Performance Of Detection System Research Articles

An Efficient Forest Fire Detection Algorithm Using Improved YOLOv5

Forest fires result in severe disaster, causing significant ecological damage and substantial economic losses. Flames and smoke represent the predominant characteristics of forest fires. However, these flames and smoke often exhibit irregular shapes, rendering them susceptible to erroneous positive or negative identifications, consequently compromising the overall performance of detection systems. To enhance the average precision and recall rates of detection, this paper introduces an enhanced iteration of the You Only Look Once version 5 (YOLOv5) algorithm. This advanced algorithm aims to achieve more effective fire detection. First, we use Switchable Atrous Convolution (SAC) in the backbone network of the traditional YOLOv5 to enhance the capture of a larger receptive field. Then, we introduce Polarized Self-Attention (PSA) to improve the modeling of long-range dependencies. Finally, we incorporate Soft Non-Maximum Suppression (Soft-NMS) to address issues related to missed detections and repeated detections of flames and smoke by the algorithm. Among the plethora of models explored, our proposed algorithm achieves a 2.0% improvement in mean Average Precision@0.5 (mAP50) and a 3.1% enhancement in Recall when compared with the YOLOv5 algorithm. The integration of SAC, PSA, and Soft-NMS significantly enhances the precision and efficiency of the detection algorithm. Moreover, the comprehensive algorithm proposed here can identify and detect key changes in various monitoring scenarios.

MTSA-Net: A multiscale time self-attention network for ship radiated self-noise reduction

MTSA-Net: A multiscale time self-attention network for ship radiated self-noise reduction

Spatial-temporal knowledge distillation for lightweight network traffic anomaly detection

Spatial-temporal knowledge distillation for lightweight network traffic anomaly detection

Experimental study on phase compensation method based on autocorrelation in laser heterodyne detection

Experimental study on phase compensation method based on autocorrelation in laser heterodyne detection

The synchronous detection technique for the accurate monitoring of high-energy pulsed X-rays

The synchronous detection technique for the accurate monitoring of high-energy pulsed X-rays

A CrowdSensing-based approach for proximity detection in indoor museums with Bluetooth tags

A CrowdSensing-based approach for proximity detection in indoor museums with Bluetooth tags

Performance Metrics of an Intrusion Detection System Through Window-Based Deep Learning Models

Intrusion and prevention technologies perform reliably in harsh conditions by fortifying many of the world's highest security sites with few defects in high performance. This paper aims to contribute by designing an intrusion/preventive system using a window-based convolutional neural network (CNN), an integrated recurrent neural network (RNN), and autoencoders (AutoE) to detect and test the performance of the intrusion detection system. The data packets were converted to images where the pixels were used as input. The CNN architecture shows a three-layer model with high predictive performance. The result shows high performance on CNN as compared to both RNN and AutoE; CNN seems to resist overfitting more than the rest of the models. The future perspective would be to test the model on other standard methods such as support vector machine (SVM) and dynamic control systems.

Background studies and normalization of signal events in the Mu2e experiment

The Mu2e experiment is currently being constructed at Fermilab to search for the neutrino-less conversion of negative muons into electrons in the field of an aluminum nucleus. The experiment aims at a sensitivity of four orders of magnitude higher than previous related experiments, which implies highly demanding accuracy requirements both in the design and during the operation. To achieve such a goal, two important tasks should be accomplished. First, it is essential to estimate precisely the particle yields and all the backgrounds that could mimic the monoenergetic conversion electron signal. Second, it is necessary to normalize the signal events accurately. The normalization of the signal events is planned to be done using a detector system made of an HPGe detector and a Lanthanum Bromide detector, which will measure the rate of muons stopped on the aluminum target by looking at the emitted characteristic X-and γ-rays of energies up to 1809 keV. Therefore, it is essential to evaluate the detector system's performance before the start of the actual experiment. In this study, the first task was addressed by an extensive campaign of Monte Carlo simulations to investigate the relevant parameters and their impact on the experiment's sensitivity. The second task was handled by taking advantage of the Helmholtz-Zentrum Dresden-Rossendorf (HZDR) pulsed Bremsstrahlung photon beam at the ELBE facility. The detector system was tested at the ELBE facility under timing and background conditions similar to the ones expected at the Mu2e experiment. The study presents and discusses the simulation results and the detector system testing campaign.

Distributed adaptive event-triggered fault detection filter of positive semi-Markovian jump systems

Distributed adaptive event-triggered fault detection filter of positive semi-Markovian jump systems

Ensemble Machine Learning to Detect Sarcasm in English on Twitter Social Media

Detecting sarcasm in English tweets on social media platforms like Twitter is a complex task due to its subtle and ambiguous nature. This study explores the use of ensemble machine learning techniques, including Logistic Regression, Naive Bayes, Decision Tree, and Support Vector Machine (SVM), to effectively identify sarcasm. A dataset containing sarcastic and non-sarcastic English tweets was collected and preprocessed. Features representing lexical, syntactic, and semantic information were extracted to train and evaluate the ensemble models. The Support Vector Machine method demonstrated the highest performance among the techniques employed, achieving an accuracy of 80% and an F1-score of 80% for sarcasm detection. This highlights the efficacy of Support Vector Machines in capturing complex patterns and differentiating between sarcastic and non-sarcastic tweets. By leveraging the strengths of multiple machine learning algorithms, the ensemble approach enhances the overall performance of the sarcasm detection system. It provides a more robust and accurate detection of sarcasm, thereby improving the understanding of user sentiments and opinions in online conversations. This research contributes to sentiment analysis and natural language processing, offering valuable insights into sarcasm detection in social media. The findings have practical implications for interpreting user-generated content on platforms like Twitter, enabling a better understanding of user sentiments and facilitating more meaningful interactions.

Assessing the alerting capabilities of the Earthquake Network early warning system in Haiti with Monte Carlo simulations

Smartphone-based earthquake early warning systems implemented by citizen science initiatives are characterised by significant variability in their smartphone network geometry. This has a direct impact on the earthquake detection capability and system performance. Here, a Monte Carlo-based simulation framework is implemented to infer relevant earthquake detection quantities such as detection distance from the epicentre, detection delay, and warning time for people exposed to high ground shaking levels. The framework is applied to Haiti, which has experienced deadly earthquakes in recent decades, and to the Earthquake Network citizen science initiative, which is popular in the country. It is found that warning times of up to 12 s for people exposed to intensities between 7.5 and 8.5 on the modified Mercalli scale are possible starting from a relatively low involvement of citizens in the initiative (i.e., from 1 Haitian in 10,000).

The diagnostic performance and clinical value of deep learning-based nodule detection system concerning influence of location of pulmonary nodule

BackgroundThe deep learning-based nodule detection (DLD) system improves nodule detection performance of observers on chest radiographs (CXRs). However, its performance in different pulmonary nodule (PN) locations remains unknown.MethodsWe divided the CXR intrathoracic region into non-danger zone (NDZ) and danger zone (DZ). The DZ included the lung apices, paramediastinal areas, and retrodiaphragmatic areas, where nodules could be missed. We used a dataset of 300 CXRs (100 normal and 200 abnormal images with 216 PNs [107 NDZ and 109 DZ nodules]). Eight observers (two thoracic radiologists [TRs], two non-thoracic radiologists [NTRs], and four radiology residents [RRs]) interpreted each radiograph with and without the DLD system. The metric of lesion localization fraction (LLF; the number of correctly localized lesions divided by the total number of true lesions) was used to evaluate the diagnostic performance according to the nodule location.ResultsThe DLD system demonstrated a lower LLF for the detection of DZ nodules (64.2) than that of NDZ nodules (83.2, p = 0.008). For DZ nodule detection, the LLF of the DLD system (64.2) was lower than that of TRs (81.7, p < 0.001), which was comparable to that of NTRs (56.4, p = 0.531) and RRs (56.7, p = 0.459). Nonetheless, the LLF of RRs significantly improved from 56.7 to 65.6 using the DLD system (p = 0.021) for DZ nodule detection.ConclusionThe performance of the DLD system was lower in the detection of DZ nodules compared to that of NDZ nodules. Nonetheless, RR performance in detecting DZ nodules improved upon using the DLD system.Critical relevance statementDespite the deep learning-based nodule detection system’s limitations in detecting danger zone nodules, it proves beneficial for less-experienced observers by providing valuable assistance in identifying these nodules, thereby advancing nodule detection in clinical practice.Key points• The deep learning-based nodule detection (DLD) system can improve the diagnostic performance of observers in nodule detection.• The DLD system shows poor diagnostic performance in detecting danger zone nodules.• For less-experienced observers, the DLD system is helpful in detecting danger zone nodules.Graphical

Pipelined deep learning architecture for the detection of Alzheimer’s disease

Pipelined deep learning architecture for the detection of Alzheimer’s disease

A differential evolution-based algorithm with maturity extension for feature selection in intrusion detection system

A differential evolution-based algorithm with maturity extension for feature selection in intrusion detection system

OPTIMASI KLASIFIKASI DETEKSI SENJATA API PADA ANIMASI BLACK LAGOON DENGAN ALGORITMA DEEPSORT DAN KALMAN FILTER

As technology develops, especially in the field of artificial intelligence, many new systems and programs are created with artificial intelligence, one of which is the field of computer vision, namely the detection of objects ranging from transportation, humans, masks, animals, and even firearms detection systems. However, this system still has deficiencies in the level of accuracy when complex conditions such as lighting and others. Currently firearms detection systems use datasets in the form of photos of real weapons, rarely or even no one uses other datasets such as games or animations to determine the performance of the detection system. This study aims to determine the performance of the firearms detection and classification system by experimenting with objects from an animation with a public dataset from the website www.imfdb.org/Black_Lagoon as well as several scenes from other animations and also an animation entitled Black Lagoon, then optimization is carried out.with the DeepSORT algorithm and Kalman Filter assisted by the YOLOv3 detection system, several percentage values were obtained such as a mAP value of 82.13%, 81.19% precision, and 78.53% recall in the training dataset using 100 iterations of epochs, and in the training model got a loss value of 8.27% and 6.67% for the val_loss value using 50 iterations of epochs.

Factors influencing the performance of optical heterodyne detection system

Factors influencing the performance of optical heterodyne detection system

Application of GA-WELM Model Based on Stratified Cross-Validation in Intrusion Detection

Aiming at the problem of poor detection performance under the environment of imbalanced type distribution, an intrusion detection model of genetic algorithm to optimize weighted extreme learning machine based on stratified cross-validation (SCV-GA-WELM) is proposed. In order to solve the problem of imbalanced data types in cross-validation subsets, SCV is used to ensure that the data distribution in all subsets is consistent, thus avoiding model over-fitting. The traditional fitness function cannot solve the problem of small sample classification well. By designing a weighted fitness function and giving high weight to small sample data, the performance of the model can be effectively improved in the environment of imbalanced type distribution. The experimental results show that this model is superior to other intrusion detection models in recall and McNemar hypothesis test. In addition, the recall of the model for small sample data is higher, reaching 91.5% and 95.1%, respectively. This shows that it can effectively detect intrusions in an environment with imbalanced type distribution. Therefore, the model has practical application value in the field of intrusion detection, and can be used to improve the performance of intrusion detection systems in the actual environment. This method has a wide application prospect, such as network security, industrial control system, and power system.

MAFSIDS: a reinforcement learning-based intrusion detection model for multi-agent feature selection networks

Large unbalanced datasets pose challenges for machine learning models, as redundant and irrelevant features can hinder their effectiveness. Furthermore, the performance of intrusion detection systems (IDS) can be further degraded by the emergence of new network attack types. To address these issues, we propose MAFSIDS (Multi-Agent Feature Selection Intrusion Detection System), a DQL (Deep Q-Learning) based IDS.MAFSIDS comprises a feature self-selection algorithm and a DRL (Deep Reinforcement Learning) attack detection module. The feature self-selection algorithm leverages a multi-agent reinforcement learning framework, which redefines the feature selection problem by converting the traditional {2}^{N} feature selection space into N agent representations. This approach reduces model complexity and enhances the search strategy for feature selection. To ensure accurate feature representation and expedite the feature selection process, we have also developed a GCN (Graph Convolutional Network) method that extracts deeper features from the data. The DRL attack detection module utilizes the Mini-Batchs technique to encode the data, allowing reinforcement learning to be applied in a supervised learning context. This integration improves accuracy. Additionally, the policy network in this module is designed to be minimalist, enhancing model efficiency. To evaluate the performance of our model, we conducted comprehensive simulation experiments using Python. We tested the model using the CSE-CIC-IDS2018 and NSL-KDD datasets, achieving impressive accuracy rates of 96.8% and 99.1%, as well as F1-Scores of 96.3% and 99.1%, respectively. The selected feature subset successfully eliminates approximately 80% of redundant features compared to the original feature set. Furthermore, we compared our proposed model with other popular machine-learning models.

Improving Anomalous Sound Detection by Distance Matrix-Based Visualization of Measurement Flaws

Although recent DNN-based methods have improved the performance of anomalous sound detection systems, it is still difficult to deploy a system in a real environment without performance degradation. This is often due to measurement flaws such as sensor variability, poor setup, or environmental noise. Since such adverse effects are difficult to model by machine learning, a practical approach to this issue is for humans to identify such flaws and correct them. To this end, we propose a method to visualize measurement flaws as a heatmap based on the distance matrix of the samples in the dataset. This method is designed to find unexpected flaws in the measurement process. Using this method, we were able to identify measurement flaws of anomalous sound detection systems in real production lines. The robustness of anomalous sound detection can be improved by correcting the flaws found by our method.

Performance of coherent FSO systems with adaptive M-PSK modulation

Performance of coherent FSO systems with adaptive M-PSK modulation