Detection Efficiency Research Articles

DVCW-YOLO for Printed Circuit Board Surface Defect Detection

The accurate and efficient detection of printed circuit board (PCB) surface defects is crucial to the electronic information manufacturing industry. However, current approaches to PCB defect detection face challenges, including large model sizes and difficulties in balancing detection accuracy with speed. To address these challenges, this paper proposes a novel PCB surface defect detection algorithm, named DVCW-YOLO. First, all standard convolutions in the backbone and neck networks of YOLOv8n are replaced with lightweight DWConv convolutions. In addition, a self-designed C2fCBAM module is introduced to the backbone network for extracting features. Next, within the neck structure, the C2f module is substituted with the more lightweight VOVGSCSP module, thereby reducing model redundancy, simplifying model complexity, and enhancing detection speed. By enhancing prominent features and suppressing less important ones, this modification allows the model to better focus on key regions, thereby improving feature representation capabilities. Finally, the WIoU loss function is implemented to replace the traditional CIoU function in YOLOv8n. This adjustment addresses issues related to low generalization and poor detection performance for small objects or complex backgrounds, while also mitigating the impact of low-quality or extreme samples on model accuracy. Experimental results demonstrate that the DVCW-YOLO model achieves a mean average precision (mAP) of 99.3% and a detection speed of 43.3 frames per second (FPS), which represent improvements of 4% and 4.08%, respectively, over the YOLOv8n model. These results confirm that the proposed model meets the real-time PCB defect detection requirements of small and medium-sized enterprises.

Optimizing Stereolithography Printing Parameters for Enhanced Microfluidic Chip Quality

Abstract In the pursuit of innovative biosensing technologies for critical applications such as early breast cancer detection, the development of efficient and portable devices is crucial. This work describes a unique stereolithography (SLA)-based three-dimensional–printed microfluidic device intended particularly for optofluidic biosensing with just microliter quantities of blood, similar to diabetes monitoring devices. Unlike typical cumbersome lab equipment such as the Biacore machine, which needs large blood sample volumes and laboratory processing, microfluidic technology allows for patient-operated, at-home testing, decreasing the requirement for hospital visits. The main contribution of this study is to optimize the SLA printing parameters, namely the exposure duration, in order to improve the microfluidic chip’s transparency and channel quality. This improvement allows for the exact immobilization of biorecognition components within the channels, resulting in sensitive and efficient biomarker detection. By extending the exposure duration, we considerably increase the structural integrity and optical clarity of the microfluidic channels, which are critical for successful biosignal transduction in labeled sensing applications. This development not only leads to a cheaper cost and faster manufacturing compared with conventional technologies but also offers increased performance in real bio-sensing applications. Thus, our work represents a big step forward in the development of accessible, efficient, and compact devices for early-stage illness diagnosis, outperforming existing lab-based diagnostics.

Neuromorphic computing for real-time adaptive penetration testing: analysis of human intuition in AI-dominated work space

This research investigates the revolutionary integration of neuromorphic computing with human intuition in the domain of real-time adaptive penetration testing, addressing the critical challenges facing modern cybersecurity in AI-dominated workspaces. The study presents a novel approach that combines the parallel processing capabilities of neuromorphic architectures with the nuanced decision-making abilities of human security experts, resulting in a hybrid system that significantly enhances threat detection and response capabilities. Our research implements a sophisticated neuromorphic computing model featuring 1024 input neurons, 2048 hidden layer neurons, and 512 output neurons, utilizing modified leaky integrate-and-fire algorithms for spike processing. The system incorporates real-time adaptive mechanisms that enable dynamic threat modeling and immediate response generation, while simultaneously learning from human expert intuition through a carefully designed collaborative interface. This architecture demonstrates remarkable improvements in both processing efficiency and threat detection accuracy, achieving a 94.7% true positive rate while maintaining an exceptionally low 2.3% false positive rate. The experimental methodology encompassed extensive testing across a diverse range of attack scenarios, including advanced persistent threats, zero-day vulnerabilities, and sophisticated social engineering attacks. The test environment comprised 500 virtual nodes distributed across multiple security zones, providing a realistic platform for comprehensive system evaluation. Performance metrics revealed significant improvements over traditional approaches, including an 89% success rate in adapting to previously unseen attack patterns and a 76% reduction in decision-making time for complex threats. Quantitative analysis demonstrates the system's superior capabilities in real-time threat detection and response, with average detection latencies of 1.2 milliseconds and consistent performance maintaining up to 100,000 concurrent connections. Qualitative assessment of human-AI synergy, conducted with 50 experienced penetration testers, revealed that 92% reported enhanced decision-making capabilities, while 95% experienced reduced cognitive load during complex scenarios. The research contributes significantly to the field by establishing a new paradigm for adaptive penetration testing that effectively combines neuromorphic computing efficiency with human intuitive expertise. The findings demonstrate that this integration not only enhances detection and response capabilities but also provides a more sustainable approach to cybersecurity in increasingly complex technological environments. Furthermore, the study opens new avenues for research in human-AI collaboration within cybersecurity, suggesting promising directions for future development in adaptive security systems. The implications of this research extend beyond immediate cybersecurity applications, offering insights into the broader field of human-AI collaboration in critical decision-making scenarios. The study also addresses important ethical considerations regarding AI autonomy in security operations, providing guidelines for responsible implementation of AI-driven security solutions while maintaining essential human oversight.

Study of neutron production for 360 GeV cosmic muons

The China Jinping underground Laboratory (CJPL) is an excellent location for studying solar, terrestrial, and supernova neutrinos due to its 2400-meter vertical rock overburden. Its unparalleled depth gives an opportunity to investigate the cosmic-ray muons with exceptionally high average energy at ∼360 GeV. This paper details a study of muon-related backgrounds based on 1178 days of data collected by the 1-ton prototype neutrino detector used for the Jinping Neutrino Experiment (JNE) since 2017. The apparent effects for the leakage of muons’ secondary particles due to detector’s finite size on the measured neutron yield are first discussed in detail. The analysis of 493 cosmic-ray muon candidates and 13.6±5.7 cosmogenic neutron candidates, along with a thorough evaluation of detection efficiency and uncertainties, gives a muon flux of (3.56±0.16stat±0.10syst)×10−10 cm−2 s−1 and a cosmogenic neutron yield of (3.37±1.41stat±0.31syst)×10−4 μ−1 g−1 cm2 in LAB-based liquid scintillator. Published by the American Physical Society 2024

Vibrating screen fault detection based on video frame prediction

ABSTRACT As a key technical equipment for improving coal quality, efficiency and promoting clean and efficient utilization, it is important for vibrating screens to detect their faults quickly and accurately. Traditional vibrating screen fault detection methods usually rely on the vibration signals collected by sensors, and identify the faults through spectral analysis, time domain feature extraction, etc. However, these methods face the problems of inaccurate detection, insufficient real-time performance and poor adaptability to environmental changes under complex working conditions. To address this limitation, this paper proposes a video frame prediction model, DST-UNet, to detect the operation of vibrating screens through video. The model combines convolutional units with different kernel sizes to simultaneously extract local detail information and global structural information in the operation of vibrating screens; uses ECA as a temporal attention mechanism to improve the model’s understanding and sensitivity to temporal sequences and employs convolutional units to generate spatio-temporal weights to dynamically fuse spatial and temporal features. In addition, an evaluation metric MTR is proposed to assess the model performance. The experimental results show that, compared with models such as SimVP, 3D convolution and PredRNN, the method improves the sensitivity, accuracy and processing speed of fault detection while reducing the model complexity and can realize efficient detection with limited data sets, which helps to identify equipment faults quickly and reduce the risk of damage.

ADVANCING EARLY SKIN CANCER DETECTION: A COMPARATIVE ANALYSIS OF MACHINE LEARNING ALGORITHMS FOR MELANOMA DIAGNOSIS USING DERMOSCOPIC IMAGES

Early detection of skin cancer, particularly melanoma, is crucial for improving patient outcomes and survival rates. Traditional diagnostic methods often require subjective interpretation by dermatologists, which can lead to inconsistent results. In recent years, machine learning algorithms, especially deep learning models such as Convolutional Neural Networks (CNNs), have shown promise in automating the analysis of medical images, enabling more accurate and efficient detection of skin cancer. This study investigates the performance of various machine learning models for skin cancer detection using the ISIC dataset, which consists of dermoscopic images. Six machine learning algorithms were evaluated: Logistic Regression, Decision Trees, Random Forest, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and CNN. The models were assessed based on their accuracy, precision, recall, F1-score, and Area Under the Curve (AUC). The results demonstrated that CNN outperformed all other models in terms of accuracy, AUC, and F1-score, making it the most effective algorithm for skin cancer detection in this study. While traditional machine learning algorithms like Random Forest and SVM showed promising results, CNNs' ability to automatically extract relevant features from complex images provided a significant advantage. The findings suggest that CNNs are particularly well-suited for early-stage skin cancer detection, although challenges related to model interpretability and dataset variability remain. This study highlights the potential of machine learning in revolutionizing skin cancer diagnosis and paves the way for future research focused on improving model robustness and clinical integration.

Atomistic insights on P implantation by using molecular ion beams for scalable spin-qubit arrays

Abstract Quantum information technologies hold immense promise, with quantum computers poised to revolutionize problem-solving capabilities. Among the leading contenders are solid-state spin-qubits, particularly those utilizing the spin of phosphorous donors (31P). While significant progress has been made in enhancing quantum coherence and qubit control, challenges persist, notably in achieving precise and scalable P placement in Si substrate. This paper investigates by means of molecular dynamics the use of molecular PF2 ions for implantation, aiming to reduce placement uncertainty while maintaining detection efficiency. We examine energy transfer, molecule integrity, implantation profiles, electronic signal components, and stable damage. Among other things we find that the assumption that the molecule only breaks apart immediately due to the presence of an a-SiO2 layer on the surface of the crystal and that the intensity of the electronic signal from ion-solid interactions does not correlate necessarily with the penetration depth of P.

Multi-Zone Fence Perimeter Surveillance: A New Edge-FOG Architecture for Efficient Detection and Classification of Intrusion

Multi-Zone Fence Perimeter Surveillance: A New Edge-FOG Architecture for Efficient Detection and Classification of Intrusion

Enhanced high-sensitivity multi-layer neutron detector based on LiF:ZnS(Ag) scintillator

This study proposes a novel, highly sensitive neutron detector design utilizing a unique multi-layered configuration. Each layer consists of a LiF: ZnS(Ag) scintillator coupled with a transparent neutron moderator that also functions as a light guide for the Silicon Photomultiplier (SiPM) light sensor. This design offers a cost-effective and readily available alternative for existing neutron detectors. The research focused on optimizing a single layer for maximum sensitivity and proper gamma rejection capabilities. This optimized configuration is then replicated to form the multi-layer detector. A primary challenge with the LiF:ZnS(Ag) scintillator is its inherent opacity, limiting its width and consequently, its detection efficiency. Our proposed multi-layer structure addresses this limitation by employing a thin scintillator in each layer. This strategic design minimizes emitted light attenuation while simultaneously enhancing sensitivity through the cumulative effect of multiple layers. Our experiments demonstrate a significant improvement in detection efficiency compared to the single-layer setup. Additionally, our architecture offers an actual improvement in differentiating between gamma and neutron signals. By analyzing count rates across the detector’s layers, we gain valuable operational insights, such as the ability to predict the source direction. Our finding demonstrates an improved sensitivity achieved by minimal loss of neutrons during the moderation of 329% compared to a single layer, aligned with the potential range of improvement while maintaining extremely high gamma rejection. Supported by the presented findings, this design represents a noticeable advancement over existing solutions, offering scalable customization to user requirements. Notably, it outperforms traditional 3He tube-based detection configurations, positioning it as a compelling and advantageous replacement option.

Secure cloud computing: leveraging GNN and leader K-means for intrusion detection optimization

Over the past two decades, cloud computing has experienced exponential growth, becoming a critical resource for organizations and individuals alike. However, this rapid adoption has introduced significant security challenges, particularly in intrusion detection, where traditional systems often struggle with low detection accuracy and high processing times. To address these limitations, this research proposes an optimized Intrusion Detection System (IDS) that leverages Graph Neural Networks and the Leader K-means clustering algorithm. The primary aim of the study is to enhance both the accuracy and efficiency of intrusion detection within cloud environments. Key contributions of this work include the integration of the Leader K-means algorithm for effective data clustering, improving the IDS’s ability to differentiate between normal and malicious activities. Additionally, the study introduces an optimized Grasshopper Optimization algorithm, which enhances the performance of the Optimal Neural Network, further refining detection accuracy. For added data security, the system incorporates Advanced Encryption Standard encryption and steganography, ensuring robust protection of sensitive information. The proposed solution has been implemented on the Java platform with CloudSim support, and the findings demonstrate a significant improvement in both detection accuracy and processing efficiency compared to existing methods. This research presents a comprehensive solution to the ongoing security challenges in cloud computing, offering a valuable contribution to the field.

Enhancing practicality and efficiency of deepfake detection

The proliferation of deepfake generation has become increasingly widespread. Current solutions for automatically detecting and classifying generated content require substantial computational resources, making them impractical for use by the average non-expert individual, particularly from edge computing applications. In this paper, we propose a series of techniques to accelerate the inference speed of deepfake detection on video data. We also draw inspiration from steganalysis approaches to expose deepfakes as any secret payloads encoded in the image. Furthermore, some key considerations were identified to significantly reduce the size of the core convolutional neural network. The experiment yielded competitive results when evaluated on two second-generation deepfake datasets, namely Celeb-DFv2 and DFDC, while requiring only a fraction of the typical computational cost and resources.

Fast Rumor Detection in Social Networks Through Large Language Models-Based Semantic Enhancement

With the popularity of social media, the rapid spread of rumors has brought great challenges to social stability and personal safety. Therefore, it is of great significance to develop an efficient and accurate rumor detection method. This study first reviews this research status of social network rumor detection, analyzes the advantages and disadvantages of traditional methods and deep learning-based methods, and then completes network language feature extraction based on Transformer, uses KIMI large language model (LLM) for semantic enhancement, and then completes the establishment of semantic feature fusion module. Finally, KIMI and GPT are integrated to construct a hybrid LLM for fast detection of Internet rumors. In order to verify the effectiveness of this method, this study obtains data sets from MicroBlog community management center, Twitter and TikTok for experiments. The experimental results show that the semantic enhancement technology based on the mixed LLM performs well in the rumor detection task, which is significantly better than the single LLM. In addition, we also discuss the effect of semantic enhancement strategy on model performance, and find that this strategy can further improve the accuracy and robustness of the model. The experimental results verify the effectiveness and practicability of this method, and provide a new idea and method for the rapid detection of social network rumors.

Onco-mNGS facilitates rapid and precise identification of the etiology of fever of unknown origin: a single-centre prospective study in North China

ObjectivesDelayed diagnosis of patients with Fever of Unknown Origin has long been a daunting clinical challenge. Onco-mNGS, which can accurately diagnose infectious agents and identify suspected tumor signatures by analyzing host chromosome copy number changes, has been widely used to assist identifying complex etiologies. However, the application of Onco-mNGS to improve FUO etiological screening has never been studied before.MethodsIn this single-centre prospective study, we included 65 patients with classic FUO, who were randomly divided into control group (sample cultivation) and mNGS group (cultivation + Onco-mNGS). We analyzed the infectious agents and symbiotic microbiological, tumor and clinical data of both groups.ResultsInfection-related pathogenic detection efficiency rose from 25% (control group) to 48.48% (experimental group). Seven patients with chromosome copy number changes had later been confirmed tumors, indicating a 100% of clinical concordance rate of Onco-mNGS for tumors. In addition, the time frame for diagnosing or ruling out infection/tumor with Onco-mNGS had greatly reduced to approximately 2 days, which was 7.34 days earlier than that in the control group.ConclusionsOnco-mNGS is an ideal rapid diagnostic aid to assist improving the early diagnostic efficiency of FUO-associated diseases.

A framework for hardware trojan detection based on contrastive learning

With the rapid development of the semiconductor industry, Hardware Trojans (HT) as a kind of malicious function that can be implanted at will in all processes of integrated circuit design, manufacturing, and deployment have become a great threat in the field of hardware security. Side-channel analysis is widely used in the detection of HT due to its high efficiency, non-contact nature, and accuracy. In this paper, we propose a framework for HT detection based on contrastive learning using power consumption information in unsupervised or weakly supervised scenarios. First, the framework augments the data, such as creatively using a one-dimensional discrete chaotic mapping to disturb the data to achieve data augmentation to improve the generalization capabilities of the model. Second, the model representation is learned by comparing the similarities and differences between samples, freeing it from the dependence on labels. Finally, the detection of HT is accomplished more efficiently by categorizing the side information during circuit operation through the backbone network. Experiments on data from nine different public HTs show that the proposed method exhibits better generalization capabilities using the same network model within a comparative learning framework. The model trained on the dataset of small Trojan T100 has a detection efficiency advantage of up to 44% in detecting large Trojans, while the model trained on the dataset of large Trojan T2100 has a detection efficiency advantage of up to 10% in detecting small Trojans. The results in data imbalanced and noisy environments also show that the contrastive learning framework in this paper can better fulfill the requirements of detecting unknown HT in unsupervised or weakly supervised scenarios.

IoT Security: Botnet Detection Using Self-Organizing Feature Map and Machine Learning

The rapid advancement of Internet of Things (IoT) technology has created potential for progress in various aspects of life. However, the increasing number of IoT devices also raises the risk of cyberattacks, particularly IoT botnets often exploited by attackers. This is largely due to the limitations of IoT devices, such as constraints in capacity, power, and memory, necessitating an efficient detection system. This study aims to develop a resource-efficient botnet detection system by using the Self-Organizing Feature Map (SOFM) dimensionality reduction method in combination with machine learning algorithms. The proposed method includes a feature engineering process using SOFM to address high-dimensional data, followed by classification with various machine learning algorithms. The experiments evaluate performance based on accuracy, sensitivity, specificity, False Positive Rate (FPR), and False Negative Rate (FNR). Results show that the Decision Tree algorithm achieved the highest accuracy rate of 97.24%, with a sensitivity of 0.9523, specificity of 0.9932, and a fast execution time of 100.66 seconds. The use of SOFM successfully reduced memory consumption from 3.08 GB to 923MB. Experimental results indicate that this approach is effective for enhancing IoT security in resource-constrained devices.

Non-destructive inspection of interfacial debonding in concrete-filled steel tube bridges

ABSTRACT Concrete-filled steel tube (CFST) structures are widely used in the construction of large-scale structures owing to its exceptional load-bearing capacity. However, inadequate construction quality and improper maintenance may cause debonding defects at the steel-concrete interface. Two non-destructive testing methods, i.e. the ultrasonic phased array (UPA) method and the impact acoustic method, are developed for deboning detection. A combined inspection procedure is proposed to balance the detection accuracy and efficiency, where the impact acoustic method is utilised for a rapid preliminary inspection to pinpoint the suspected debonding areas, while the UPA method is sequentially applied to accurately evaluate the debonding rates in the suspected areas. The impact acoustic method analyzes the time-frequency spectrum from the echo signals using wavelet transform, and calculates the vibration energy ratio for debonding detection. The UPA method applies the total focusing method for image reconstruction, and extracts the reflection coefficient at the steel–concrete interface for debonding evaluation. The laboratory experimental results show that the combined method possesses a high accuracy and accurately delineates the range of the debonding defect while preserving the detection efficiency. Field tests conducted on an arch bridge in Foshan, China, verified the effectiveness of the proposed evaluation method.

Challenges in Damage Detection for Real-time Monitoring and Predictive Maintenance Methodology

Damage detection techniques offer a sophisticated and effective way to monitor and assess the health of engineering systems, but they come with challenges related to model accuracy, sensor data quality, and the detection of subtle or distributed damage. Advances in computational methods, sensor technology, and hybrid techniques continue to improve the practical application of these methods, making them increasingly relevant in real-time condition monitoring and predictive maintenance systems. The proposed methodology aims to develop damage detection systems that play a key role in enabling real-time monitoring and early identification of potential problems. This approach significantly reduces the need for conventional checks, leading to time and cost savings. By predicting the remaining service life of components, the methodology helps optimize maintenance schedules and prevent unexpected failures. Additionally, the use of smart materials and sensors enhances the accuracy and efficiency of damage detection, allowing more proactive, data-driven damage detection and monitoring systems.

An Efficient Group Convolution and Feature Fusion Method for Weed Detection

Weed detection is a crucial step in achieving intelligent weeding for vegetables. Currently, research on vegetable weed detection technology is relatively limited, and existing detection methods still face challenges due to complex natural conditions, resulting in low detection accuracy and efficiency. This paper proposes the YOLOv8-EGC-Fusion (YEF) model, an enhancement based on the YOLOv8 model, to address these challenges. This model introduces plug-and-play modules: (1) The Efficient Group Convolution (EGC) module leverages convolution kernels of various sizes combined with group convolution techniques to significantly reduce computational cost. Integrating this EGC module with the C2f module creates the C2f-EGC module, strengthening the model’s capacity to grasp local contextual information. (2) The Group Context Anchor Attention (GCAA) module strengthens the model’s capacity to capture long-range contextual information, contributing to improved feature comprehension. (3) The GCAA-Fusion module effectively merges multi-scale features, addressing shallow feature loss and preserving critical information. Leveraging GCAA-Fusion and PAFPN, we developed an Adaptive Feature Fusion (AFF) feature pyramid structure that amplifies the model’s feature extraction capabilities. To ensure effective evaluation, we collected a diverse dataset of weed images from various vegetable fields. A series of comparative experiments was conducted to verify the detection effectiveness of the YEF model. The results show that the YEF model outperforms the original YOLOv8 model, Faster R-CNN, RetinaNet, TOOD, RTMDet, and YOLOv5 in detection performance. The detection metrics achieved by the YEF model are as follows: precision of 0.904, recall of 0.88, F1 score of 0.891, and mAP0.5 of 0.929. In conclusion, the YEF model demonstrates high detection accuracy for vegetable and weed identification, meeting the requirements for precise detection.

Using a YOLO Deep Learning Algorithm to Improve the Accuracy of 3D Object Detection by Autonomous Vehicles

This study presents an adaptation of the YOLOv4 deep learning algorithm for 3D object detection, addressing a critical challenge in autonomous vehicle (AV) systems: accurate real-time perception of the surrounding environment in three dimensions. Traditional 2D detection methods, while efficient, fall short in providing the depth and spatial information necessary for safe navigation. This research modifies the YOLOv4 architecture to predict 3D bounding boxes, object depth, and orientation. Key contributions include introducing a multi-task loss function that optimizes 2D and 3D predictions and integrating sensor fusion techniques that combine RGB camera data with LIDAR point clouds for improved depth estimation. The adapted model, tested on real-world datasets, demonstrates a significant increase in 3D detection accuracy, achieving a mean average precision (mAP) of 85%, intersection over union (IoU) of 78%, and near real-time performance at 93–97% for detecting vehicles and 75–91% for detecting people. This approach balances high detection accuracy and real-time processing, making it highly suitable for AV applications. This study advances the field by showing how an efficient 2D detector can be extended to meet the complex demands of 3D object detection in real-world driving scenarios without sacrificing computational efficiency.

Addressing Class Imbalance in Intrusion Detection: A Comprehensive Evaluation of Machine Learning Approaches

The ever-growing number of cyber attacks in today’s digitally interconnected world requires highly efficient intrusion detection systems (IDSs), which accurately identify both frequent and rare network intrusions. One of the most important challenges in IDSs is the class imbalance problem of network traffic flow data, where benign traffic flow significantly outweighs attack instances. This directly affects the ability of machine learning models to identify minority class threats. This paper is intended to evaluate various machine learning algorithms under different levels of class imbalances, using resampling as a strategy for this problem. The paper will provide an experimental comparison by combining various algorithms for classification and class imbalance learning, assessing the performance through the F1-score and geometric mean (G-mean). The work will contribute to creating robust and adaptive IDS through the judicious integration of resampling with machine learning models, thus helping the domain of cybersecurity to become resilient.