High Recall Rate Research Articles

Comparison of machine learning algorithms and feature visualization analysis for diabetes risk prediction

There are already 537 million people suffering from diabetes worldwide, according to the research data released by the International Diabetes Federation in 2022. It is a frequent and pervasive condition that has overtaken cancer as the second leading cause of death from modern diseases and is second only to it in terms of the harm it causes to the human body. The excessive morbidity, mortality, shortened life expectancy, and economic and other costs associated with it make it an important condition that plagues people. This study selected the Diabetes Health Indicators Dataset from the Kaggle database based on machine learning algorithm prediction models in predicting the risk of diabetes. 70% of the data were randomly chosen from the presence of high blood pressure (High BP), high cholesterol (High Chol), body mass index (BMI), smoking status (Smoker), and the risk of diabetes, smoking status (Smoker), health status, age, education, income, and gender as independent variables, and whether or not they had diabetes as dependent variables, and built diabetes prediction models based on five traditional machine algorithms such as random forests, logistic regression, decision trees, Support Vector Machine (SVM), XGBoost, and deep neural networks (DNN). The models were then validated with 30% of the samples. Three indicators, Accuracy, AUC, and Recall, were compared and analyzed, and DNN was found to have a high prediction accuracy (75.49%) and recall rate (79%).

Lightweight Intrusion Detection Method of Vehicle CAN Bus under Computational Resource Constraints

In order to improve the security protection performance of the vehicle Controller Area Network (CAN) bus, the research builds an adaptive lightweight intrusion detection algorithm based on the limited computing and storage resources of the on-board ECU environment and the message cycle characteristics to supervise and detect the vehicle CAN bus intrusion. The results showed that the message cycle-based adaptive intrusion detection algorithm had high accuracy and recall rate, and fast computational search efficiency, with a stable detection time of less than 3 seconds. The intrusion detection capability is continuously optimized as the training time increases, and after stabilization, the resource utilization rate reaches over 95% with a throughput of 100Mb/s. The algorithm has strong protection capabilities. The average vehicle CPU usage of the algorithm is only 4.76%, which is 10.17% lower than the intrusion detection algorithm based on support vector machines. It can effectively prevent interference with the normal operation of the vehicle CAN bus. The algorithm has high detection accuracy for interrupt type attacks, and there are no false positives or missed alarms. For injection type attacks, the probability of missed alarms is less than 1%. The intrusion detection of vehicle CAN bus based on the message cycle characteristics provides technical reference for the safety and stability of the vehicle network, and has important practical value for the intelligent and networked development of the automobile industry.

Fusion extracted features from deep learning for identification of multiple positioning errors in dental panoramic imaging.

Dental panoramic imaging plays a pivotal role in dentistry for diagnosis and treatment planning. However, correctly positioning patients can be challenging for technicians due to the complexity of the imaging equipment and variations in patient anatomy, leading to positioning errors. These errors can compromise image quality and potentially result in misdiagnoses. This research aims to develop and validate a deep learning model capable of accurately and efficiently identifying multiple positioning errors in dental panoramic imaging. This retrospective study used 552 panoramic images selected from a hospital Picture Archiving and Communication System (PACS). We defined six types of errors (E1-E6) namely, (1) slumped position, (2) chin tipped low, (3) open lip, (4) head turned to one side, (5) head tilted to one side, and (6) tongue against the palate. First, six Convolutional Neural Network (CNN) models were employed to extract image features, which were then fused using transfer learning. Next, a Support Vector Machine (SVM) was applied to create a classifier for multiple positioning errors, using the fused image features. Finally, the classifier performance was evaluated using 3 indices of precision, recall rate, and accuracy. Experimental results show that the fusion of image features with six binary SVM classifiers yielded high accuracy, recall rates, and precision. Specifically, the classifier achieved an accuracy of 0.832 for identifying multiple positioning errors. This study demonstrates that six SVM classifiers effectively identify multiple positioning errors in dental panoramic imaging. The fusion of extracted image features and the employment of SVM classifiers improve diagnostic precision, suggesting potential enhancements in dental imaging efficiency and diagnostic accuracy. Future research should consider larger datasets and explore real-time clinical application.

Population-wide evaluation of artificial intelligence and radiologist assessment of screening mammograms

ObjectivesTo validate an AI system for standalone breast cancer detection on an entire screening population in comparison to first-reading breast radiologists.Materials and methodsAll mammography screenings performed between August 4, 2014, and August 15, 2018, in the Region of Southern Denmark with follow-up within 24 months were eligible. Screenings were assessed as normal or abnormal by breast radiologists through double reading with arbitration. For an AI decision of normal or abnormal, two AI-score cut-off points were applied by matching at mean sensitivity (AIsens) and specificity (AIspec) of first readers. Accuracy measures were sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and recall rate (RR).ResultsThe sample included 249,402 screenings (149,495 women) and 2033 breast cancers (72.6% screen-detected cancers, 27.4% interval cancers). AIsens had lower specificity (97.5% vs 97.7%; p < 0.0001) and PPV (17.5% vs 18.7%; p = 0.01) and a higher RR (3.0% vs 2.8%; p < 0.0001) than first readers. AIspec was comparable to first readers in terms of all accuracy measures. Both AIsens and AIspec detected significantly fewer screen-detected cancers (1166 (AIsens), 1156 (AIspec) vs 1252; p < 0.0001) but found more interval cancers compared to first readers (126 (AIsens), 117 (AIspec) vs 39; p < 0.0001) with varying types of cancers detected across multiple subgroups.ConclusionStandalone AI can detect breast cancer at an accuracy level equivalent to the standard of first readers when the AI threshold point was matched at first reader specificity. However, AI and first readers detected a different composition of cancers.Clinical relevance statementReplacing first readers with AI with an appropriate cut-off score could be feasible. AI-detected cancers not detected by radiologists suggest a potential increase in the number of cancers detected if AI is implemented to support double reading within screening, although the clinicopathological characteristics of detected cancers would not change significantly.Key Points• Standalone AI cancer detection was compared to first readers in a double-read mammography screening population.• Standalone AI matched at first reader specificity showed no statistically significant difference in overall accuracy but detected different cancers.• With an appropriate threshold, AI-integrated screening can increase the number of detected cancers with similar clinicopathological characteristics.

Construction of a cross-domain integration teaching model for physical education oriented to the characteristics of disciplines

Abstract In this paper, we first use SemiSDAE and matrix decomposition to model the source and target domains according to the subject features and convert the encoder feature space to the matrix decomposition feature space through the feature transformation function to deeply integrate the features extracted from both. Combining the implicit feedback information of users for cross-domain recommendation and retaining the semantic information, the features of various disciplines are integrated into the online and offline cross-domain teaching based on the SPOC platform. The method in this paper has a high recall rate of 0.99 in 60% of training data, and 0.96 in 80% of training data, and the results of various sports programs are excellent, which can organically combine sports with other disciplines and promote the development and innovation of sports teaching.

Ensemble-Based Machine Learning Approach for Real-Time Person Counting in an Instant Attendance System

Real-time attendance systems have become indispensable in various domains, including educational institutions and workplaces, as they automate attendance tracking and improve efficiency. This paper introduces a robust real-time attendance system that combines OpenCV and the You Only Look Once (YOLO) model. By integrating computer vision and deep learning techniques, the system achieves accurate and rapid face detection and recognition. Our proposed system utilizes OpenCV, a powerful computer vision library, to capture video streams from cameras. The YOLO model, a cutting-edge real-time object detection algorithm, is employed to identify and localize faces within the video frames. Thanks to YOLO's efficiency, the system ensures real-time processing, enabling seamless attendance recording. To enhance accuracy, the system employs a two-step approach consisting of face detection and face recognition. During the face detection phase, the YOLO model detects bounding boxes around faces. Subsequently, the system matches these detected faces against a pre-existing database of enrolled individuals using face recognition techniques. To improve performance, transfer learning techniques are applied to fine-tune the YOLO model on a diverse dataset containingvarious face images. This adaptation process ensures high precision and recall rates, even in challenging conditions such as varying lighting and occlusion. Experimental results demonstrate the effectiveness of the proposed real-time attendance system, achieving a high accuracy rate suitable for practical applications. Its real-time performance allows for seamless integration into existing attendance management workflows, resulting in time savings and improved administrative processes. The core intention of this paperwork is to develop a GUI page using Python programming, which will display the number of students who are present and number of students who are absent along with the number of students when section details are entered manually. In addition to this, pictures of the students who are present are displayed on the framework. It is called the real-time face detection which is very beneficial for managingacademic activity.

Cluster analysis and its application in teaching resources of university curriculum: A personalized method

AbstractThe current personalized recommendation methods for teaching resources of university courses suffer from poor recommendation effectiveness due to the absence of user tags. To address this issue, a new personalized recommendation method based on cluster analysis is proposed. The proposed method leverages web crawler technology to obtain user tags, followed by processing the tags to remove meaningless terms, normalize word forms, and perform data processing. The processed tags are used to calculate user interest preferences for each tag cluster generated by clustering. Based on this, a user interest model is built, and user similarity is calculated to determine the recommendation score of each resource. The recommended resources are then ranked according to their recommendation score and presented to the target user. Experimental results demonstrate that the proposed method achieves high accuracy, recall rate, and F1 value for personalized recommendation of teaching resources in colleges and universities. In comparison, the method proposed in this paper has a significantly shorter recommendation time of 10.65 s. Further, the proposed model not only takes less time but also has higher recommendation efficiency when compared with existing personalized recommendation methods.

DDoS Attack Detection in SDN using ML Techniques

Abstract: The increasing prevalence of DDoS attacks poses a serious threat to modern network infrastructures. SDN has been proposed as a promising solution for enhancing network security. However, detecting and mitigating DDoS attack in software definednetwork remains a challenging task. In this research paper, suggest an innovative approach in order to identify DDoS assaults in software-defined networks using (ML) techniques. Ourmethod entails gathering and analyzing network data. Traffic data using SDN controllers. We use variety of ML techniques analyze the traffic information to discover unexpected traffic patterns that might point to the presence of a DDoS attack. Random Forest, DecisionTree, K-Means clustering are among the algorithms used. We evaluate our approach using a real-world dataset and compare it to existing DDoS detection techniques in SDN. Our results show that our approach achieves high accuracy, precision, and recall rates indetecting DDoS attacks. We also demonstrate that our technique can detect either known and unexpected DDoS assaults with low false-positive rates. Overall, our study indicates thepotency of applying machine learning methods to SDN DDoS attack detection. Our methodoffers a promising remedy for boosting network security in contemporary infrastructures.

Augmented reality, deep learning and vision-language query system for construction worker safety

Low situational awareness contributes to safety incidents in construction. Existing Deep Learning (DL)-based applications lack the capability to provide context-specific and interactive feedback that is essential for workers to fully understand their surrounding environments. This paper proposes the Visual Construction Safety Query (VCSQ) system. The system encompasses real-time Image Captioning (IC), safety-centric Visual Question Answering (VQA), and keyword-based Image-Text Retrieval (ITR), integrated with head-mounted Augmented Reality (AR) devices. System validation includes benchmarks and real-world images. The ITR module posted high recall rates of 0.801 and 0.835 for Recall@5 and @10. The VQA module achieved an 89.7% accuracy rate, and the IC module had a SPICE score of 0.449. Feasibility tests and surveys confirmed the system's practical advantages in different construction scenarios. This study establishes an integration roadmap adaptable to future advancements in interactive DL and immersive AR.

MGS2AMR: a gene-centric mining of metagenomic sequencing data for pathogens and their antimicrobial resistance profile

BackgroundIdentification of pathogenic bacteria from clinical specimens and evaluating their antimicrobial resistance (AMR) are laborious tasks that involve in vitro cultivation, isolation, and susceptibility testing. Recently, a number of methods have been developed that use machine learning algorithms applied to the whole-genome sequencing data of isolates to approach this problem. However, making AMR assessments from more easily available metagenomic sequencing data remains a big challenge.ResultsWe present the Metagenomic Sequencing to Antimicrobial Resistance (MGS2AMR) pipeline, which detects antibiotic resistance genes (ARG) and their possible organism of origin within a sequenced metagenomics sample. This in silico method allows for the evaluation of bacterial AMR directly from clinical specimens, such as stool samples. We have developed two new algorithms to optimize and annotate the genomic assembly paths within the raw Graphical Fragment Assembly (GFA): the GFA Linear Optimal Path through seed segments (GLOPS) algorithm and the Adapted Dijkstra Algorithm for GFA (ADAG). These novel algorithms improve the sensitivity of ARG detection and aid in species annotation. Tests based on 1200 microbiome samples show a high ARG recall rate and correct assignment of the ARG origin. The MGS2AMR output can further be used in many downstream applications, such as evaluating AMR to specific antibiotics in samples from emerging intestinal infections. We demonstrate that the MGS2AMR-derived data is as informative for the entailing prediction models as the whole-genome sequencing (WGS) data. The performance of these models is on par with our previously published method (WGS2AMR), which is based on the sequencing data of bacterial isolates.ConclusionsMGS2AMR can provide researchers with valuable insights into the AMR content of microbiome environments and may potentially improve patient care by providing faster quantification of resistance against specific antibiotics, thereby reducing the use of broad-spectrum antibiotics. The presented pipeline also has potential applications in other metagenome analyses focused on the defined sets of genes.6MP9Kjm5NAthB1ioNbPxBHVideo

An Approach to Re-evaluate the Reference Cutoff of the Parameters of Newborn Screening: An Observational Study.

Background Unless a cutoff level of the parameters of newborn screening (NBS) is defined, a screening test's results would end in high recall rates and apprehensive parents. The study aimed to establish a cutoff level of the healthy term newborns. Materials and methods The study was a retrospective observational data analysis on a cohort of 1158 term newborns who underwent NBS in our institute. The percentile distribution of the NBS parameters was computed and the 99th percentile value was considered the new cutoff. For lower values, such as neonatal glucose 6-phosphate dehydrogenase (nG6PD) and neonatal biotinidase (nBIOT), low percentile values were regarded as new cutoff value. Results Neonatal thyroid stimulating hormone (nTSH), nG6PD, neonatal immunoreactive trypsinogen (nIRT), and nBIOT showed a wide variation in the distribution. Most newborns had neonatal galactose (nGAL), nIRT, and nBIOT values above the median. The 99th percentile value of nTSH was 14.5 mIU/L, and that of neonatal 17-hydroxyprogesterone (n17-OHP) was 43.7 nmol/L. The 1.0th percentile value for nG6PD was decreased to 2.18 IU/gHb. The new cutoff values for nBIOT, nIRT, neonatal phenylketonuria (nPKU) and nGAL were 48.59 U, 95.3 µg/L, 2.3 mg/dL and 15.9 mg/dL. The mean and median nTSH values did not significantly differ (p=0.99) in the first five days of birth. On the contrary, the study population depicted considerably raised levels of n17-OHP on day 3, followed by a sharp decrease (p=0.029). Similarly, nIRT displayed significant differences in the first five days (p=0.017). Conclusion Using the 99th percentile values of the NBS parameters as the new cutoff levels might be beneficial in terms of the recall rates and cost burden.

An Improved Object Detection Algorithm for Thyroid Nodule Ultrasound Image Based on Faster R-CNN

To propose an improved algorithm for thyroid nodule object detection based on Faster R-CNN so as to improve the detection precision of thyroid nodules in ultrasound images. The algorithm used ResNeSt50 combined with deformable convolution (DC) as the backbone network to improve the detection effect of irregularly shaped nodules. Feature pyramid networks (FPN) and Region of Interest (RoI) Align were introduced in the back of the trunk network. The former was used to reduce missed or mistaken detection of thyroid nodules, and the latter was used to improve the detection precision of small nodules. To improve the generalization ability of the model, parameters were updated during backpropagation with an optimizer improved by Sharpness-Aware Minimization (SAM). In this experiment, 6 261 thyroid ultrasound images from the Affiliated Hospital of Xuzhou Medical University and the First Hospital of Nanjing were used to compare and evaluate the effectiveness of the improved algorithm. According to the findings, the algorithm showed optimization effect to a certain degree, with the AP50 of the final test set being as high as 97.4% and AP@50:5:95 also showing a 10.0% improvement compared with the original model. Compared with both the original model and the existing models, the improved algorithm had higher detection precision and improved capacity to detect thyroid nodules with better accuracy and precision. In particular, the improved algorithm had a higher recall rate under the requirement of lower detection frame precision. The improved method proposed in the study is an effective object detection algorithm for thyroid nodules and can be used to detect thyroid nodules with accuracy and precision.

A Predictive Model for Early Detection of Oral Cancer via Adaboost Classification Technique

Oral cancer is a serious and potentially life-threatening disease that affects millions of people around the world. Early detection is critical for improving treatment outcomes and reducing mortality rates. This study aims to develop a predictive model for early detection of oral cancer using the AdaBoost classification technique. A dataset containing various risk factors associated with oral cancer was used to train and test the model. The results show that the AdaBoost algorithm was able to accurately classify oral cancer patients and non-cancer patients with high precision and recall rates. The developed predictive model could be used as a tool for early detection of oral cancer, thus improving patient outcomes and reducing mortality rates. The significance of oral cancer prediction and classification by improvised AdaBoost proposed technique produce potential to improve patient outcomes, facilitate early detection and treatment of oral cancer, and increase the efficiency and accuracy of the diagnostic process.

Nontargeted screening of aldehydes and ketones by chemical isotope labeling combined with ultra-high performance liquid chromatography-high resolution mass spectrometry followed by hybrid filtering of features

Aldehydes and ketones are important carbonyl compounds that are widely present in foodstuffs, biological organisms and human living environment. However, it is still challenging to comprehensively detect and capture them using liquid chromatography – mass spectrometry. In this work, a chemical isotope labeling (CIL) coupled with ultra-high performance liquid chromatography - high resolution mass spectrometry (UHPLC–HRMS) strategy was developed for the capture and detection of this class of compounds. 2,4-Dinitrophenylhydrazine (DNPH) and isotope-labeled DNPH (DNPH-d3) were utilized to selectively label the target analytes. To address the difficulties in processing UHPLC–HRMS data, a post-acquisition data processing method called MSFilter was proposed to facilitate the screening and identification aldehydes and ketones in complex matrices. The MSFilter consists of four independent filters, namely statistical characteristic-based filtering, mass defect filtering, CIL paired peaks filtering, and diagnostic fragmentation ion filtering. These filters can be used individually or in combination to eliminate unrelated interfering MS features and efficiently detect DNPH-labeled aldehydes and ketones. The results of a mixture containing 48 model compounds showed that although all individual filtering methods could significantly reduce more than 95% of the raw MS features with acceptable recall rates above 85%, but they had relatively high false positive ratios of over 90%. In comparison, the hybrid filtering method combining four filters is able to eliminate massive interfering features (> 99.5%) with a high recall rate of 81.25% and a much lower false positive ratio of 15.22%. By implementing the hybrid filtering method in MSFilter, a total of 154 features were identified as potential signals of CCs from the original 45,961 features of real tobacco samples, of which 70 were annotated. We believe that the proposed strategy is promising to analyze the potential CCs in complex samples by UHPLC–HRMS.

Why Start Now? Retrospective Study Evaluating Baseline Screening Mammography in Patients Age 60 and Older

PurposeExtensive data exist regarding the importance of baseline mammography and screening recommendations in the age range of 40-50 years old, however, less is known about women who start screening at age 60. The purpose of this retrospective study is to assess the characteristics and outcomes of women aged 60 years and older presenting for baseline mammographic screening. MethodsThis is an IRB-approved single institution retrospective review of data from patients aged 60+ receiving baseline screening mammograms between 2010 and 2022 was obtained. Information regarding patient demographics, breast density, and BI-RADS assessment was acquired from Cerner EHR. Of patients with a BI-RADS 0 assessment, imaging, and chart review was performed. Family history, gynecologic history, prior breast biopsy or surgery, and hormone use was reviewed. For those with a category 4 or 5 assessment after diagnostic work-up, biopsy outcomes were reported. Cancer detection rate (CDR), recall rate (RR), positive predictive value 1 (PPV1), PPV2, and PPV3 were calculated. ResultsData was analyzed from 1409 women over age 60 who underwent breast cancer screening. The recall rate was 29.3% (413/1409). The CDR, PPV1, PPV2, and PPV3 were calculated as 15/1000, 5.2% (21/405), 29.2% (21/72), and 31.8% (21/66), respectively. After work-up, 224 diagnostic patients had a 1-year follow-up and none were diagnosed with breast cancer. One (1.4%, 1/71) of the BI-RADS 3 lesions was malignant at 2-year follow-up. Of the patients recalled from screening, 29.6% had a family history of breast cancer, and the majority of both recalled and nonrecalled patients had Category B breast density. There was no statistically significant difference in breast density or race of patients recalled vs not recalled. 93.2% of recalled cases were given BI-RADS descriptors, with mass and focal asymmetry being the most common lesions, and 22.1% of recalled cases included more than one lesion. ConclusionInitiating screening mammography for patients over 60 years old may result in higher recall rates, but also leads to a high CDR of potentially clinically relevant invasive cancers. After a diagnostic work-up, BI-RADS 3 assessments are within standard guidelines. This study provides guidance for radiologists reading baseline mammograms and clinicians making screening recommendations in patients over age 60.

Smart Contract Vulnerability Detection Based on Deep Learning and Multimodal Decision Fusion.

With the rapid development and widespread application of blockchain technology in recent years, smart contracts running on blockchains often face security vulnerability problems, resulting in significant economic losses. Unlike traditional programs, smart contracts cannot be modified once deployed, and vulnerabilities cannot be remedied. Therefore, the vulnerability detection of smart contracts has become a research focus. Most existing vulnerability detection methods are based on rules defined by experts, which are inefficient and have poor scalability. Although there have been studies using machine learning methods to extract contract features for vulnerability detection, the features considered are singular, and it is impossible to fully utilize smart contract information. In order to overcome the limitations of existing methods, this paper proposes a smart contract vulnerability detection method based on deep learning and multimodal decision fusion. This method also considers the code semantics and control structure information of smart contracts. It integrates the source code, operation code, and control-flow modes through the multimodal decision fusion method. The deep learning method extracts five features used to represent contracts and achieves high accuracy and recall rates. The experimental results show that the detection accuracy of our method for arithmetic vulnerability, re-entrant vulnerability, transaction order dependence, and Ethernet locking vulnerability can reach 91.6%, 90.9%, 94.8%, and 89.5%, respectively, and the detected AUC values can reach 0.834, 0.852, 0.886, and 0.825, respectively. This shows that our method has a good vulnerability detection effect. Furthermore, ablation experiments show that the multimodal decision fusion method contributes significantly to the fusion of different modalities.

Mixing detection on Bitcoin transactions using statistical patterns

AbstractCryptocurrencies, particularly Bitcoin, have garnered attention for their potential in anonymous transactions. However, their anonymity has often been compromised by deanonymization attacks. To counter this, mixing services have been introduced. While they enhance privacy, they obscure fund traceability. This study seeks to demystify transactions linked to these services, shedding light on pathways of concealed and laundered money. We propose a method to identify and classify transactions and addresses of major mixing services in Bitcoin. Unlike previous research focusing on older techniques like CoinJoin, we emphasize modern mixing services. We gathered labelled data by transacting with three prominent mixers (MixTum, Blemder, and CryptoMixer) and identified recurring patterns. Using these patterns, an algorithm was created to pinpoint mixing transactions and distinguish mixer‐related addresses.The algorithm achieved a remarkable recall rate of 100%. Given the lack of clear ground truth and the vast number of unlabelled transactions, ensuring accuracy was a challenge. However, by analyzing a set of non‐mixing transactions with our model, it was confirmed that the high recall rate was not misleading. This work provides a significant advancement in monitoring mixing transactions, presenting a valuable tool against fraud and money laundering in cryptocurrency networks.

Dynamic Adaptation Attack Detection Model for a Distributed Multi-Access Edge Computing Smart City.

The internet of things (IoT) technology presents an intelligent way to improve our lives and contributes to many fields such as industry, communications, agriculture, etc. Unfortunately, IoT networks are exposed to many attacks that may destroy the entire network and consume network resources. This paper aims to propose intelligent process automation and an auto-configured intelligent automation detection model (IADM) to detect and prevent malicious network traffic and behaviors/events at distributed multi-access edge computing in an IoT-based smart city. The proposed model consists of two phases. The first phase relies on the intelligent process automation (IPA) technique and contains five modules named, specifically, dataset collection and pre-processing module, intelligent automation detection module, analysis module, detection rules and action module, and database module. In the first phase, each module composes an intelligent connecting module to give feedback reports about each module and send information to the next modules. Therefore, any change in each process can be easily detected and labeled as an intrusion. The intelligent connection module (ICM) may reduce the search time, increase the speed, and increase the security level. The second phase is the dynamic adaptation of the attack detection model based on reinforcement one-shot learning. The first phase is based on a multi-classification technique using Random Forest Trees (RFT), k-Nearest Neighbor (K-NN), J48, AdaBoost, and Bagging. The second phase can learn the new changed behaviors based on reinforced learning to detect zero-day attacks and malicious events in IoT-based smart cities. The experiments are implemented using a UNSW-NB 15 dataset. The proposed model achieves high accuracy rates using RFT, K-NN, and AdaBoost of approximately 98.8%. It is noted that the accuracy rate of the J48 classifier achieves 85.51%, which is lower than the others. Subsequently, the accuracy rates of AdaBoost and Bagging based on J48 are 98.9% and 91.41%, respectively. Additionally, the error rates of RFT, K-NN, and AdaBoost are very low. Similarly, the proposed model achieves high precision, recall, and F1-measure high rates using RFT, K-NN, AdaBoost, and Bagging. The second phase depends on creating an auto-adaptive model through the dynamic adaptation of the attack detection model based on reinforcement one-shot learning using a small number of instances to conserve the memory of any smart device in an IoT network. The proposed auto-adaptive model may reduce false rates of reporting by the intrusion detection system (IDS). It can detect any change in the behaviors of smart devices quickly and easily. The IADM can improve the performance rates for IDS by maintaining the memory consumption, time consumption, and speed of the detection process.

Fire Detection of yolov8 Model based on Integrated SE Attention Mechanism

Our proposed method utilizes YOLOv8 and SE attention mechanism for detecting fires, which is crucial for early detection and prevention. Our method balances accuracy and real-time performance while detecting fires in various scenarios. The proposed method achieves an average mAP0.5 value of 0.730, with an improvement from the original model's mAP0.5 value of 0.707 after incorporating the SE attention mechanism. We evaluated our model on a benchmark dataset and demonstrated its effectiveness in accurately detecting and localizing fires with high precision and recall rates. Experimental results confirm the effectiveness of our proposed method in accurately detecting and localizing fires, demonstrating its potential for wide application and promotion in the fire safety industry.

Real-Time Intelligent Detection System for Illegal Wearing of On-Site Power Construction Worker Based on Edge-YOLO and Low-Cost Edge Devices

Ensuring personal safety and preventing accidents are critical aspects of power construction safety supervision. However, current monitoring methods are inefficient and unreliable as most of them rely on manual monitoring and transmission, which results in slow detection and delayed warnings regarding violations. To overcome these challenges, we propose an intelligent detection system that can accurately identify instances of illegal wearing of power construction workers in real-time. Firstly, we integrated the squeeze-and-excitation (SE) module into our convolutional neural network to enhance detection accuracy. This module effectively prioritizes informative features while suppressing less relevant ones, resulting in improved overall performance. Secondly, we present an embedded real-time detection system that utilizes Jetson Xavier NX and Edge-YOLO. This system promptly detects and alerts power construction workers of instances of illegal wearing behavior. To ensure a lightweight implementation, we design appropriate detection heads based on target size and distribution, reducing model parameters while enhancing detection speed and minimizing accuracy loss. Additionally, we employed data augmentation to enhance the system’s robustness. Our experimental results demonstrate that our improved Edge-YOLO model achieves high detection precision and recall rates of 0.964 and 0.966, respectively, with a frame rate of 35.36 frames per second when deployed on Jetson Xavier NX. Therefore, Edge-YOLO proves to be an ideal choice for intelligent real-time detection systems, providing superior accuracy and speed performance compared to the original YOLOv5s model and other models in the YOLO series for safety monitoring at construction sites.