An Adaptive Method for Detecting and Removing EEG Noise / 中国医疗器械杂志

To solve the problem of real-time detection and removal of EEG signal noise in anesthesia depth monitoring, we proposed an adaptive EEG signal noise detection and removal method. This method uses discrete wavelet transform to extract the low-frequency energy and high-frequency energy of a segment of EEG signals, and sets two sets of thresholds for the low-frequency band and high-frequency band of the EEG signal. These two sets of thresholds can be updated adaptively according to the energy situation of the most recent EEG signal. Finally, we judge the level of signal interference according to the range of low-frequency energy and high-frequency energy, and perform corresponding denoising processing. The results show that the method can more accurately detect and remove the noise interference in the EEG signal, and improve the stability of the calculated characteristic parameters.

An anesthesia depth computing method study based on wavelet transform and artificial neural network / 生物医学工程学杂志

General anesthesia is an essential part of surgery to ensure the safety of patients. Electroencephalogram (EEG) has been widely used in anesthesia depth monitoring for abundant information and the ability of reflecting the brain activity. The paper proposes a method which combines wavelet transform and artificial neural network (ANN) to assess the depth of anesthesia. Discrete wavelet transform was used to decompose the EEG signal, and the approximation coefficients and detail coefficients were used to calculate the 9 characteristic parameters. Kruskal-Wallis statistical test was made to these characteristic parameters, and the test showed that the parameters were statistically significant for the differences of the four levels of anesthesia: awake, light anesthesia, moderate anesthesia and deep anesthesia (

Mejoramiento de contraste y segmentación en imágenes de úlceras del pie diabético / Contrast enhancement and segmentation in diabetic foot ulcer images

Introducción: El 3 a 5 por ciento de los pacientes diabéticos en Cuba sufren úlcera del pie diabético. Las imágenes fotográficas de estas úlceras permiten hacer evaluaciones cuantitativas de los tratamientos. En Cuba, dicha evaluación se hace manual o semiautomáticamente. No se registra software cubano que automáticamente realice la medición de las áreas de la lesión y permita conocer las características de la úlcera, antes y después de la aplicación de un tratamiento. Objetivo: Comparar cualitativamente métodos de preprocesamiento y segmentación de la úlcera, dada la ausencia de una regla de oro. Método: Estudio descriptivo y transversal en 6 pacientes diabéticos del Instituto Nacional de Angiología y Cirugía Vascular en octubre de 2018, con lesiones de grado I-IV en la escala de Wagner. Se utilizó el marco estereotáxico para extremidades FrameHeber03® para obtener imágenes planimétricas estandarizadas de las úlceras. Se obtuvieron 51 imágenes de úlceras que se preprocesaron mediante el algoritmo Transformada Wavelet Discreta Logarítmica en un modelo S-LIP y se determinó su borde mediante los métodos de segmentación Chan-Vese, modelo de mezclas gaussianas y GrabCut. Resultados: Se mostró la utilidad de preprocesar las imágenes para lograr mejores resultados en la segmentación. El mejor y más factible método de segmentación fue el de mezclas gaussianas. Los algoritmos resultaron ser más precisos en pacientes de piel oscura, debido al mayor contraste entre la piel y el borde de la úlcera. Conclusiones: El algoritmo de segmentación automática de mezclas gaussianas. puede incluirse en un software para medir el área de la úlcera(AU)

Introduction: The 3 to 5 percent of Cuban diabetic patients suffer from diabetic foot ulcer. The diabetic foot ulcer photographic images allow quantitative evaluations of a treatment. In Cuba, the ulcer area measurement is done manually or semi-automatically. There is no Cuban software reported that automatically measures the area, and allows knowing the state of the foot ulcer before and after a treatment. Goal: To compare qualitatively (given the absence of a gold standard) ulcer´s pre-processing and segmentation methods. Method: We develop a descriptive and transversal study with 6 diabetic patients from Nacional Institute of Angiology and Vascular Surgery during October, 2018, with lesions of degree I-IV in the Wagner scale. The stereotaxic frame FrameHeber03® was used for obtaining planimetric images of the ulcers. In all, 51 ulcer images were obtained, and then we pre-processed it by Logarithmic Discrete Wavelet Transform under a S-LIP model, and found the ulcer border with the segmentation methods Chan-Vese, Gaussian Mixture Model (GMM), and GrabCut. Results: The pre-processing step was crutial for obtaining good results in the segmentation step. The best performance was reached by the GMM segmentation method. The algorithms were more accurate in images with black skin patients, due to the high contrast between the skin and the ulcer border. Conclusions: The automatic segmentation method (GMM) could be included in a software for detecting the border of the diabetic foot ulcer(AU)

Multi class disorder detection of magnetic resonance brain images using composite features and neural network

Brain disorder recognition has becoming a promising area of study. In reality, some disorders share similar features and signs, making the task of diagnosis and treatment challenging. This paper presents a rigorous and robust computer aided diagnosis system for the detection of multiple brain abnormalities which can assist physicians in the diagnosis and treatment of brain diseases. In this system, we used energy of wavelet sub bands, textural features of gray level co-occurrence matrix and intensity feature of MR brain images. These features are ranked using Wilcoxon test. The composite features are classifi ed using back propagation neural network. Bayesian regulation is adopted to fi nd the optimal weights of neural network. The experimentation is carried out on datasets DS-90 and DS-310 of Harvard Medical School. To enhance the generalization capability of the network, fi vefold stratifi ed cross validation technique is used. The proposed system yields multi class disease classifi cation accuracy of 100% in diff erentiating 90 MR brain images into 18 classes and 97.81% in diff erentiating 310 MR brain images into 6 classes. The experimental results reveal that the composite features along with BPNN classifi er create a competent and reliable system for the identifi cation of multiple brain disorders which can be used in clinical applications. The Wilcoxon test outcome demonstrates that standard deviation feature along with energies of approximate and vertical sub bands of level 7 contribute the most in achieving enhanced multi class classifi cation performance results.

Real-time premature ventricular contractions detection based on Redundant Discrete Wavelet Transform

Abstract Introduction Premature Ventricular Contraction (PVC) is among the most common types of ventricular cardiac arrhythmia. However, it only poses danger if the person suffers from a heart disease, such as heart failure. Hence, this is an important factor to consider in heart disease people. This paper presents an ECG real-time analysis system for PVC detection. Methods This system is based on threshold adaptive methods and Redundant Discrete Wavelet Transform (RDWT), with a real-time approach. This analysis is based on wavelet coefficients energy for PVC detection. It is presented also a study to find the most indicated wavelet mother for ECG analysis application among the following wavelet families: Daubechies, Coiflets and Symlets. The system detection performance was validated on the MIT-BIH Arrhythmia Database. Results The best results were verified with db2 wavelet mother: the Sensitivity Se = 99.18%, Positive Predictive Value P+ = 99.15% and Specificity Sp = 99.94%, on 80.872 annotated beats, and 61.2 s processing speed for a half-hour record. Conclusion The proposed system exhibits reliable PVC detection, with real-time approach, and a simple algorithmic structure that can be implemented in many platforms.

Influences of the signal border extension in the discrete wavelet transform in EEG spike detection

Abstract Introduction The discrete wavelet transform is used in many studies as signal preprocessor for EEG spike detection. An inherent process of this mathematical tool is the recursive wavelet convolution over the signal that is decomposed into detail and approximation coefficients. To perform these convolutions, firstly it is necessary to extend signal borders. The selection of an unsuitable border extension algorithm may increase the false positive rate of an EEG spike detector. Methods In this study we analyzed nine different border extensions used for convolution and 19 mother wavelets commonly seen in other EEG spike detectors in the literature. Results The border extension may degrade an EEG spike detector up to 44.11%. Furthermore, results behave differently for distinct number of wavelet coefficients. Conclusion There is not a best border extension to be used with any EEG spike detector based on the discrete wavelet transform, but the selection of the most adequate border extension is related to the number of coefficients of a mother wavelet.

Study on improved algorithm for multimodal medical image fusion based on discrete wavelet transform / 重庆医学

Objective To propose an improved discrete wavelet transform (DWT ) and to apply it in multimodal medical im‐age fusion .Methods Firstly ,the source medical images were initially transformed into the high frequency and low frequency images by DWT ;then the high frequency part adopted the big direction absolute values ,which effectively preserved the detailed informa‐tion of image ,while the low frequency part used the fusion rule of local energy ratio for preserving the most of image information ;finally ,the discrete wavelet reverse transform was used for reconstructing the fusion sub‐images into fusion image .Results By comparing the fusion images by 3 groups of medical images ,this proposed algorithm was superior to other existing algorithms in the aspects of subjective visual effect and objective evaluation indicators .Conclusion The proposed algorithm of medical image fusion is rapid and accurate ,has excellent performance in the noise environment and clinical examples ,can obtain the high quality fusion im‐age and has higher clinical application value .

An Efficient Watermarking Scheme for Medical Data Security With the Aid of Neural Network

ABSTRACT Digital watermarking has emerged as major technique for ensuring security for various types of data like medical data, digital copyright protection, transaction tracing and so on. With the advancement in digital data distribution over the network there has been increase in the need for protection of such data from unauthorized copying or usages. Watermarking helps in providing the security to some extent. Robustness against any sort of unauthenticated attack is the major requirement of watermarking. In this paper we proposed an efficient watermarking technique for medical data security with the aid of neural network. Usage of neural network is generally used to create and control watermarking strength automatically. This method provides better watermarked data which can be highly secured to unauthorized usage. It is observed that the proposed method provides better security for the multimedia data when compared with other data security methods.

Pulmonary crackle characterization: approaches in the use of discrete wavelet transform regarding border effect, mother-wavelet selection, and subband reduction

Introduction Crackles are discontinuous, non-stationary respiratory sounds and can be characterized by their duration and frequency. In the literature, many techniques of filtering, feature extraction, and classification were presented. Although the discrete wavelet transform (DWT) is a well-known tool in this area, issues like signal border extension, mother-wavelet selection, and its subbands were not properly discussed. Methods In this work, 30 different mother-wavelets 8 subbands were assessed, and 9 border extension modes were evaluated. The evaluations were done based on the energy representation of the crackle considering the mother-wavelet and the border extension, allowing a reduction of not representative subbands. Results Tests revealed that the border extension mode considered during the DWT affects crackle characterization, whereas SP1 (Smooth-Padding of order 1) and ASYMW (Antisymmetric-Padding (whole-point)) modes shall not be used. After DWT, only 3 subbands (D3, D4, and D5) were needed to characterize crackles. Finally, from the group of mother-wavelets tested, Daubechies 7 and Symlet 7 were found to be the most adequate for crackle characterization. Discussion DWT can be used to characterize crackles when proper border extension mode, mother-wavelet, and subbands are taken into account.

CLASIFICACIÓN DE LATIDOS DE CONTRACCIÓN VENTRICULAR PREMATURA BASADOS EN MÉTODOS DE APRENDIZAJE NO SUPERVISADO / CLASSIFICATION OF PREMATURE VENTRICULAR CONTRACTION BEATS BASED ON UNSUPERVISED LEARNING METHODS / CLASSIFICAÇÃO DE PREMATUROS BATIDAS CONTRAÇÃO VENTRICULAR BASEADO SUPERVISIONADO MÉTODOS DE APRENDIZAGEM

Las enfermedades cardiovasculares son la principal causa de mortalidad en el mundo, por lo que el desarrollo de algoritmos que detecten arritmias cardíacas en tiempo real se ha convertido en un campo de investigación importante. El desarrollo de estos algoritmos ha conllevado a la mejora de dispositivos cardiacos portátiles. Este artículo presenta el desempeño de dos algoritmos basados en aprendizaje de máquina no supervisado para la detección de latidos de contracción ventricular prematura en la señal ECG. Los latidos se extraen de las bases de datos del MIT-BIH, los cuales fueron pre-procesados y segmentados por el grupo de investigación de Dinámica Cardiovascular de la UPB. La Transformada Wavelet Discreta, el Análisis de Componentes Principales y un método híbrido propuesto son implementados para la extracción de características y reducción de dimensiones, a partir de los cuales se generan 8 espacios de características para la evaluación de los algoritmos. Kmeans y Mapas auto-organizados son desarrollados y comparados en términos de precisión y costo computacional. Se logró una especificidad del 96.22 % y una sensibilidad del 95.04 % con un tiempo de ejecución de 79.41µs por latido. Los resultados permiten concluir que estos métodos pueden implementarse en aplicaciones de detección de arritmias en tiempo real debido a su bajo costo computacional.

Cardiovascular diseases are the principal cause of mortality in the world, so that the development of algorithms that detect cardiac arrhythmias in real time has become an important field of research. The development of these algorithms has led to the improvement of wearable cardiac devices. This paper presents the performance of two algorithms based in unsupervised learning methods for the detection of Premature Ventricular Contraction in the ECG signal. The beats are extracted from MIT-BIH databases, which were preprocessed and segmented by the UPB’s Dynamic Cardiovascular research group. The Discrete Wavelet Transform (DWT), Principal Component Analysis (PCA) and a proposed hybrid method are implemented for the feature extraction and dimension reduction, from which 8 feature spaces are generated and tested. Kmeans and Self Organizing Maps are developed and compared in terms of accuracy and computational cost. Specificity of 96.22 % and sensitivity of 95.94% with 79.41µs per beat are accomplished. The results show that these methods can be implemented in applications of real time arrhythmia detection because of their low computational cost.

A doença cardiovascular é a principal causa de morte em todo o mundo, de modo que o desenvolvimento de algoritmos para detectar arritmias cardíacas, em tempo real, tornou-se um importante campo de pesquisa. O desenvolvimento desses algoritmos tem levado a melhores dispositivos cardíacos portáteis. Este artigo apresenta o desempenho dos dois com base na aprendizagem de máquina sem supervisão para detecção de batidas de contração ventriculares prematuras nos algoritmos de sinais de ECG. As batidas são extraídos das bases de dados do MIT-BIH, que foram pré-processados e segmentado pelo grupo da UPB Cardiovasculares Dynamics pesquisa. A Transformada Wavelet Discreta, Análise de Componentes Principais e uma abordagem híbrida proposta são implementadas para extração de características e redução de dimensão, a partir do qual 8 espaços de recursos para a avaliação dos algoritmos são gerados. Kmeans e mapas de auto-organização são desenvolvidos e comparados em termos de precisão e custo computacional. A especificidade de 96,22% e uma sensibilidade de 95,04% com um tempo de execução de 79.41µs por batida foi alcançado. Os resultados mostram que estes métodos podem ser implementados em aplicações de detecção de arritmia em tempo real, devido ao seu baixo custo computacional.

DETECCIÓN DE POTENCIALES EVOCADOS RELACIONADOS A EVENTOS EN INTERFACES CEREBRO-COMPUTADORA MEDIANTE TRANSFORMADA WAVELET / DETECTION OF EVENT-RELATED POTENTIALS IN BRAIN-COMPUTER INTERFACES USING THE WAVELET TRANSFORM / DETECÇÃO DE POTENCIAIS EVOCADOS RELACIONADOS A EVENTOS EM INTERFACES CÉREBRO-COMPUTADOR USANDO TRANSFORMADA WAVELET

Una interfaz cerebro computadora (ICC) es un sistema que provee una forma de comunicación directa entre el cerebro de una persona y el mundo exterior. Para el presente trabajo se utilizaron ICC basadas en EEG utilizando el paradigma de potenciales evocados relacionados con eventos (PRE). El objetivo de este trabajo es resolver en forma eficiente el problema de clasificación, en el cual se tienen dos clases posibles: registros con respuesta (PRE) y registros sin respuesta. Para esto se propone evaluar el desempeño de una ICC utilizando la transformada wavelet diádica discreta (DDWT, del inglés Dyadic Discrete Wavelet Transform) y la transformada wavelet packet (WPT, del inglés Wavelet Packet Transform) como métodos de extracción de características para la detección de la señal de PRE. La base de datos utilizada posee registros de EEG de época única de diez sujetos sanos. A partir de los patrones temporales (registros sin post-procesamiento) se generaron cinco conjuntos de patrones wavelet luego de aplicar la DDWT y WPT mediante diferentes técnicas. Se evaluó el desempeño de cada conjunto de patrones wavelet y de los patrones temporales mediante un clasificador lineal de Fisher. Se encontró que los patrones DDWT filtrados a 16 Hz presentan resultados de clasificación superiores a los patrones temporales. De esta manera al mejorar la etapa de extracción de características se mejora la clasificación, y consecuentemente, el desempeño del sistema completo de una ICC.

A brain-computer interface (BCI) is a system that provides a direct communication between the brain of a person and the outside world. For the present work we used an EEG-based event-related evoked potentials BCI. This paper aims to efficiently solve the problem of classification, which has two possible classes: recordings with evoked-potentials (ERP) and recordings without them. We proposed to evaluate the performance of a BCI using the discrete dyadic wavelet transform (DDWT) and the wavelet packet transform (WPT) as feature extraction methods for ERP signal detection. The database consisted of single-epoch EEG recordings from ten healthy subjects. From temporal patterns (recordings without any post-processing), five wavelet patterns were generated after applying DDWT and WPT via different techniques. The performance of the wavelet and temporal patterns were analyzed with the Fisher linear classifier finding that DDWT patterns, filtered at 16 Hz, presented better classification results than temporal patterns. This means that improving the feature extraction step, improves classification, and consequently, the performance of the entire BCI system.

Uma interface cérebro-computador (BCI) é um sistema que fornece uma forma de comunicação direta entre o cérebro de uma pessoa e o mundo exterior. Para este trabalho foram utilizados ICC baseado EEG evocados usando o paradigma de potenciais relacionados a eventos (ERP). O objetivo deste trabalho é resolver de forma eficiente o problema de classificação, em que há duas classes possíveis: registros Respondidas (PRE) e registros sem resposta. Para isso é avaliar o desempenho de uma ICC usando a wavelet diádica transformada discreta (DDWT, Discrete Wavelet Diádica Inglês Transform) e transformar pacote wavelet (WPT Transformada Wavelet Packet Inglês) como métodos de extração de características para a detecção de sinal PRE. A base de dados utilizada tem apenas EEG registra o tempo de dez indivíduos saudáveis. A partir dos padrões temporais (sem registros de pósprocessamento), cinco conjuntos de padrões após a aplicação wavelet e WPT DDWT gerado por várias técnicas. O desempenho de cada conjunto de padrões de wavelet e padrões temporais usando um classificador linear Fisher foi avaliado. Descobrimos que os padrões DDWT filtrados para 16 Hz apresentaram resultados acima da classificação padrões temporais. Assim, para melhorar a classificação de estágio de extração de características é melhorada, e, consequentemente, o desempenho de todo o sistema no ICC.

Análisis de señales electroencefalográficas para la clasificación de habla imaginada / Electroencephalographic Signals Analysis for Imagined Speech Classification

El presente trabajo tiene como objetivo interpretar las señales de EEG registradas durante la pronunciación imaginada de palabras de un vocabulario reducido, sin emitir sonidos ni articular movimientos (habla imaginada o no pronunciada) con la intención de controlar un dispositivo. Específicamente, el vocabulario permitiría controlar el cursor de la computadora, y consta de las palabras del lenguaje español: "arriba", "abajo", "izquierda", "derecha", y "seleccionar". Para ello, se registraron las señales de EEG de 27 individuos utilizando un protocolo básico para saber a priori en qué segmentos de la señal la persona imagina la pronunciación de la palabra indicada. Posteriormente, se utiliza la transformada wavelet discreta (DWT) para extraer características de los segmentos que son usados para calcular la energía relativa wavelet (RWE) en cada una de los niveles en los que la señal es descompuesta, y se selecciona un subconjunto de valores RWE provenientes de los rangos de frecuencia menores a 32 Hz. Enseguida, éstas se concatenan en dos configuraciones distintas: 14 canales (completa) y 4 canales (los más cercanos a las áreas de Broca y Wernicke). Para ambas configuraciones se entrenan tres clasificadores: Naive Bayes (NB), Random Forest (RF) y Máquina de vectores de soporte (SVM). Los mejores porcentajes de exactitud se obtuvieron con RF cuyos promedios fueron 60.11% y 47.93% usando las configuraciones de 14 canales y 4 canales, respectivamente. A pesar de que los resultados aún son preliminares, éstos están arriba del 20%, es decir, arriba del azar para cinco clases. Con lo que se puede conjeturar que las señales de EEG podrían contener información que hace posible la clasificación de las pronunciaciones imaginadas de las palabras del vocabulario reducido.

This work aims to interpret the EEG signals associated with actions to imagine the pronunciation of words that belong to a reduced vocabulary without moving the articulatory muscles and without uttering any audible sound (imagined or unspoken speech). Specifically, the vocabulary reflects movements to control the cursor on the computer, and consists of the Spanish language words: "arriba", "abajo", "izquierda", "derecha", and "seleccionar". To do this, we have recorded EEG signals from 27 subjects using a basic protocol to know a priori in what segments of the signal a subject imagines the pronunciation of the indicated word. Subsequently, discrete wavelet transform (DWT) is used to extract features from the segments. These are used to compute relative wavelet energy (RWE) in each of the levels in that EEG signal is decomposed and, it is selected a RWE values subset with the frequencies smaller than 32 Hz. Then, these are concatenated in two different configurations: 14 channels (full) and 4 channels (the channels nearest to the brain areas of Wernicke and Broca). The following three classifiers were trained using both configurations: Naive Bayes (NB), Random Forest (RF) and support vector machines (SVM). The best accuracies were obtained by RF whose averages were 60.11% and 47.93% using both configurations, respectively. Even though, the results are still preliminary, these are above 20%, this means they are more accurate than chance for five classes. Based on them, we can conjecture that the EEG signals could contain information needed for the classification of the imagined pronunciations of the words belonging to a reduced vocabulary.