A Dual Transformer-Based Deep Learning Model for Passenger Anomaly Behavior Detection in Elevator Cabs DOI Creative Commons

Yijin Ji,

Haoxiang Sun, Benlian Xu

et al.

International Journal of Swarm Intelligence Research, Journal Year: 2024, Volume and Issue: 15(1), P. 1 - 14

Published: Nov. 22, 2024

Effective detection of abnormal behaviors within elevator cabins is critical to ensure safety. While existing deep learning based anomaly methods mainly focus on convolutional neural networks for spatial feature extraction and recurrent temporal learning, recent advancements in the Transformer architecture have demonstrated its power time series predictions, extended capabilities vision tasks. In this study, we present a duel transformer-based framework that can proficiently detect falling fighting events cabs. The proposed solution leverages transformer (ViT) extract frame-level features, followed by identify abnormalities surveillance videos. A comprehensive comparison between method other traditional network variants carried out validate effectiveness method.

Language: Английский

Recurrent Neural Networks: A Comprehensive Review of Architectures, Variants, and Applications DOI Creative Commons
Ibomoiye Domor Mienye, Theo G. Swart, George Obaido

et al.

Information, Journal Year: 2024, Volume and Issue: 15(9), P. 517 - 517

Published: Aug. 25, 2024

Recurrent neural networks (RNNs) have significantly advanced the field of machine learning (ML) by enabling effective processing sequential data. This paper provides a comprehensive review RNNs and their applications, highlighting advancements in architectures, such as long short-term memory (LSTM) networks, gated recurrent units (GRUs), bidirectional LSTM (BiLSTM), echo state (ESNs), peephole LSTM, stacked LSTM. The study examines application to different domains, including natural language (NLP), speech recognition, time series forecasting, autonomous vehicles, anomaly detection. Additionally, discusses recent innovations, integration attention mechanisms development hybrid models that combine with convolutional (CNNs) transformer architectures. aims provide ML researchers practitioners overview current future directions RNN research.

Language: Английский

Citations

46
Recurrent Neural Networks: A Comprehensive Review of Architectures, Variants, and Applications DOI Open Access
Ibomoiye Domor Mienye, Theo G. Swart, George Obaido

et al.

Published: Aug. 12, 2024

Recurrent Neural Networks (RNNs) have significantly advanced the field of machine learning by enabling effective processing sequential data. This paper provides a comprehensive review RNNs and their applications, highlighting advancements in architectures such as Long Short-Term Memory (LSTM) networks, Gated Units (GRUs), Bidirectional LSTM (BiLSTM), stacked LSTM. The study examines application different domains, including natural language (NLP), speech recognition, financial time series forecasting, bioinformatics, autonomous vehicles, anomaly detection. Additionally, discusses recent innovations, integration attention mechanisms development hybrid models that combine with convolutional neural networks (CNNs) transformer architectures. aims to provide researchers practitioners overview current state future directions RNN research.

Language: Английский

Citations

25
The explainable potential of coupling hybridized metaheuristics, XGBoost, and SHAP in revealing toluene behavior in the atmosphere DOI
Nebojša Bačanin, Mirjana Perišić, Gordana Jovanović

et al.

The Science of The Total Environment, Journal Year: 2024, Volume and Issue: 929, P. 172195 - 172195

Published: April 15, 2024

Language: Английский

Citations

18
Cardiovascular care with digital twin technology in the era of generative artificial intelligence DOI
Phyllis Thangaraj, S. Benson, Evangelos K. Oikonomou

et al.

European Heart Journal, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 26, 2024

Abstract Digital twins, which are in silico replications of an individual and its environment, have advanced clinical decision-making prognostication cardiovascular medicine. The technology enables personalized simulations scenarios, prediction disease risk, strategies for trial augmentation. Current applications digital twins integrated multi-modal data into mechanistic statistical models to build physiologically accurate cardiac replicas enhance phenotyping, enrich diagnostic workflows, optimize procedural planning. twin is rapidly evolving the setting newly available modalities advances generative artificial intelligence, enabling dynamic comprehensive unique individual. These fuse physiologic, environmental, healthcare machine learning real-time patient predictions that can model interactions with environment accelerate care. This review summarizes medicine their potential future by incorporating new modalities. It examines technical deep intelligence broaden scope predictive power twins. Finally, it highlights societal challenges as well ethical considerations essential realizing vision cardiology

Language: Английский

Citations

12
Construction and verification of a nomogram model for the risk of death in sepsis patients DOI Creative Commons
Yanjie Yang,

Huiling Zhao,

Ling Ge

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Feb. 11, 2025

At present, there is insufficient evidence to evaluate the prognosis of patients with sepsis. This study anazed clinical data 822 sepsis in ICU a tertiary Grade A hospital construct and validate nomogram model for predicting 28-day mortality risk patients. The was constructed using multivariate logistic regression analysis screen independent factors affecting prognosis, prediction built based on these factors. performance evaluated Hosmer–Lemeshow test, receiver operating characteristic curve (ROC), calibration plot, decision (DCA). Multivariate identified five patients: Age, SOFA score, CRP, Mechanical ventilation, use Vasoactive drugs. odds ratios (OR) 95% confidence intervals (95% CI) were 1.037 (1.024–1.050), 1.093 (1.044–1.145), 1.034 (1.026–1.042), 1.967 (1.176–3.328), 2.515 (1.611–3.941), respectively, all P-values < 0.05. Based factors, constructed, area under ROC (AUC) training set external validation being 0.849 CI 0.818–0.880) 0.837 0.887–0.886), respectively. Both DCA plot confirmed that has good efficacy. established this excellent predictive ability, which can help clinicians identify high-risk early provide guidance decision-making.

Language: Английский

Citations

1
Association between the non-high-density lipoprotein cholesterol to high-density lipoprotein cholesterol ratio and mortality among hypertension patients DOI Creative Commons
Xiaozhou Su, Hengyi Rao, Chunli Zhao

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Feb. 19, 2025

The non-high-density lipoprotein cholesterol to high-density ratio (NHHR) reflects the balance between pro- and anti-atherogenic lipoproteins. This study aims explore relationship NHHR mortality among hypertension patients. Data from 17,075 hypertensive adults in National Health Nutrition Examination Survey (NHANES) were analyzed. Multivariate Cox regression restricted cubic splines used assess correlation mortality. A segmented model evaluated threshold effects, sensitivity analyses confirmed result robustness. Machine learning algorithms establish a prediction model. Over median follow-up of 84 months, 3625 deaths occurred. U-shaped association was observed both all-cause cardiovascular mortality, with values at 2.32 2.65. Below these thresholds, negatively associated while above thresholds positively associated. classified as an important variable model, random survival forest (rsf) algorithm showing superior performance. identified patients, points 2.65, indicating that is potential predictor patients hypertension.

Language: Английский

Citations

1
Respiratory Condition Detection Using Audio Analysis and Convolutional Neural Networks Optimized by Modified Metaheuristics DOI Creative Commons
Nebojša Bačanin, Luka Jovanovic, Ruxandra Stoean

et al.

Axioms, Journal Year: 2024, Volume and Issue: 13(5), P. 335 - 335

Published: May 18, 2024

Respiratory conditions have been a focal point in recent medical studies. Early detection and timely treatment are crucial factors improving patient outcomes for any condition. Traditionally, doctors diagnose respiratory through an investigation process that involves listening to the patient’s lungs. This study explores potential of combining audio analysis with convolutional neural networks detect patients. Given significant impact proper hyperparameter selection on network performance, contemporary optimizers employed enhance efficiency. Moreover, modified algorithm is introduced tailored specific demands this study. The proposed approach validated using real-world dataset has demonstrated promising results. Two experiments conducted: first tasked models condition when observing mel spectrograms patients’ breathing patterns, while second experiment considered same data format multiclass classification. Contemporary optimize architecture training parameters both cases. Under identical test conditions, best optimized by metaheuristic, accuracy 0.93 detection, slightly reduced 0.75 identification.

Language: Английский

Citations

7
Computer-Vision Unmanned Aerial Vehicle Detection System Using YOLOv8 Architectures DOI Open Access
Aleksandar Petrović, Nebojša Bačanin, Luka Jovanovic

et al.

International Journal of Robotics and Automation Technology, Journal Year: 2024, Volume and Issue: 11, P. 1 - 12

Published: May 22, 2024

Abstract: This work aims to test the performance of you only look once version 8 (YOLOv8) model for problem drone detection. Drones are very slightly regulated and standards need be established. With a robust system detecting drones possibilities regulating their usage becoming realistic. Five different sizes were tested determine best architecture size this problem. The results indicate high across all models that each is used specific case. Smaller suited lightweight approaches where some false identification tolerable, while largest with stationary systems require precision.

Language: Английский

Citations

6
Emerging analytical approaches for personalized medicine using machine learning in pediatric and congenital heart disease DOI
Bhargava Chinni, Cedric Manlhiot

Canadian Journal of Cardiology, Journal Year: 2024, Volume and Issue: 40(10), P. 1880 - 1896

Published: Aug. 7, 2024

Language: Английский

Citations

5
Optimizing renewable energy utilization through production and demand forecasting using optimized recurrent networks DOI
Nebojša Bačanin, M. Pavlov, Stanislava Kozakijevic

et al.

Elsevier eBooks, Journal Year: 2025, Volume and Issue: unknown, P. 203 - 225

Published: Jan. 1, 2025

Language: Английский

Citations

0