Real-time Detection Research Articles

MAFIKD: A Real-Time Pest Detection Method Based on Knowledge Distillation

MAFIKD: A Real-Time Pest Detection Method Based on Knowledge Distillation

Adaptive sensor attack detection and defense framework for autonomous vehicles based on density

The security of autonomous vehicles heavily depends on localization systems that integrate multiple sensors, which are vulnerable to sensor attacks and increase the risk of accidents. Given the diversity of sensor attacks and the dynamic changing of driving scenarios of autonomous vehicles, an adaptive and effective attack detection and defense framework faces a considerable challenge. This paper proposes a novel real-time adaptive attack detection and defense framework based on density, which can detect and identify attacked sensors and effectively recover data. We first develop a reinforcement learning multi-armed Bandit-based Density-Based Spatial Clustering of Applications with Noise (BDBSCAN) algorithm that selects hyperparameters adaptively. The Adaptive Extended Kalman Filter (AEKF) combines with the vehicle dynamic model on the localization system and extracts data features used for the BDBSCAN algorithm to monitor potential sensor attacks. If attack detection indicates possible system compromise, AEKF is further employed on localization sensors with anomalies identified through the BDBSCAN algorithm of the attacked sensors. To ensure precision and reliability, the data recovery incorporates a redundancy mechanism to apply a decision tree to select the optimal state estimation between AEKF and Extended Kalman Filter (EKF) to replace corrupted sensor data. To evaluate the effectiveness and adaptability of the proposed framework, we conducted 15,000 experiments using the real-world KITTI and V2V4Real datasets across various driving and sensor attack scenarios. The results demonstrate that our proposed framework achieves 100% accuracy and 0% false alarm rate in various driving scenarios for attack detection within 0.15 s, with a recovery time of 0.08 s.

Modeling and experimental characterization of a compact Mach–Zehnder electro-optic modulator on the SOI

The Mach–Zehnder electro-optic modulator (MZM) plays a crucial role in photonics integration technology during signal transmission. We propose the design and fabrication of a silicon-based electro-optic modulator based on the M-Z structure and design a modulator using silicon as the waveguide core layer on a silicon dioxide substrate. The detailed design involves a 1×2 splitter, branch waveguides, modulating arm waveguides, and a 2×1 combiner. The MZM fabrication and characterization results reveal that the half-wave voltage of the silicon-based MZM is 2 V, with an optical loss of −2.464dB and a device core size of 450µm×2800µm. The experimental results indicate that the MZM with a low half-wave voltage, a low loss, and high integration has significant value. The proposed MZM exhibits considerable improvements, featuring a low half-wave voltage and a low loss, and this device may serve as a fundamental component in wearable large-scale photonic integrated circuits for weak ECG real-time detection.

Application of multimodal intraoperative neurophysiological monitoring technology in neurofibromatosis type 1 related peripheral nerve tumor surgery

To summarize application effect and clinical experience of multimodal intraoperative neurophysiological monitoring (IONM) technology in the surgery of neurofibromatosis type 1 (NF1) related peripheral nerve tumors. A retrospective study was conducted on NF1 patients, who admitted between January 2019 and December 2023 and treated with peripheral nerve tumor resection surgery assisted by multimodal IONM technology. There were 49 males and 45 females. The age ranged from 5 to 78 years, with an average of 33.7 years. Tumor morphological classification included 71 cases of nodular type, 13 cases of diffuse type, and 10 cases of mixed type. Target tumors were distributed in craniofacial region (47 cases), neck (11 cases), trunk (12 cases), and limbs (24 cases). Preoperatively, 44 cases had no obvious neurological symptoms, while the remaining patients had neurological symptoms, including 15 cases of visual impairment, 5 cases of hearing impairment, 16 cases of somatic movement disorders, and 31 cases of somatic sensory disorders, of which 7 cases had more than one symptom. IONM plans were selected based on the relevant nerves and adjacent important structures of the target tumor, including visual evoked potential (17 cases), somatosensory evoked potential (44 cases), motor evoked potential (88 cases), and electromyogram (94 cases). All surgeries were successfully completed. Ninety-three patients underwent total/near total resection and 1 patient underwent palliative resection. Pathological examination showed 80 cases of neurofibroma and 14 cases of malignant peripheral nerve sheath tumors. Complications included 2 cases of hematoma and 3 cases of incision infection. All patients were followed up 3-61 months (median, 15 months). During follow-up, no significant changes in neurological symptoms or tumor recurrence were found. Among the patients with preoperative visual impairment, there were 14 cases with no improvement in symptoms and 1 with improvement after surgery. Among the patients with somatic movement disorders, there were 11 cases with no improvement in symptoms, 3 cases with improvement, 2 cases with aggravation, 4 newly onset cases, and 1 case with significant impact on daily life after surgery. Among the patients with somatic sensory disorders, there were 17 cases with no improvement in symptoms, 14 cases with improvement, and 13 newly onset cases. The patients with hearing impairment showed improvement after surgery. The clinical manifestations of NF1 related peripheral nerve tumors are complex. Multimodal IONM technology can provide real-time detection of nerve provocation and damage. Surgical treatment with multimodal IONM technology is safe and can reduce complications.

Efficient Credit Card Fraud Detection System Using Big Data and Machine Learning

This review offers a detailed strategy to address the growing threat of credit card fraud in today's digital landscape. By utilizing Big Data analytics alongside machine learning methods, the system aims to transform fraud detection processes. It tackles the challenges arising from the increasing volume and complexity of credit card transactions, enabling the real-time detection and prevention of fraudulent actions. The system employs sophisticated machine learning algorithms to identify patterns and anomalies linked to fraudulent activities, allowing for proactive responses to emerging fraud tactics. Additionally, the system is optimized to handle and analyze large datasets efficiently, ensuring timely and precise detection of fraud. It also incorporates strong security protocols to protect sensitive customer data while adhering to privacy regulations. This review ultimately seeks to enhance the safety and reliability of electronic payments, protecting financial institutions and consumers from the harmful effects of credit card fraud.

Preparation of a V–COF@SWCNTs-COOH/SPCE supported molecularly imprinted electrochemical sensor for real-time detection of trace sulfadimidine

To overcome the limitations of insufficient sensitivity and poor specificity of portable screen-printed carbon electrode-electrochemical sensors (SPCE-EC) in practical applications, we prepared carrier composites of carboxylic single-walled carbon nanotubes vertically grafted by covalent organic frameworks (v-COF@SWCNTs-COOH) and coated with a molecularly imprinted polymer (MIP) of sulfadimidine (SM2). 55 °C hot steam elution is more eco-friendly than traditional organic solvent elution. The results showed that when the mass ratio of DBA to DBA-SWCNTs was 1:1, the v-COF@SWCNTs-COOH obtained by the two-step synthesis method could increase the electrical signal up to 2.33-fold of the bare electrode. The bifunctional monomer MIP prepared on the above structure enhanced the signal response by 2.91-fold, with a high imprint factor of 20. The assembled MIP/v-COF@SWCNTs-COOH/SPCE were analyzed by differential pulse voltammetry (DPV) with a high sensitivity of 0.21 nM for LOD and 0.70 nM for LOQ. In milk and fish samples, the recovery rate was 95.0 %–104.8 %. The validation of authentic pork samples with the statutory LC-MS/MS method showed no significant difference (P > 0.05). The sensor's performance indicators remained robust after five repeated uses. Therefore, the MIP/v-COF@SWCNTs-COOH/SPCE combines the cheapness and portability of SPCE, while the sensitivity and specificity of small molecule detection were significantly improved.

A Low-Power IoT Architecture for Real-Time Detection and Notification of Unauthorized Tree Poaching

A Low-Power IoT Architecture for Real-Time Detection and Notification of Unauthorized Tree Poaching

Machine learning applications of network security enhancement: review

Machine learning (ML) is being used to improve intrusion detection mechanisms and identification in cyber security. Network data volume scaling (with the help of Machine learning) — Automated analysis and pattern recognition for large amounts of network-data, thereby detection of anomalies / potentially malicious activities that escape current rule-based techniques. By training ML models on historical data, these models can learn benign network behavior as well the anomalies in them that may result from malicious activities. The purpose of this essay/report is to begin taking a look under the hood at how ML can be used for security threat detection and analysis in-networking on an ongoing basis. This paper covers the automation of ML algorithms to enhance network security. Given the current state of electronic threats and their evolution, traditional security methods are typically insufficient. ML can analyze large volumes of data, learn patterns from it and this makes it suitable to complement network defense mechanisms. It then discusses different ML applications in network cybersecurity: intrusion detection, anomaly detection, spam and malware analysis; which the paper characterizes. It analyzes the potential, benefits and constrains of major ML methods in network security like supervised learning; unsupervised learning and reinforcement-learning. Finally, this paper represents recent progress in the use and impact of ML techniques along with case studies.The paper discusses the existing difficulties in the field such as the necessity for datasets and the vulnerability of machine learning models to adversarial attacks.T he paper also highlights avenues for exploration by focusing on developing scalable security solutions based on machine learning that are resilient and flexible.The goal of this examination is to offer both researchers and industry professionals valuable perspectives into the opportunities and obstacles linked to utilizing machine learning, in the domain of network security. ML methods have potential to improve network security by addressing the challenges posed by the increasing cyber threats that traditional security measures struggle to combat effectively over time. One key strength of ML lies in its capacity to analyze datasets and identify intricate patterns efficiently. The research paper delves into applications of ML in enhancing network security. The list covers security tools like Intrusion Detection Systems (IDS) examining malware and phishing attempts as well as anomalies in network activity and user behavior analysis (UEBA). The study explores both supervised and unsupervised learning methods. How they are used for quick threat detection and response in real time scenarios.You will find case studies and recent developments that showcase the implementation and effectiveness of these strategies.In addition the article delves into the obstacles linked to using machine learning techniques in network security including the necessity, for datasets, The paper's goal is to give an enlightening summary to scholars and practitioners about how machine learning can be applied to network security in order to provide solutions that are robust, adaptive, and scalable. To this end, it touches on several relevant aspects. One is the threat posed by adversarial attacks on the sorts of models that are likely to be used in this context. Another is the imperative, deriving from both adversarial threat and model drift, that models needed in this context be available in a form usable for continuous update. Keywords: Network Security, machine learning (ML), Intrusion Detection Systems (IDS), entity behavior analytics (UEBA).

Integrating Security Information and Event Management (SIEM) with Data Lakes and AI: Enhancing Threat Detection and Response

The evolving threat landscape in cybersecurity necessitates more advanced and efficient solutions for threat detection and response. Traditional Security Information and Event Management (SIEM) systems have limitations in handling large volumes of data and identifying sophisticated threats. This research explores the integration of SIEM solutions with data lakes, offering a scalable and flexible approach to managing security data. By leveraging artificial intelligence (AI) and machine learning (ML) algorithms, SIEM platforms can enhance their capabilities in real-time threat detection, automated response, and advanced analytics. This integration enables organizations to process vast amounts of structured and unstructured data from various sources, improving both the speed and accuracy of identifying security threats. The article examines the architecture, benefits, and challenges of combining SIEM with data lakes and AI, providing insights into how these technologies can collectively strengthen organizational security postures.

CMUT for ultrafast passive cavitation detection during ultrasound-induced blood–brain barrier disruption: proof of concept study

Objective.Cavitation dose monitoring plays a key role in ultrasound drug delivery to the brain. The use of capacitive micromachined ultrasonic transducer (CMUT) technology has a great potential for passive cavitation detection (PCD).Approach.Here, a circular (diameter 7 mm) CMUT centered at 5 MHz was designed to be inserted into a therapeutic transducer (1.5 MHz) used for ultrasound-induced blood-brain barrier (BBB) disruption on mice. CMUT-based real-time cavitation detection was performed during the ultrasound procedure (50μl intravenous injection of SonoVue microbubbles, frequency 1.5 MHz, PNP 480 kPa, duty Cycle 10%, PRF 10 Hz, duration 60 s). BBB disruption were confirmed by contrast-enhanced 7T-MRI.Main results.The CMUT device has a fractional bandwidth of 140%, almost twice a conventional piezocomposite PCD transducer. As expected, the CMUT device was able to detect the occurrence of harmonic, subharmonic and ultraharmonic frequencies as well as the increase of broadband signal indicating inertial cavitation in a wide frequency range (from 0.75 to 6 MHz). Signal-to-noise ratio was high enough (>40 dB) to perform ultrafast monitoring and follow the subtle intrapulse variations of frequency components at a rate of 10 kHz.Significance. This firstin vivoproof of concept demonstrates the interest of CMUT for PCD and encourages us to develop devices for PCD in larger animals by integrating an amplifier directly to the CMUT front-end to considerably increase the signal-to-noise ratio.

Hair-YOLO: a hair follicle detection model based on YOLOv8

Abstract Hair follicle detection technology has developed rapidly in recent years. Traditional manual detection methods are labor-intensive and inefficient. To address this problem, we propose a real-time hair follicle detection model called Hair-YOLO based on YOLOv8. This model focuses on accurately identifying the number of hairs within each follicle, providing precise data that help doctors assess hair density and follicle health of patients. First of all, we incorporate the re-parameterization Ghost module into the backbone, reducing the parameters and computational load. Then, the deformable convolution v3 operator is integrated into the neck network, enhancing adaptation to follicle shapes. Next, we propose a novel multi-scale feature perception separated and enhancement attention (Multi-SEAM) module, building upon the SEAM module, to address complex scalp scenarios. Furthermore, we enhance bounding box regression by replacing the standard complete intersection over union loss, with a modified point distance intersection over union loss. Finally, we construct a new hair follicle dataset and use it for a comparative analysis of Hair-YOLO and established models. Our model shows excellent performance, with a 14.26% increase in mAP@0.5:0.95, a 2.98% increase in Recall, and a 4.31% increase in Precision compared to the baseline.

Advancing Multi-Ion Sensing with Poly-Octylthiophene: 3D-Printed Milker-Implantable Microfluidic Device.

On-site rapid multi-ion sensing accelerates early identification of environmental pollution, water quality, and disease biomarkers in both livestock and humans. This study introduces a pocket-sized 3D-printed sensor, manufactured using additive manufacturing, specifically designed for detecting iron (Fe2+), nitrate (NO3 -), calcium (Ca2+), and phosphate (HPO4 2-). A unique feature of this device is its utilization of a universal ion-to-electron transducing layer made from highly redox-active poly-octylthiophene (POT), enabling an all-solid-state electrode tailored to each ion of interest. Manufactured with an extrusion-based 3D printer, the device features a periodic pattern of lateral layers (width = 80µm), including surface wrinkles. The superhydrophobic nature of the POT prevents the accumulation of nonspecific ions at the interface between the gold and POT layers, ensuring exceptional sensor selectivity. Lithography-free, 3D-printed sensors achieve sensitivity down to 1ppm of target ions in under a minute due to their 3D-wrinkled surface geometry. Integrated seamlessly with a microfluidic system for sample temperature stabilization, the printed sensor resides within a robust, pocket-sized 3D-printed device. This innovation integrates with milking parlors for real-time calcium detection, addressing diagnostic challenges in on-site livestock health monitoring, and has the capability to monitor water quality, soil nutrients, and human diseases.

Intelligent Vision System with Pruning and Web Interface for Real-Time Defect Detection on African Plum Surfaces

Agriculture stands as the cornerstone of Africa’s economy, supporting over 60% of the continent’s labor force. Despite its significance, the quality assessment of agricultural products remains a challenging task, particularly at a large scale, consuming valuable time and resources. The African plum is an agricultural fruit that is widely consumed across West and Central Africa but remains underrepresented in AI research. In this paper, we collected a dataset of 2892 African plum samples from fields in Cameroon representing the first dataset of its kind for training AI models. The dataset contains images of plums annotated with quality grades. We then trained and evaluated various state-of-the-art object detection and image classification models, including YOLOv5, YOLOv8, YOLOv9, Fast R-CNN, Mask R-CNN, VGG-16, DenseNet-121, MobileNet, and ResNet, on this African plum dataset. Our experimentation resulted in mean average precision scores ranging from 88.2% to 89.9% and accuracies between 86% and 91% for the object detection models and the classification models, respectively. We then performed model pruning to reduce model sizes while preserving performance, achieving up to 93.6% mean average precision and 99.09% accuracy after pruning YOLOv5, YOLOv8 and ResNet by 10–30%. We deployed the high-performing YOLOv8 system in a web application, offering an accessible AI-based quality assessment tool tailored for African plums. To the best of our knowledge, this represents the first such solution for assessing this underrepresented fruit, empowering farmers with efficient tools. Our approach integrates agriculture and AI to fill a key gap.

Insight of Employing Molecular Junctions for Sensor Applications.

Molecular junctions (MJs) exhibit distinct charge transport properties and have the potential to become the next generation of electronic devices. Advancing molecular electronics for practical uses, such as sensors, is crucial to propel its progress to the next level. In this review, we discussed how MJs can serve as a sensor for detecting a wide range of analytes with exceptional sensitivity and specificity. The primary advances and potential of molecular junctions for the various kinds of sensors including photosensors, explosives (DNTs, TNTs), cancer biomarker detection (DNA, mRNA), COVID detection, biogases (CO, NO, NH), environmental pH, practical chemicals, and water pollutants are listed and examined here. The fundamental ideas of molecular junction formation as well as the sensing mechanism have been examined here. This review demonstrates that MJ-based sensors hold significant promise for real-time and on-site detection. It provides valuable insights into current research and outlines potential future directions for advancing molecular junction-based sensors for practical applications.

All-weather airport runway foreign object detection based on mixed attention

Foreign object detection on airport runways plays a vital role in the safe takeoff and landing of aircraft. However, the existing detection algorithms have the phenomenon of missed detection and false detection under different lighting and weather runway environments. Therefore, a YOLOv5 foreign object detection algorithm suitable for all-weather airport runway environment is proposed. Aiming at the problem of a certain degree of feature loss in the pooling process of the original network, a cross-stage local spatial pyramid pooling module is designed, which can adaptively extract deep feature semantic information and enhance the multi-scale representation ability of the network; a hybrid attention module is introduced in the feature fusion part, and the channel and spatial feature weights are redistributed to strengthen the feature difference between foreign objects and irrelevant background elements; in view of the fact that small foreign objects are difficult to identify and locate, resulting in missed detection, a multi-scale positioning loss function is designed, and the detection ability of small targets is improved by adding similarity measurement; the optimized training strategy is used to train the MS-FOD dataset. Experimental results show that the improved algorithm achieves 95.83%mAP and Recall, which are respectively improved 94.31% by and 15.69% compared with the original YOLOv5 3.68% , and the detection speed is 68 frames /s , which meets the needs of real-time foreign object detection. The effectiveness of the proposed algorithm for foreign object detection on airport runways is effectively verified.

Potato Leaf Disease Detection Based on a Lightweight Deep Learning Model

Traditional methods of agricultural disease detection rely primarily on manual observation, which is not only time-consuming and labor-intensive, but also prone to human error. The advent of deep learning has revolutionized plant disease detection by providing more accurate and efficient solutions. The management of potato diseases is critical to the agricultural industry, as these diseases can lead to substantial losses in crop production. The prompt identification and classification of potato leaf diseases are essential to mitigating such losses. In this paper, we present a novel approach that integrates a lightweight convolutional neural network architecture, RegNetY-400MF, with transfer learning techniques to accurately identify seven different types of potato leaf diseases. The proposed method not only enhances the precision of potato leaf disease detection but also reduces the computational and storage demands, with a mere 0.40 GFLOPs and a model size of 16.8 MB. This makes it well-suited for use on edge devices with limited resources, enabling real-time disease detection in agricultural environments. The experimental results demonstrated that the accuracy of the proposed method in identifying seven potato leaf diseases was 90.68%, providing a comprehensive solution for potato crop management.

Efficient detection method for tiny defects on printed circuit boards based on focus information transfer and structurally flexible head

Abstract Aiming at the conflict between the defect feature recognition capability and the detection speed of current vision inspection techniques in the task of detecting tiny defects in printed circuit boards (PCBs). In this paper, we proposed an EPD-YOLO with a focus information transfer (FIT) structure and a structurally flexible head (SFHead). While improving the network’s ability to recognize tiny defects in similar PCBs through FIT, the feature information capturing capability of the dual SFHead structure is utilized to ensure detection accuracy and improve real-time detection speed. Experimental results show that the proposed EPD-YOLO has a mAP of 97.6 % , while the number of network parameters is only 5.13M, and it takes only 7.4 ms to detect an image, which achieves a better balance between accuracy and speed.

IRodd (intelligent-road damage detection) for real-time infrastructure preservation in detection, classification, calculation, and visualization

Preserving roads involves regularly evaluating government policy through advanced assessments using vehicles with specialized capabilities and high-resolution scanning technology. However, the cost is often not affordable due to a limited budget. Road surface surveys are highly expected to use low-cost tools and methods capable of being carried out comprehensively. This research aims to create a road damage detection application system by identifying and qualifying precisely the type of damage that occurs using a single CNN to detect objects in real time. Especially for the type of pothole, further analysis is to measure the volume or dimensions of the hole with a LiDAR smartphone. The study area is 38 province’s representative area in Indonesia. This research resulted in the iRodd (intelligent-road damage detection) for detection and classification per type of road damage in real-time object detection. Especially for the type of pothole damage, further analysis is carried out to obtain a damage volume calculation model and 3D visualization. The resulting iRodd model contributes in terms of completion (analyzing the parameters needed to be related to the road damage detection process), accuracy (precision), reliability (the level of reliability has high precision and is still within the limits of cost-effective), correct prediction (four-fifths of all positive objects that should be identified), efficient (object detection models strike a good balance between being able to recognize objects with high precision and being able to capture most objects that would otherwise be detected-high sensitivity), meanwhile, in the calculation of pothole volume, where the precision level is established according to the volume error value, comparing the derived data to the reference data with an average error of 5.35% with an RMSE value of 6.47 mm. The advanced iRodd model with LiDAR smartphone devices can present visualization and precision in efficiently calculating the volume of asphalt damage (potholes).

DETER Monitoring on Non-Forest Vegetation in the Brazilian Amazon

In this study, we present the initial results of the pilot project applying the Near Real-Time Deforestation Detection System (DETER), initially designed for monitoring forest ecosystems (DETER Amazon), recently expanded to surveil an area of approximately 280,000 km² of non-forest natural vegetation (NF) in the Brazilian Amazon. The system issues two types of alerts: 1) NF suppression alerts, categorized into three classes (suppression with exposed soil, suppression with cultivated area, and suppression by mining) and 2) burn scars, based on an adapted methodology from DETER Amazon and images with low spatial resolution (64 m) but high temporal resolution (two or three days to assess the entire biome). The pilot project period was from August 2022 to July 2023, with a total area of suppression alerts covering 575.22 km², predominantly composed of the NF suppression class with exposed soil, and 8,036.99 km² of burn scar areas. The majority of NF suppression alerts occurred in the states of Roraima, Mato Grosso, Rondônia, and Pará. Additionally, Amazonas and Amapá stood out among the states with a higher area of burn scar alerts. Savannas were the most affected non-forest physiognomies, with 5,037 km² (6.5%) of their area impacted by suppression and fire. Comparative analysis with other monitoring systems in the Amazon demonstrated the effectiveness of DETER NF in monitoring non-forest areas in the biome.

Secured Cyber-Internet Security in Intrusion Detection with Machine Learning Techniques

The rapid proliferation of Internet-connected devices has elevated the significance of cybersecurity, making intrusion detection a critical aspect of maintaining network integrity. Traditional security measures often fail to provide adequate protection against sophisticated attacks, necessitating advanced and robust solutions. This paper introduces a comprehensive cyber-internet security framework that leverages machine learning techniques for real-time intrusion detection and prevention. The proposed methodology employs a hybrid approach, integrating supervised and unsupervised learning models to detect anomalies and classify intrusions effectively. Specifically, a combination of Support Vector Machine (SVM), Decision Trees (DT), and K-means clustering is used to enhance detection accuracy and reduce false-positive rates. The experimental results demonstrate that the proposed model achieved a detection accuracy of 97.8%, a precision of 96.5%, and a recall of 95.2% on the NSL-KDD dataset. The implementation also reduced the false-positive rate to 1.2% and the computational overhead by 15% compared to traditional detection systems. Additionally, the proposed system was tested on real-time traffic data, where it successfully identified and mitigated various cyber threats, including Distributed Denial of Service (DDoS) attacks and network infiltrations, with minimal latency and high reliability. In conclusion, the study presents an efficient and secured cyber-internet security framework that significantly enhances intrusion detection capabilities using machine learning techniques. The proposed system provides a scalable and adaptive solution for securing critical infrastructure and networks against evolving cyber threats, making it an ideal candidate for deployment in real-world cybersecurity applications.