15% higher than that of random detection methods. The experimental results show that the algorithm has good time performance and high detection hit rate.展开更多
Quality of Service (QoS) generally refers to measurable like latency and throughput, things that directly affect the user experience. Queuing (the most popular QoS tool) involves choosing the packets to be sent based ...Quality of Service (QoS) generally refers to measurable like latency and throughput, things that directly affect the user experience. Queuing (the most popular QoS tool) involves choosing the packets to be sent based on something other than arrival time. The Active queue management is important subject to manage this queue to increase the effectiveness of Transmission Control Protocol networks. Active queue management (AQM) is an effective means to enhance congestion control, and to achieve trade-off between link utilization and delay. The de facto standard, Random Early Detection (RED), and many of its variants employ queue length as a congestion indicator to trigger packet dropping. One of these enhancements of RED is FRED or Fair Random Early Detection attempts to deal with a fundamental aspect of RED in that it imposes the same loss rate on all flows, regardless of their bandwidths. FRED also uses per-flow active accounting, and tracks the state of active flows. FRED protects fragile flows by deterministically accepting flows from low bandwidth connections and fixes several shortcomings of RED by computing queue length during both arrival and departure of the packet. Unlike FRED, we propose a new scheme that used hazard rate estimated packet dropping function in FRED. We call this new scheme Enhancement Fair Random Early Detection. The key idea is that, with EFRED Scheme change packet dropping function, to get packet dropping less than RED and other AQM algorithms like ARED, REM, RED, etc. Simulations demonstrate that EFRED achieves a more stable throughput and performs better than current active queue management algorithms due to decrease the packets loss percentage and lowest in queuing delay, end to end delay and delay variation (JITTER).展开更多
To maintain the reliability of power systems,routine inspections using drones equipped with advanced object detection algorithms are essential for preempting power-related issues.The increasing resolution of drone-cap...To maintain the reliability of power systems,routine inspections using drones equipped with advanced object detection algorithms are essential for preempting power-related issues.The increasing resolution of drone-captured images has posed a challenge for traditional target detection methods,especially in identifying small objects in high-resolution images.This study presents an enhanced object detection algorithm based on the Faster Regionbased Convolutional Neural Network(Faster R-CNN)framework,specifically tailored for detecting small-scale electrical components like insulators,shock hammers,and screws in transmission line.The algorithm features an improved backbone network for Faster R-CNN,which significantly boosts the feature extraction network’s ability to detect fine details.The Region Proposal Network is optimized using a method of guided feature refinement(GFR),which achieves a balance between accuracy and speed.The incorporation of Generalized Intersection over Union(GIOU)and Region of Interest(ROI)Align further refines themodel’s accuracy.Experimental results demonstrate a notable improvement in mean Average Precision,reaching 89.3%,an 11.1%increase compared to the standard Faster R-CNN.This highlights the effectiveness of the proposed algorithm in identifying electrical components in high-resolution aerial images.展开更多
In smart driving for rail transit,a reliable obstacle detection system is an important guarantee for the safety of trains.Therein,the detection of the rail area directly affects the accuracy of the system to identify ...In smart driving for rail transit,a reliable obstacle detection system is an important guarantee for the safety of trains.Therein,the detection of the rail area directly affects the accuracy of the system to identify dangerous targets.Both the rail line and the lane are presented as thin line shapes in the image,but the rail scene is more complex,and the color of the rail line is more difficult to distinguish from the background.By comparison,there are already many deep learning-based lane detection algorithms,but there is a lack of public datasets and targeted deep learning detection algorithms for rail line detection.To address this,this paper constructs a rail image dataset RailwayLine and labels the rail line for the training and testing of models.This dataset contains rich rail images including single-rail,multi-rail,straight rail,curved rail,crossing rails,occlusion,blur,and different lighting conditions.To address the problem of the lack of deep learning-based rail line detection algorithms,we improve the CLRNet algorithm which has an excellent performance in lane detection,and propose the CLRNet-R algorithm for rail line detection.To address the problem of the rail line being thin and occupying fewer pixels in the image,making it difficult to distinguish from complex backgrounds,we introduce an attention mechanism to enhance global feature extraction ability and add a semantic segmentation head to enhance the features of the rail region by the binary probability of rail lines.To address the poor curve recognition performance and unsmooth output lines in the original CLRNet algorithm,we improve the weight allocation for line intersection-over-union calculation in the original framework and propose two loss functions based on local slopes to optimize the model’s local sampling point training constraints,improving the model’s fitting performance on curved rails and obtaining smooth and stable rail line detection results.Through experiments,this paper demonstrates that compared with other mainstream lane detection algorithms,the algorithm proposed in this paper has a better performance for rail line detection.展开更多
In foggy traffic scenarios, existing object detection algorithms face challenges such as low detection accuracy, poor robustness, occlusion, missed detections, and false detections. To address this issue, a multi-scal...In foggy traffic scenarios, existing object detection algorithms face challenges such as low detection accuracy, poor robustness, occlusion, missed detections, and false detections. To address this issue, a multi-scale object detection algorithm based on an improved YOLOv8 has been proposed. Firstly, a lightweight attention mechanism, Triplet Attention, is introduced to enhance the algorithm’s ability to extract multi-dimensional and multi-scale features, thereby improving the receptive capability of the feature maps. Secondly, the Diverse Branch Block (DBB) is integrated into the CSP Bottleneck with two Convolutions (C2F) module to strengthen the fusion of semantic information across different layers. Thirdly, a new decoupled detection head is proposed by redesigning the original network head based on the Diverse Branch Block module to improve detection accuracy and reduce missed and false detections. Finally, the Minimum Point Distance based Intersection-over-Union (MPDIoU) is used to replace the original YOLOv8 Complete Intersection-over-Union (CIoU) to accelerate the network’s training convergence. Comparative experiments and dehazing pre-processing tests were conducted on the RTTS and VOC-Fog datasets. Compared to the baseline YOLOv8 model, the improved algorithm achieved mean Average Precision (mAP) improvements of 4.6% and 3.8%, respectively. After defogging pre-processing, the mAP increased by 5.3% and 4.4%, respectively. The experimental results demonstrate that the improved algorithm exhibits high practicality and effectiveness in foggy traffic scenarios.展开更多
Accurate and real-time road defect detection is essential for ensuring traffic safety and infrastructure maintenance.However,existing vision-based methods often struggle with small,sparse,and low-resolution defects un...Accurate and real-time road defect detection is essential for ensuring traffic safety and infrastructure maintenance.However,existing vision-based methods often struggle with small,sparse,and low-resolution defects under complex road conditions.To address these limitations,we propose Multi-Scale Guided Detection YOLO(MGD-YOLO),a novel lightweight and high-performance object detector built upon You Only Look Once Version 5(YOLOv5).The proposed model integrates three key components:(1)a Multi-Scale Dilated Attention(MSDA)module to enhance semantic feature extraction across varying receptive fields;(2)Depthwise Separable Convolution(DSC)to reduce computational cost and improve model generalization;and(3)a Visual Global Attention Upsampling(VGAU)module that leverages high-level contextual information to refine low-level features for precise localization.Extensive experiments on three public road defect benchmarks demonstrate that MGD-YOLO outperforms state-of-the-art models in both detection accuracy and efficiency.Notably,our model achieves 87.9%accuracy in crack detection,88.3%overall precision on TD-RD dataset,while maintaining fast inference speed and a compact architecture.These results highlight the potential of MGD-YOLO for deployment in real-time,resource-constrained scenarios,paving the way for practical and scalable intelligent road maintenance systems.展开更多
Detecting individuals wearing safety helmets in complex environments faces several challenges.These factors include limited detection accuracy and frequent missed or false detections.Additionally,existing algorithms o...Detecting individuals wearing safety helmets in complex environments faces several challenges.These factors include limited detection accuracy and frequent missed or false detections.Additionally,existing algorithms often have excessive parameter counts,complex network structures,and high computational demands.These challenges make it difficult to deploy such models efficiently on resource-constrained devices like embedded systems.Aiming at this problem,this research proposes an optimized and lightweight solution called FGP-YOLOv8,an improved version of YOLOv8n.The YOLOv8 backbone network is replaced with the FasterNet model to reduce parameters and computational demands while local convolution layers are added.This modification minimizes computational costs with only a minor impact on accuracy.A new GSTA(GSConv-Triplet Attention)module is introduced to enhance feature fusion and reduce computational complexity.This is achieved using attention weights generated from dimensional interactions within the feature map.Additionally,the ParNet-C2f module replaces the original C2f(CSP Bottleneck with 2 Convolutions)module,improving feature extraction for safety helmets of various shapes and sizes.The CIoU(Complete-IoU)is replaced with the WIoU(Wise-IoU)to boost performance further,enhancing detection accuracy and generalization capabilities.Experimental results validate the improvements.The proposedmodel reduces the parameter count by 19.9% and the computational load by 18.5%.At the same time,mAP(mean average precision)increases by 2.3%,and precision improves by 1.2%.These results demonstrate the model’s robust performance in detecting safety helmets across diverse environments.展开更多
Aiming at the problems of insufficient feature extraction ability for small targets,complex image background,and low detection accuracy in marine life detection,this paper proposes a marine life detection algorithm SG...Aiming at the problems of insufficient feature extraction ability for small targets,complex image background,and low detection accuracy in marine life detection,this paper proposes a marine life detection algorithm SGW-YOLOv8 based on the improvement of YOLOv8.First,the Adaptive Fine-Grained Channel Attention(FCA)module is fused with the backbone layer of the YOLOv8 network to improve the feature extraction ability of the model.This paper uses the YOLOv8 network backbone layer to improve the feature extraction capability of the model.Second,the Efficient Multi-Scale Attention(C2f_EMA)module is replaced with the C2f module in the Neck layer of the network to improve the detection performance of the model for small underwater targets.Finally,the loss function is optimized to Weighted Intersection over Union(WIoU)to replace the original loss function,so that the model is better adapted to the target detection task in the complex ocean background.The improved algorithm has been experimented with on the Underwater Robot Picking Contest(URPC)dataset,and the results show that the improved algorithm achieves a detection accuracy of 84.5,which is 2.3%higher than that before the improvement,and at the same time,it can accurately detect the small-target marine organisms and adapts to the task of detecting marine organisms in various complex environments.展开更多
Plant diseases present a significant threat to global agricultural productivity, endangering both crop yields and quality. Traditional detection methods largely rely on manual inspection, a process that is not only la...Plant diseases present a significant threat to global agricultural productivity, endangering both crop yields and quality. Traditional detection methods largely rely on manual inspection, a process that is not only labor-intensive and time-consuming but also subject to subjective biases and dependent on operators’ expertise. Recent advancements in Transformer-based architectures have shown substantial progress in image classification tasks, particularly excelling in global feature extraction. However, despite their strong performance, the high computational complexity and large parameter requirements of Transformer models limit their practical application in plant disease detection. To address these constraints, this study proposes an optimized Efficient Swin Transformer specifically engineered to reduce computational complexity while enhancing classification accuracy. This model is an improvement over the Swin-T architecture, incorporating two pivotal modules: the Selective Token Generator and the Feature Fusion Aggregator. The Selective Token Generator minimizes the number of tokens processed, significantly increasing computational efficiency and facilitating multi-scale feature extraction. Concurrently, the Feature Fusion Aggregator adaptively integrates static and dynamic features, thereby enhancing the model’s ability to capture complex details within intricate environmental contexts.Empirical evaluations conducted on the PlantDoc dataset demonstrate the model’s superior classification performance, achieving a precision of 80.14% and a recall of 76.27%. Compared to the standard Swin-T model, the Efficient Swin Transformer achieves approximately 20.89% reduction in parameter size while improving precision by 4.29%. This study substantiates the potential of efficient token conversion techniques within Transformer architectures, presenting an effective and accurate solution for plant disease detection in the agricultural sector.展开更多
In order to solve the problems of limited performance of single and homogeneous sensors and high false alarm rate caused by environmental noise,heterogeneous sensor data fusion(HSDF),a fall detection framework based o...In order to solve the problems of limited performance of single and homogeneous sensors and high false alarm rate caused by environmental noise,heterogeneous sensor data fusion(HSDF),a fall detection framework based on heterogeneous sensor fusion of acceleration and audio signals,is proposed in this paper.By analyzing the heterogeneity of acceleration data and audio data,the framework uses Dempstershafer Theory(D-S Theory)to integrate the output of acceleration and audio data at decision level.Firstly,a normalized window interception algorithm——anomaly location window algorithm(ALW)is proposed by analyzing the fall process and the characteristics of acceleration changes based on acceleration data.Secondly,the one-dimensional residual convolutional network(1D-ReCNN)is designed for fall detection based on the audio data.Finally,it is verified that the HSDF framework has good advantages in terms of sensitivity and false alarm rate by the collection of volunteers’simulated fall data and free living data in real environment.展开更多
In response to challenges posed by complex backgrounds,diverse target angles,and numerous small targets in remote sensing images,alongside the issue of high resource consumption hindering model deployment,we propose a...In response to challenges posed by complex backgrounds,diverse target angles,and numerous small targets in remote sensing images,alongside the issue of high resource consumption hindering model deployment,we propose an enhanced,lightweight you only look once version 8 small(YOLOv8s)detection algorithm.Regarding network improvements,we first replace tradi-tional horizontal boxes with rotated boxes for target detection,effectively addressing difficulties in feature extraction caused by varying target angles.Second,we design a module integrating convolu-tional neural networks(CNN)and Transformer components to replace specific C2f modules in the backbone network,thereby expanding the model’s receptive field and enhancing feature extraction in complex backgrounds.Finally,we introduce a feature calibration structure to mitigate potential feature mismatches during feature fusion.For model compression,we employ a lightweight channel pruning technique based on localized mean average precision(LMAP)to eliminate redundancies in the enhanced model.Although this approach results in some loss of detection accuracy,it effec-tively reduces the number of parameters,computational load,and model size.Additionally,we employ channel-level knowledge distillation to recover accuracy in the pruned model,further enhancing detection performance.Experimental results indicate that the enhanced algorithm achieves a 6.1%increase in mAP50 compared to YOLOv8s,while simultaneously reducing parame-ters,computational load,and model size by 57.7%,28.8%,and 52.3%,respectively.展开更多
Defect detection is vital in the nonwoven material industry,ensuring surface quality before producing finished products.Recently,deep learning and computer vision advancements have revolutionized defect detection,maki...Defect detection is vital in the nonwoven material industry,ensuring surface quality before producing finished products.Recently,deep learning and computer vision advancements have revolutionized defect detection,making it a widely adopted approach in various industrial fields.This paper mainly studied the defect detection method for nonwoven materials based on the improved Nano Det-Plus model.Using the constructed samples of defects in nonwoven materials as the research objects,transfer learning experiments were conducted based on the Nano DetPlus object detection framework.Within this framework,the Backbone,path aggregation feature pyramid network(PAFPN)and Head network models were compared and trained through a process of freezing,with the ultimate aim of bolstering the model's feature extraction abilities and elevating detection accuracy.The half-precision quantization method was used to optimize the model after transfer learning experiments,reducing model weights and computational complexity to improve the detection speed.Performance comparisons were conducted between the improved model and the original Nano Det-Plus model,YOLO,SSD and other common industrial defect detection algorithms,validating that the improved methods based on transfer learning and semi-precision quantization enabled the model to meet the practical requirements of industrial production.展开更多
Real-time anomaly detection of massive data streams is an important research topic nowadays due to the fact that a lot of data is generated in continuous temporal processes. There is a broad research area, covering ma...Real-time anomaly detection of massive data streams is an important research topic nowadays due to the fact that a lot of data is generated in continuous temporal processes. There is a broad research area, covering mathematical, statistical, information theory methodologies for anomaly detection. It addresses various problems in a lot of domains such as health, education, finance, government, etc. In this paper, we analyze the state-of-the-art of data streams anomaly detection techniques and algorithms for anomaly detection in data streams (time series data). Critically surveying the techniques’ performances under the challenge of real-time anomaly detection of massive high-velocity streams, we conclude that the modeling of the normal behavior of the stream is a suitable approach. We evaluate Holt-Winters (HW), Taylor’s Double Holt-Winters (TDHW), Hierarchical temporal memory (HTM), Moving Average (MA), Autoregressive integrated moving average (ARIMA) forecasting models, etc. Holt-Winters (HW) and Taylor’s Double Holt-Winters (TDHW) forecasting models are used to predict the normal behavior of the periodic streams, and to detect anomalies when the deviations of observed and predicted values exceeded some predefined measures. In this work, we propose an enhancement of this approach and give a short description about the algorithms and then they are categorized by type of pre-diction as: predictive and non-predictive algorithms. We implement the Genetic Algorithm (GA) to periodically optimize HW and TDHW smoothing parameters in addition to the two sliding windows parameters that improve Hyndman’s MASE measure of deviation, and value of the threshold parameter that defines no anomaly confidence interval [1]. We also propose a new optimization function based on the input training datasets with the annotated anomaly intervals, in order to detect the right anomalies and minimize the number of false ones. The proposed method is evaluated on the known anomaly detection benchmarks NUMENTA and Yahoo datasets with annotated anomalies and real log data generated by the National education information system (NEIS)1 in Macedonia.展开更多
基金supported by the National Natural Science Foundation of China(No.62103298)。
文摘Aiming at the problems of low detection accuracy and large model size of existing object detection algorithms applied to complex road scenes,an improved you only look once version 8(YOLOv8)object detection algorithm for infrared images,F-YOLOv8,is proposed.First,a spatial-to-depth network replaces the traditional backbone network's strided convolution or pooling layer.At the same time,it combines with the channel attention mechanism so that the neural network focuses on the channels with large weight values to better extract low-resolution image feature information;then an improved feature pyramid network of lightweight bidirectional feature pyramid network(L-BiFPN)is proposed,which can efficiently fuse features of different scales.In addition,a loss function of insertion of union based on the minimum point distance(MPDIoU)is introduced for bounding box regression,which obtains faster convergence speed and more accurate regression results.Experimental results on the FLIR dataset show that the improved algorithm can accurately detect infrared road targets in real time with 3%and 2.2%enhancement in mean average precision at 50%IoU(mAP50)and mean average precision at 50%—95%IoU(mAP50-95),respectively,and 38.1%,37.3%and 16.9%reduction in the number of model parameters,the model weight,and floating-point operations per second(FLOPs),respectively.To further demonstrate the detection capability of the improved algorithm,it is tested on the public dataset PASCAL VOC,and the results show that F-YOLO has excellent generalized detection performance.
基金supported by the Shanxi Agricultural University Science and Technology Innovation Enhancement Project。
文摘This paper proposes a lightweight traffic sign detection system based on you only look once(YOLO).Firstly,the classification to fusion(C2f)structure is integrated into the backbone network,employing deformable convolution and bi-directional feature pyramid network(BiFPN)_Concat to improve the adaptability of the network.Secondly,the simple attention module(SimAm)is embedded to prioritize key features and reduce the complexity of the model after the C2f layer at the end of the backbone network.Next,the focal efficient intersection over union(EloU)is introduced to adjust the weights of challenging samples.Finally,we accomplish the design and deployment for the mobile app.The results demonstrate improvements,with the F1 score of 0.8987,mean average precision(mAP)@0.5 of 98.8%,mAP@0.5:0.95 of 75.6%,and the detection speed of 50 frames per second(FPS).
基金supported by the National Natural Science Foundation of China(Grant Nos.62101275 and 62101274).
文摘UAV-based object detection is rapidly expanding in both civilian and military applications,including security surveillance,disaster assessment,and border patrol.However,challenges such as small objects,occlusions,complex backgrounds,and variable lighting persist due to the unique perspective of UAV imagery.To address these issues,this paper introduces DAFPN-YOLO,an innovative model based on YOLOv8s(You Only Look Once version 8s).Themodel strikes a balance between detection accuracy and speed while reducing parameters,making itwell-suited for multi-object detection tasks from drone perspectives.A key feature of DAFPN-YOLO is the enhanced Drone-AFPN(Adaptive Feature Pyramid Network),which adaptively fuses multi-scale features to optimize feature extraction and enhance spatial and small-object information.To leverage Drone-AFPN’smulti-scale capabilities fully,a dedicated 160×160 small-object detection head was added,significantly boosting detection accuracy for small targets.In the backbone,the C2f_Dual(Cross Stage Partial with Cross-Stage Feature Fusion Dual)module and SPPELAN(Spatial Pyramid Pooling with Enhanced LocalAttentionNetwork)modulewere integrated.These components improve feature extraction and information aggregationwhile reducing parameters and computational complexity,enhancing inference efficiency.Additionally,Shape-IoU(Shape Intersection over Union)is used as the loss function for bounding box regression,enabling more precise shape-based object matching.Experimental results on the VisDrone 2019 dataset demonstrate the effectiveness ofDAFPN-YOLO.Compared to YOLOv8s,the proposedmodel achieves a 5.4 percentage point increase inmAP@0.5,a 3.8 percentage point improvement in mAP@0.5:0.95,and a 17.2%reduction in parameter count.These results highlight DAFPN-YOLO’s advantages in UAV-based object detection,offering valuable insights for applying deep learning to UAV-specific multi-object detection tasks.
基金supported by the Natural Science Foundation Project of Fujian Province,China(Grant No.2023J011439 and No.2019J01859).
文摘With the rapid urbanization and exponential population growth in China,two-wheeled vehicles have become a popular mode of transportation,particularly for short-distance travel.However,due to a lack of safety awareness,traffic violations by two-wheeled vehicle riders have become a widespread concern,contributing to urban traffic risks.Currently,significant human and material resources are being allocated to monitor and intercept non-compliant riders to ensure safe driving behavior.To enhance the safety,efficiency,and cost-effectiveness of traffic monitoring,automated detection systems based on image processing algorithms can be employed to identify traffic violations from eye-level video footage.In this study,we propose a robust detection algorithm specifically designed for two-wheeled vehicles,which serves as a fundamental step toward intelligent traffic monitoring.Our approach integrates a novel convolutional and attention mechanism to improve detection accuracy and efficiency.Additionally,we introduce a semi-supervised training strategy that leverages a large number of unlabeled images to enhance the model’s learning capability by extracting valuable background information.This method enables the model to generalize effectively to diverse urban environments and varying lighting conditions.We evaluate our proposed algorithm on a custom-built dataset,and experimental results demonstrate its superior performance,achieving an average precision(AP)of 95%and a recall(R)of 90.6%.Furthermore,the model maintains a computational efficiency of only 25.7 GFLOPs while achieving a high processing speed of 249 FPS,making it highly suitable for deployment on edge devices.Compared to existing detection methods,our approach significantly enhances the accuracy and robustness of two-wheeled vehicle identification while ensuring real-time performance.
基金supported in part by the National Natural Science Foundation of China under Grants 61861007in part by the Guizhou Province Science and Technology Planning Project ZK[2021]303in part by the Guizhou Province Science Technology Support Plan under Grants[2022]264,[2023]096,[2023]412 and[2023]409.
文摘Brain tumors are neoplastic diseases caused by the proliferation of abnormal cells in brain tissues,and their appearance may lead to a series of complex symptoms.However,current methods struggle to capture deeper brain tumor image feature information due to the variations in brain tumor morphology,size,and complex background,resulting in low detection accuracy,high rate of misdiagnosis and underdiagnosis,and challenges in meeting clinical needs.Therefore,this paper proposes the CMS-YOLO network model for multi-category brain tumor detection,which is based on the You Only Look Once version 10(YOLOv10s)algorithm.This model innovatively integrates the Convolutional Medical UNet extended block(CMUNeXt Block)to design a brand-new CSP Bottleneck with 2 convolutions(C2f)structure,which significantly enhances the ability to extract features of the lesion area.Meanwhile,to address the challenge of complex backgrounds in brain tumor detection,a Multi-Scale Attention Aggregation(MSAA)module is introduced.The module integrates features of lesions at different scales,enabling the model to effectively capture multi-scale contextual information and enhance detection accuracy in complex scenarios.Finally,during the model training process,the Shape-IoU loss function is employed to replace the Complete-IoU(CIoU)loss function for optimizing bounding box regression.This ensures that the predicted bounding boxes generated by the model closely match the actual tumor contours,thereby further enhancing the detection precision.The experimental results show that the improved method achieves 94.80%precision,93.60%recall,96.20%score,and 79.60%on the MRI for Brain Tumor with Bounding Boxes dataset.Compared to the YOLOv10s model,this represents improvements of 1.0%,1.1%,1.0%,and 1.1%,respectively.The method can achieve automatic detection and localization of three distinct categories of brain tumors—glioma,meningioma,and pituitary tumor,which can accurately detect and identify brain tumors,assist doctors in early diagnosis,and promote the development of early treatment.
文摘Pulmonary nodules represent an early manifestation of lung cancer.However,pulmonary nodules only constitute a small portion of the overall image,posing challenges for physicians in image interpretation and potentially leading to false positives or missed detections.To solve these problems,the YOLOv8 network is enhanced by adding deformable convolution and atrous spatial pyramid pooling(ASPP),along with the integration of a coordinate attention(CA)mechanism.This allows the network to focus on small targets while expanding the receptive field without losing resolution.At the same time,context information on the target is gathered and feature expression is enhanced by attention modules in different directions.It effectively improves the positioning accuracy and achieves good results on the LUNA16 dataset.Compared with other detection algorithms,it improves the accuracy of pulmonary nodule detection to a certain extent.
文摘Load time series analysis is critical for resource management and optimization decisions,especially automated analysis techniques.Existing research has insufficiently interpreted the overall characteristics of samples,leading to significant differences in load level detection conclusions for samples with different characteristics(trend,seasonality,cyclicality).Achieving automated,feature-adaptive,and quantifiable analysis methods remains a challenge.This paper proposes a Threshold Recognition-based Load Level Detection Algorithm(TRLLD),which effectively identifies different load level regions in samples of arbitrary size and distribution type based on sample characteristics.By utilizing distribution density uniformity,the algorithm classifies data points and ultimately obtains normalized load values.In the feature recognition step,the algorithm employs the Density Uniformity Index Based on Differences(DUID),High Load Level Concentration(HLLC),and Low Load Level Concentration(LLLC)to assess sample characteristics,which are independent of specific load values,providing a standardized perspective on features,ensuring high efficiency and strong interpretability.Compared to traditional methods,the proposed approach demonstrates better adaptive and real-time analysis capabilities.Experimental results indicate that it can effectively identify high load and low load regions in 16 groups of time series samples with different load characteristics,yielding highly interpretable results.The correlation between the DUID and sample density distribution uniformity reaches 98.08%.When introducing 10% MAD intensity noise,the maximum relative error is 4.72%,showcasing high robustness.Notably,it exhibits significant advantages in general and low sample scenarios.
基金Supported by the National Natural Science Foundation of China(61202458,61403109)the Natural Science Foundation of Heilongjiang Province of China(F2017021)and the Harbin Science and Technology Innovation Research Funds(2016RAQXJ036)
文摘Combining the principle of antibody concentration with the idea of biological evolution, this paper proposes an adaptive target detection algorithm for cloud service security based on Bio-Inspired Performance Evaluation Process Algebra(Bio-PEPA). The formal modelling of cloud services is formally modded by Bio-PEPA and the modules are transformed between cloud service internal structures and various components. Then, the security adaptive target detection algorithm of cloud service is divided into two processes, the short-term optimal action selection process which selects the current optimal detective action through the iterative operation of the expected function and the adaptive function, and the long-term detective strategy realized through the updates and eliminations of action planning table. The combination of the two processes reflects the self-adaptability of cloud service system to target detection. The simulating test detects three different kinds of security risks and then analyzes the relationship between the numbers of components with time in the service process. The performance of this method is compared with random detection method and three anomaly detection methods by the cloud service detection experiment. The detection time of this method is 50.1% of three kinds of detection methods and 86.3% of the random detection method. The service success rate is about 15% higher than that of random detection methods. The experimental results show that the algorithm has good time performance and high detection hit rate.
文摘Quality of Service (QoS) generally refers to measurable like latency and throughput, things that directly affect the user experience. Queuing (the most popular QoS tool) involves choosing the packets to be sent based on something other than arrival time. The Active queue management is important subject to manage this queue to increase the effectiveness of Transmission Control Protocol networks. Active queue management (AQM) is an effective means to enhance congestion control, and to achieve trade-off between link utilization and delay. The de facto standard, Random Early Detection (RED), and many of its variants employ queue length as a congestion indicator to trigger packet dropping. One of these enhancements of RED is FRED or Fair Random Early Detection attempts to deal with a fundamental aspect of RED in that it imposes the same loss rate on all flows, regardless of their bandwidths. FRED also uses per-flow active accounting, and tracks the state of active flows. FRED protects fragile flows by deterministically accepting flows from low bandwidth connections and fixes several shortcomings of RED by computing queue length during both arrival and departure of the packet. Unlike FRED, we propose a new scheme that used hazard rate estimated packet dropping function in FRED. We call this new scheme Enhancement Fair Random Early Detection. The key idea is that, with EFRED Scheme change packet dropping function, to get packet dropping less than RED and other AQM algorithms like ARED, REM, RED, etc. Simulations demonstrate that EFRED achieves a more stable throughput and performs better than current active queue management algorithms due to decrease the packets loss percentage and lowest in queuing delay, end to end delay and delay variation (JITTER).
基金supported by the Shanghai Science and Technology Innovation Action Plan High-Tech Field Project(Grant No.22511100601)for the year 2022 and Technology Development Fund for People’s Livelihood Research(Research on Transmission Line Deep Foundation Pit Environmental Situation Awareness System Based on Multi-Source Data).
文摘To maintain the reliability of power systems,routine inspections using drones equipped with advanced object detection algorithms are essential for preempting power-related issues.The increasing resolution of drone-captured images has posed a challenge for traditional target detection methods,especially in identifying small objects in high-resolution images.This study presents an enhanced object detection algorithm based on the Faster Regionbased Convolutional Neural Network(Faster R-CNN)framework,specifically tailored for detecting small-scale electrical components like insulators,shock hammers,and screws in transmission line.The algorithm features an improved backbone network for Faster R-CNN,which significantly boosts the feature extraction network’s ability to detect fine details.The Region Proposal Network is optimized using a method of guided feature refinement(GFR),which achieves a balance between accuracy and speed.The incorporation of Generalized Intersection over Union(GIOU)and Region of Interest(ROI)Align further refines themodel’s accuracy.Experimental results demonstrate a notable improvement in mean Average Precision,reaching 89.3%,an 11.1%increase compared to the standard Faster R-CNN.This highlights the effectiveness of the proposed algorithm in identifying electrical components in high-resolution aerial images.
基金the Sichuan Science and Technology Program(No.2022YFS0557)the National Natural Science Foundation of China(No.61972271)。
文摘In smart driving for rail transit,a reliable obstacle detection system is an important guarantee for the safety of trains.Therein,the detection of the rail area directly affects the accuracy of the system to identify dangerous targets.Both the rail line and the lane are presented as thin line shapes in the image,but the rail scene is more complex,and the color of the rail line is more difficult to distinguish from the background.By comparison,there are already many deep learning-based lane detection algorithms,but there is a lack of public datasets and targeted deep learning detection algorithms for rail line detection.To address this,this paper constructs a rail image dataset RailwayLine and labels the rail line for the training and testing of models.This dataset contains rich rail images including single-rail,multi-rail,straight rail,curved rail,crossing rails,occlusion,blur,and different lighting conditions.To address the problem of the lack of deep learning-based rail line detection algorithms,we improve the CLRNet algorithm which has an excellent performance in lane detection,and propose the CLRNet-R algorithm for rail line detection.To address the problem of the rail line being thin and occupying fewer pixels in the image,making it difficult to distinguish from complex backgrounds,we introduce an attention mechanism to enhance global feature extraction ability and add a semantic segmentation head to enhance the features of the rail region by the binary probability of rail lines.To address the poor curve recognition performance and unsmooth output lines in the original CLRNet algorithm,we improve the weight allocation for line intersection-over-union calculation in the original framework and propose two loss functions based on local slopes to optimize the model’s local sampling point training constraints,improving the model’s fitting performance on curved rails and obtaining smooth and stable rail line detection results.Through experiments,this paper demonstrates that compared with other mainstream lane detection algorithms,the algorithm proposed in this paper has a better performance for rail line detection.
基金supported by the National Natural Science Foundation of China(Grant Nos.62101275 and 62101274).
文摘In foggy traffic scenarios, existing object detection algorithms face challenges such as low detection accuracy, poor robustness, occlusion, missed detections, and false detections. To address this issue, a multi-scale object detection algorithm based on an improved YOLOv8 has been proposed. Firstly, a lightweight attention mechanism, Triplet Attention, is introduced to enhance the algorithm’s ability to extract multi-dimensional and multi-scale features, thereby improving the receptive capability of the feature maps. Secondly, the Diverse Branch Block (DBB) is integrated into the CSP Bottleneck with two Convolutions (C2F) module to strengthen the fusion of semantic information across different layers. Thirdly, a new decoupled detection head is proposed by redesigning the original network head based on the Diverse Branch Block module to improve detection accuracy and reduce missed and false detections. Finally, the Minimum Point Distance based Intersection-over-Union (MPDIoU) is used to replace the original YOLOv8 Complete Intersection-over-Union (CIoU) to accelerate the network’s training convergence. Comparative experiments and dehazing pre-processing tests were conducted on the RTTS and VOC-Fog datasets. Compared to the baseline YOLOv8 model, the improved algorithm achieved mean Average Precision (mAP) improvements of 4.6% and 3.8%, respectively. After defogging pre-processing, the mAP increased by 5.3% and 4.4%, respectively. The experimental results demonstrate that the improved algorithm exhibits high practicality and effectiveness in foggy traffic scenarios.
基金supported by Chengdu Jincheng College under the General Research Project Program(Project No.JG2024-1199)titled“Research on the Training Mechanism of Undergraduate Innovation Ability Based on Deep Integration of AI Industry-Education Collaboration”.
文摘Accurate and real-time road defect detection is essential for ensuring traffic safety and infrastructure maintenance.However,existing vision-based methods often struggle with small,sparse,and low-resolution defects under complex road conditions.To address these limitations,we propose Multi-Scale Guided Detection YOLO(MGD-YOLO),a novel lightweight and high-performance object detector built upon You Only Look Once Version 5(YOLOv5).The proposed model integrates three key components:(1)a Multi-Scale Dilated Attention(MSDA)module to enhance semantic feature extraction across varying receptive fields;(2)Depthwise Separable Convolution(DSC)to reduce computational cost and improve model generalization;and(3)a Visual Global Attention Upsampling(VGAU)module that leverages high-level contextual information to refine low-level features for precise localization.Extensive experiments on three public road defect benchmarks demonstrate that MGD-YOLO outperforms state-of-the-art models in both detection accuracy and efficiency.Notably,our model achieves 87.9%accuracy in crack detection,88.3%overall precision on TD-RD dataset,while maintaining fast inference speed and a compact architecture.These results highlight the potential of MGD-YOLO for deployment in real-time,resource-constrained scenarios,paving the way for practical and scalable intelligent road maintenance systems.
基金funded by National Natural Science Foundation of China(61741303)the Foundation Project of Guangxi Key Laboratory of Spatial Information andMapping(No.21-238-21-16).
文摘Detecting individuals wearing safety helmets in complex environments faces several challenges.These factors include limited detection accuracy and frequent missed or false detections.Additionally,existing algorithms often have excessive parameter counts,complex network structures,and high computational demands.These challenges make it difficult to deploy such models efficiently on resource-constrained devices like embedded systems.Aiming at this problem,this research proposes an optimized and lightweight solution called FGP-YOLOv8,an improved version of YOLOv8n.The YOLOv8 backbone network is replaced with the FasterNet model to reduce parameters and computational demands while local convolution layers are added.This modification minimizes computational costs with only a minor impact on accuracy.A new GSTA(GSConv-Triplet Attention)module is introduced to enhance feature fusion and reduce computational complexity.This is achieved using attention weights generated from dimensional interactions within the feature map.Additionally,the ParNet-C2f module replaces the original C2f(CSP Bottleneck with 2 Convolutions)module,improving feature extraction for safety helmets of various shapes and sizes.The CIoU(Complete-IoU)is replaced with the WIoU(Wise-IoU)to boost performance further,enhancing detection accuracy and generalization capabilities.Experimental results validate the improvements.The proposedmodel reduces the parameter count by 19.9% and the computational load by 18.5%.At the same time,mAP(mean average precision)increases by 2.3%,and precision improves by 1.2%.These results demonstrate the model’s robust performance in detecting safety helmets across diverse environments.
基金supported by 2023IT020 of the Industry-University-Research Innovation Fund for Chinese Universities-New Generation Information Technology Innovation ProgramPX-972024121 of the Education&Teaching Reform Program of Guangdong Ocean University。
文摘Aiming at the problems of insufficient feature extraction ability for small targets,complex image background,and low detection accuracy in marine life detection,this paper proposes a marine life detection algorithm SGW-YOLOv8 based on the improvement of YOLOv8.First,the Adaptive Fine-Grained Channel Attention(FCA)module is fused with the backbone layer of the YOLOv8 network to improve the feature extraction ability of the model.This paper uses the YOLOv8 network backbone layer to improve the feature extraction capability of the model.Second,the Efficient Multi-Scale Attention(C2f_EMA)module is replaced with the C2f module in the Neck layer of the network to improve the detection performance of the model for small underwater targets.Finally,the loss function is optimized to Weighted Intersection over Union(WIoU)to replace the original loss function,so that the model is better adapted to the target detection task in the complex ocean background.The improved algorithm has been experimented with on the Underwater Robot Picking Contest(URPC)dataset,and the results show that the improved algorithm achieves a detection accuracy of 84.5,which is 2.3%higher than that before the improvement,and at the same time,it can accurately detect the small-target marine organisms and adapts to the task of detecting marine organisms in various complex environments.
文摘Plant diseases present a significant threat to global agricultural productivity, endangering both crop yields and quality. Traditional detection methods largely rely on manual inspection, a process that is not only labor-intensive and time-consuming but also subject to subjective biases and dependent on operators’ expertise. Recent advancements in Transformer-based architectures have shown substantial progress in image classification tasks, particularly excelling in global feature extraction. However, despite their strong performance, the high computational complexity and large parameter requirements of Transformer models limit their practical application in plant disease detection. To address these constraints, this study proposes an optimized Efficient Swin Transformer specifically engineered to reduce computational complexity while enhancing classification accuracy. This model is an improvement over the Swin-T architecture, incorporating two pivotal modules: the Selective Token Generator and the Feature Fusion Aggregator. The Selective Token Generator minimizes the number of tokens processed, significantly increasing computational efficiency and facilitating multi-scale feature extraction. Concurrently, the Feature Fusion Aggregator adaptively integrates static and dynamic features, thereby enhancing the model’s ability to capture complex details within intricate environmental contexts.Empirical evaluations conducted on the PlantDoc dataset demonstrate the model’s superior classification performance, achieving a precision of 80.14% and a recall of 76.27%. Compared to the standard Swin-T model, the Efficient Swin Transformer achieves approximately 20.89% reduction in parameter size while improving precision by 4.29%. This study substantiates the potential of efficient token conversion techniques within Transformer architectures, presenting an effective and accurate solution for plant disease detection in the agricultural sector.
基金Supported by the National Natural Science Foundation of China(No.62172352,62171143,42306218)the Guangdong Provincial Department of Education Ocean Ranch Equipment Information and Intelligent Innovation Team Project(No.2023KCXTD016)+1 种基金the Natural Science Foundation of Hebei Province(No.2023407003)the Guangdong Ocean University Research Fund Project(No.060302102304).
文摘In order to solve the problems of limited performance of single and homogeneous sensors and high false alarm rate caused by environmental noise,heterogeneous sensor data fusion(HSDF),a fall detection framework based on heterogeneous sensor fusion of acceleration and audio signals,is proposed in this paper.By analyzing the heterogeneity of acceleration data and audio data,the framework uses Dempstershafer Theory(D-S Theory)to integrate the output of acceleration and audio data at decision level.Firstly,a normalized window interception algorithm——anomaly location window algorithm(ALW)is proposed by analyzing the fall process and the characteristics of acceleration changes based on acceleration data.Secondly,the one-dimensional residual convolutional network(1D-ReCNN)is designed for fall detection based on the audio data.Finally,it is verified that the HSDF framework has good advantages in terms of sensitivity and false alarm rate by the collection of volunteers’simulated fall data and free living data in real environment.
基金supported in part by the National Natural Foundation of China(Nos.52472334,U2368204)。
文摘In response to challenges posed by complex backgrounds,diverse target angles,and numerous small targets in remote sensing images,alongside the issue of high resource consumption hindering model deployment,we propose an enhanced,lightweight you only look once version 8 small(YOLOv8s)detection algorithm.Regarding network improvements,we first replace tradi-tional horizontal boxes with rotated boxes for target detection,effectively addressing difficulties in feature extraction caused by varying target angles.Second,we design a module integrating convolu-tional neural networks(CNN)and Transformer components to replace specific C2f modules in the backbone network,thereby expanding the model’s receptive field and enhancing feature extraction in complex backgrounds.Finally,we introduce a feature calibration structure to mitigate potential feature mismatches during feature fusion.For model compression,we employ a lightweight channel pruning technique based on localized mean average precision(LMAP)to eliminate redundancies in the enhanced model.Although this approach results in some loss of detection accuracy,it effec-tively reduces the number of parameters,computational load,and model size.Additionally,we employ channel-level knowledge distillation to recover accuracy in the pruned model,further enhancing detection performance.Experimental results indicate that the enhanced algorithm achieves a 6.1%increase in mAP50 compared to YOLOv8s,while simultaneously reducing parame-ters,computational load,and model size by 57.7%,28.8%,and 52.3%,respectively.
基金National Key Research and Development Program of China(Nos.2022YFB4700600 and 2022YFB4700605)National Natural Science Foundation of China(Nos.61771123 and 62171116)+1 种基金Fundamental Research Funds for the Central UniversitiesGraduate Student Innovation Fund of Donghua University,China(No.CUSF-DH-D-2022044)。
文摘Defect detection is vital in the nonwoven material industry,ensuring surface quality before producing finished products.Recently,deep learning and computer vision advancements have revolutionized defect detection,making it a widely adopted approach in various industrial fields.This paper mainly studied the defect detection method for nonwoven materials based on the improved Nano Det-Plus model.Using the constructed samples of defects in nonwoven materials as the research objects,transfer learning experiments were conducted based on the Nano DetPlus object detection framework.Within this framework,the Backbone,path aggregation feature pyramid network(PAFPN)and Head network models were compared and trained through a process of freezing,with the ultimate aim of bolstering the model's feature extraction abilities and elevating detection accuracy.The half-precision quantization method was used to optimize the model after transfer learning experiments,reducing model weights and computational complexity to improve the detection speed.Performance comparisons were conducted between the improved model and the original Nano Det-Plus model,YOLO,SSD and other common industrial defect detection algorithms,validating that the improved methods based on transfer learning and semi-precision quantization enabled the model to meet the practical requirements of industrial production.
文摘Real-time anomaly detection of massive data streams is an important research topic nowadays due to the fact that a lot of data is generated in continuous temporal processes. There is a broad research area, covering mathematical, statistical, information theory methodologies for anomaly detection. It addresses various problems in a lot of domains such as health, education, finance, government, etc. In this paper, we analyze the state-of-the-art of data streams anomaly detection techniques and algorithms for anomaly detection in data streams (time series data). Critically surveying the techniques’ performances under the challenge of real-time anomaly detection of massive high-velocity streams, we conclude that the modeling of the normal behavior of the stream is a suitable approach. We evaluate Holt-Winters (HW), Taylor’s Double Holt-Winters (TDHW), Hierarchical temporal memory (HTM), Moving Average (MA), Autoregressive integrated moving average (ARIMA) forecasting models, etc. Holt-Winters (HW) and Taylor’s Double Holt-Winters (TDHW) forecasting models are used to predict the normal behavior of the periodic streams, and to detect anomalies when the deviations of observed and predicted values exceeded some predefined measures. In this work, we propose an enhancement of this approach and give a short description about the algorithms and then they are categorized by type of pre-diction as: predictive and non-predictive algorithms. We implement the Genetic Algorithm (GA) to periodically optimize HW and TDHW smoothing parameters in addition to the two sliding windows parameters that improve Hyndman’s MASE measure of deviation, and value of the threshold parameter that defines no anomaly confidence interval [1]. We also propose a new optimization function based on the input training datasets with the annotated anomaly intervals, in order to detect the right anomalies and minimize the number of false ones. The proposed method is evaluated on the known anomaly detection benchmarks NUMENTA and Yahoo datasets with annotated anomalies and real log data generated by the National education information system (NEIS)1 in Macedonia.
