Article
Mathematics, Interdisciplinary Applications
Guancen Lin, Aijing Lin
Summary: This paper proposes an innovative multiscale sample entropy based on the horizontal visibility graph to measure the complexity of time series, as well as an improved multiscale cross-sample entropy method to measure the synchronization between two time series. By applying these methods to feature extraction, classification, and sleep stage division of EEG signals, it is possible to effectively monitor human health and assess physical status.
CHAOS SOLITONS & FRACTALS
(2022)
Article
Engineering, Electrical & Electronic
Kanchan Sharma, Ramesh Kumar Sunkaria
Summary: Cardiac arrhythmia can be assessed using cardiac rate variability. This study proposes Multiscale Cross Sample Entropy (MCSEn) to quantify complexity of arrhythmia at different scales, but it fails to provide complexity measures for smaller scale factors. To address this issue, a new algorithm called Multiscale E-metric Cross Sample Entropy (MECSEn) is proposed, which evaluates the complexity between arrhythmia subjects (atrial fibrillation and congestive heart failure) and healthy subjects at multiple scales using coarse-grained process. The study finds that subjects with atrial fibrillation behave as white noise, while subjects with congestive heart failure behave as pink noise. The t test shows significant differences between MCSEn, MECSEn, and multiscale sample entropy algorithm (MSEn) in evaluating the complexity of healthy and arrhythmia subjects.
SIGNAL IMAGE AND VIDEO PROCESSING
(2023)
Article
Automation & Control Systems
Jinde Zheng, Haiyang Pan, Jinyu Tong, Qingyun Liu
Summary: Extracting failure-related information from vibration signals is crucial for vibration-based fault detection in rolling bearings. This article proposes a new nonlinear dynamic parameter to enhance the measurement of data complexity and compares it with existing algorithms. Furthermore, a novel fault diagnosis approach is introduced, which achieves the highest identifying rate and the best performance among the comparative approaches.
Article
Genetics & Heredity
Hengyi Zhang
Summary: A novel feature selection algorithm based on approximate conditional entropy using fuzzy information granule is proposed and its correctness is proved by the monotonicity of entropy. Experimental results demonstrate that the algorithm significantly reduces the dimension of gene datasets and outperforms five state-of-the-art algorithms in terms of classification accuracy.
FRONTIERS IN GENETICS
(2021)
Article
Physics, Multidisciplinary
Hongjian Xiao, Danilo P. Mandic
Summary: Entropy-based methods are important in quantifying the complexity of real-world systems. We propose a new method called veMSE, which can robustly evaluate structural complexity even with shorter data. veMSE also exhibits desirable properties such as robustness to embedding dimension and noise resilience.
Article
Physics, Multidisciplinary
Ryan Furlong, Mirvana Hilal, Vincent O'Brien, Anne Humeau-Heurtier
Summary: The study explores the impact of parameters on image classification using two-dimensional fuzzy entropy and dispersion entropy. Results show that specific parameter choices are crucial for accurate image classification. By utilizing certain parameters, MFuzzyEn2D and MDispEn2D achieve outstanding performance in image classification on various datasets, with MFuzzyEn2D generally outperforming MDispEn2D.
Article
Automation & Control Systems
Renjie Zhou, Xiao Wang, Jian Wan, Naixue Xiong
Summary: Sample entropy technologies are widely used in diagnosing faults in industrial systems, but they have limitations in distance measurement and similarity evaluation. A new method called Euclidean distance based multiscale fuzzy entropy is proposed for better accuracy in measuring signal complexity and detecting bearing faults.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2021)
Article
Computer Science, Information Systems
Yan Li, Juan Liu, Chi Tang, Wei Han, Shengyi Zhou, Siqi Yang, Long He, Da Jing, Erping Luo, Kangning Xie
Summary: Research has found that multi-scale entropy curves of sleep and wakefulness EEG signals intersect, with the origin of this phenomenon remaining unexplored. Simulation results show that the crossover phenomenon is prevalent in the MSE analysis, with the crossover point moving towards larger scale factors with increasing sampling rate. Real-world EEG results indicate a higher MSE curve for wakefulness compared to sleep, showing no interception.
Article
Physics, Multidisciplinary
Dragana Bajic, Nina Japundzic-Zigon
Summary: This paper analyzes the inconsistencies of approximate and sample entropies, identifying the major problem as a coarse quantization of matching probabilities. It shows that errors in entropy arise due to the accumulation of errors in approximate entropy, while errors in sample entropy cancel each other. The length and distribution of time series also affect the errors, while self-matching has a marginal influence and interpolation introduces significant errors.
Article
Mathematics, Applied
Yi Yin, Xi Wang, Wenjing Wang, Qiang Li, Pengjian Shang
Summary: In this paper, the VGMCSE method is proposed to analyze the irreversibility of multiscale time series and gain more information from time irreversibility analysis. Numerical simulations and empirical analysis on traffic time series are conducted to validate the proposed method. The results reveal that ARFIMA and FGN series are time reversible, while Logistic map and Henon map are time irreversible. Additionally, the traffic time series recorded by detectors are classified into different groups based on their positions, indicating the presence of similar time irreversible behavior among closer-detector recordings.
COMMUNICATIONS IN NONLINEAR SCIENCE AND NUMERICAL SIMULATION
(2023)
Article
Instruments & Instrumentation
Tzu-Kang Lin, Ting-Hsuan Huang
Summary: This study combined three-dimensional printing and composite multiscale cross-sample entropy in structural health monitoring, developed a quantification criterion for single-story structural damage index, and validated the feasibility and effectiveness of the method through experiments.
SMART MATERIALS AND STRUCTURES
(2021)
Article
Engineering, Biomedical
Alessandro Mengarelli, Andrea Tigrini, Federica Verdini, Rosa Anna Rabini, Sandro Fioretti
Summary: This study investigated the postural dynamics and complexity in diabetic patients and found that diabetic neuropathy affects postural steadiness and results in a loss of complexity, particularly in the medial-lateral direction. The use of multiscale fuzzy entropy analysis provides insights into balance control mechanisms for diabetic patients.
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
(2023)
Article
Engineering, Mechanical
Jiayu Huang, Jianhua Liu, Hao Gong, Xinjian Deng
Summary: This study is the first attempt to conduct multimodal loosening detection exploiting ultrasonic and audio response signals simultaneously. A novel loosening detection method, utilizing the complementarity of multimodal signals, is proposed and proved to have excellent detection performances in the applications of two different types of threaded fasteners. The proposed method outperforms other loosening detection methods and MCFE shows great advantages in extracting representative loosening features.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2023)
Article
Computer Science, Artificial Intelligence
Zhihong Wang, Hongmei Chen, Zhong Yuan, Jihong Wan, Tianrui Li
Summary: This paper introduces a feature selection method based on multiscale fuzzy entropy, which improves the effectiveness of feature selection by fusing granule information at different scales.
IEEE TRANSACTIONS ON FUZZY SYSTEMS
(2023)
Article
Mathematics, Interdisciplinary Applications
Bhabesh Deka, Dipen Deka
Summary: The assessment of dynamical complexity is vital in various fields such as medical diagnostics, mechanical system fault analysis, and astrophysics. Traditional entropy measures are limited by short data length and sensitive predetermined parameters. Distribution entropy (DistEn) is a robust complexity estimator, but it fails to distinguish noise and chaotic signals while underestimating their complexity at higher scales. To overcome these limitations, an improved distribution entropy (ImDistEn) is proposed, which utilizes embedded vectors' orientation, ordinality, and l(1)-norm distance information. Simulation results demonstrate that ImDistEn can accurately assess the complexity of different types of signals.
CHAOS SOLITONS & FRACTALS
(2022)
Review
Physiology
Antoine Jamin, Pierre Abraham, Anne Humeau-Heurtier
Summary: The paper details the use of supervised machine learning algorithms for predictive data analytics problems in medicine. Data in the medical field can be categorized into medical images and other data. The four supervised machine learning approaches discussed are information-based, similarity-based, probability-based and error-based approaches.
CLINICAL PHYSIOLOGY AND FUNCTIONAL IMAGING
(2021)
Article
Computer Science, Information Systems
Andreia Sofia F. Gaudencio, Pedro G. Vaz, Mirvana Hilal, Joao M. Cardoso, Guillaume Mahe, Mathieu Lederlin, Anne Humeau-Heurtier
Summary: The study proposed a three-dimensional multiscale fuzzy entropy (MFE3D) algorithm to identify Idiopathic Pulmonary Fibrosis (IPF) patients from their computed tomography (CT) volumetric data. The algorithm was validated using synthetic noises, MIX(p) processes-based volumes, and texture-based volumes, showing consistency with one and two-dimensional versions. MFE3D was applied to CT scans of healthy subjects and IPF patients, revealing statistical differences in entropy values between the groups, indicating potential for IPF identification in CT scans.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2021)
Article
Engineering, Biomedical
Andreia S. Gaudencio, Pedro G. Vaz, Mirvana Hilal, Guillaume Mahe, Mathieu Lederlin, Anne Humeau-Heurtier, Joao M. Cardoso
Summary: This study aimed to quantify lung modifications in COVID-19 and IPF patients using a three-dimensional multiscale fuzzy entropy algorithm, statistically showing significant differences in 9 scale factors, with high classification accuracy and sensitivity.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2021)
Article
Physics, Multidisciplinary
Ryan Furlong, Mirvana Hilal, Vincent O'Brien, Anne Humeau-Heurtier
Summary: The study explores the impact of parameters on image classification using two-dimensional fuzzy entropy and dispersion entropy. Results show that specific parameter choices are crucial for accurate image classification. By utilizing certain parameters, MFuzzyEn2D and MDispEn2D achieve outstanding performance in image classification on various datasets, with MFuzzyEn2D generally outperforming MDispEn2D.
Article
Computer Science, Interdisciplinary Applications
Delphine Lebret, Andreia S. Gaudencio, Mirvana Hilal, Sonia Saib, Rakelle Haidar, Michel Nonent, Anne Humeau-Heurtier
Summary: This study aims to objectively quantify the health status and quality of life of patients after uterine fibroid embolization (UFE) using a novel multiscale three-dimensional entropy-based texture analysis, called multiscale 3D dispersion entropy (MDispEn(3D)). The results show that there is an inverse correlation between MDispEn(3D) entropy values and both the size and volume of fibroids, as well as the severity of symptoms. The patient age and history of fibroids also play a role in the results, and lower MDispEn(3D) values indicate larger fibroids.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2022)
Article
Biology
Hassan Serhal, Nassib Abdallah, Jean-Marie Marion, Pierre Chauvet, Mohamad Oueidat, Anne Humeau-Heurtier
Summary: Atrial fibrillation (AF) is the most common supraventricular cardiac arrhythmia, resulting in high mortality rates among affected patients. Early and automatic prediction, detection, and classification of AF are crucial for effective treatment and reducing risks for patients. The research on AF using wavelets and artificial intelligence (AI) has shown great potential for future studies, with a focus on prediction, classification, and detection of AF episodes.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Physics, Multidisciplinary
Mirvana Hilal, Andreia S. Gaudencio, Pedro G. Vaz, Joao Cardoso, Anne Humeau-Heurtier
Summary: This article introduces a new entropy measurement method for analyzing colored images, discusses its sensitivity to parameters and recognition capabilities, and applies it to dermoscopic images. The results show that this method outperforms other methods in distinguishing images with different textures.
Article
Physics, Multidisciplinary
Emma Lhermitte, Mirvana Hilal, Ryan Furlong, Vincent O'Brien, Anne Humeau-Heurtier
Summary: The researchers proposed a new entropy-based approach for RGB image classification and compared it with deep learning methods, achieving promising results.
Article
Engineering, Biomedical
Hassan Serhal, Nassib Abdallah, Jean-Marie Marion, Pierre Chauvet, Mohamad Oueidat, Anne Humeau-Heurtier
Summary: This study proposes an algorithm based on electrocardiograms, wavelet transform, and convolutional neural networks to classify between healthy subjects and patients suffering from paroxysmal atrial fibrillation (PAF). By selecting informative intrinsic mode functions and utilizing continuous wavelet transform, features are extracted and input into the CNN classification model. The algorithm achieves an accuracy of 98.8% on the PTB-XL dataset.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Biology
Tala Abdallah, Nisrine Jrad, Fahed Abdallah, Anne Humeau-Heurtier, Patrick Van Bogaert
Summary: This study proposes a data-driven model based on deep learning and self-attention mechanism for epilepsy seizure detection. The method uses a one-dimensional convolutional neural network to extract features, and combines a long short-term memory module and a self-attention mechanism for processing. Experimental results demonstrate that this method performs well in seizure recognition tasks and exhibits robustness to inter-subject variability.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Article
Computer Science, Interdisciplinary Applications
Andreia S. Gaudencio, Hamed Azami, Joao M. Cardoso, Pedro G. Vaz, Anne Humeau-Heurtier
Summary: The study introduces six bidimensional algorithms based on ensemble techniques, which offer greater stability and reduced bias in entropy estimation compared to traditional methods. These new measures were tested on synthetic images and applied to a biomedical dataset, showing improved ability to detect image dynamics and randomness levels. Results indicate higher accuracy and sensitivity in classifying patients with pulmonary emphysema, paving the way for potential clinical applications in various imaging scenarios.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2023)
Article
Computer Science, Information Systems
Anne Humeau-Heurtier
Summary: In the field of image processing, texture features and color play fundamental roles as visual cues, with complementary functions. Various methods have been proposed for extracting texture features that consider both texture and color information. This paper provides a comprehensive survey of these methods, categorizes them, and presents their concept, advantages and drawbacks, along with examples of application.
Article
Computer Science, Information Systems
Pedro G. Vaz, Andreia S. Gaudencio, L. F. Requicha Ferreira, Anne Humeau-Heurtier, Miguel Morgado, Joao Cardoso
Summary: This paper investigates the influence of sampling basis and its order on the quality of reconstructed images in single-pixel imaging. Two new orders, AS and AI, are proposed and evaluated using simulation and experimental methods. The results show that these new orders outperform existing orders in terms of image quality for low sampling ratios and low resolution images.
Article
Computer Science, Interdisciplinary Applications
Antoine Jamin, Guillaume Duval, Cedric Annweiler, Pierre Abraham, Anne Humeau-Heurtier
Summary: This study examines the impact of age on bike navigation abilities, using a bike simulator and entropy measures. The results suggest that methods based on cross-distribution entropy may be effective in highlighting the decrease in navigation capacities with age. This indicates the potential of incorporating medical benefits into leisure equipment and the value of using virtual reality for studying the impact of age.
MEDICAL & BIOLOGICAL ENGINEERING & COMPUTING
(2021)