Automatic Differentiation of Normal and Continuous Adventitious Respiratory Sounds Using Ensemble Empirical Mode Decomposition and Instantaneous Frequency

Abstract

Abstract-Differentiatingnormal from adventitious respiratory sounds (RS) is a major challenge in the diagnosis of pulmonary diseases. Particularly, continuous adventitious sounds (CAS) are of clinical interest because they reflect the severity of certain diseases. This study presents a new classifier that automatically distinguishes normal sounds from CAS. It is based on the multi-scale analysis of instantaneous frequency (IF) and envelope (IE) calculated after ensemble empirical mode decomposition (EEMD). These techniques have two major advantages over previous techniques: high temporal resolution is achieved by calculating IF-IE and a priori knowledge of signal characteristics is not required for EEMD. The classifier is based on the fact that the IF dispersion of RS signals markedly decreases when CAS appear in respiratory cycles. Therefore, CAS were detected by using a moving window to calculate the dispersion of IF sequences. The study dataset contained 1494 RS segments extracted from 870 inspiratory cycles recorded from 30 patients with asthma. All cycles and their RS segments were previously classified as containing normal sounds or CAS by a highly experienced physician to obtain a gold standard classification. A support vector machine classifier was trained and tested using an iterative procedure in which the dataset was randomly divided into training (65%) and testing (35%) sets inside a loop. The SVM classifier was also tested on 4592 simulated CAS cycles. High total accuracy was obtained with both recorded (94.6% ± 0.3%) and simulated (92.8% ± 3.6%) signals. We conclude that the proposed method is promising for RS analysis and classification.Index Terms-Asthma, continuous adventitious sounds (CAS), ensemble empirical mode decomposition (EEMD), instantaneous frequency (IF), respiratory sounds (RS), wheezes.

Performance evaluation of the Hilbert–Huang transform for respiratory sound analysis and its application to continuous adventitious sound characterization

“…Furthermore, although the properties of the HHT have led to its application to a number of biomedical signals [35][36][37][38], it has rarely been used for RS analysis, as there are only a few studies, mainly focusing on DAS detection [39][40][41]. However, we found in our previous studies that the HHT also performed well in CAS detection [42,43], which inspired us to analyze its performance for CAS characterization in depth and explore its advantages over spectrogram, which has traditionally been the most commonly used technique for this purpose.…”

“…Nevertheless, the original EMD has been used in other RS analysis approaches [45][46][47][48]. Among the proposed solutions for MM, the ensemble EMD (EEMD) [44,49] and the noise-assisted multivariate EMD (NA-MEMD) [50] are some of the most wellestablished and widely used methods, but they have rarely been applied to RS analysis [43,51]. Moreover, the implementation and performance of these methods depend on each application and a detailed analysis of the MM effect and the performance of EEMD and NA-MEMD in RS signals is lacking.…”

Detección Automática y Clasificación de Eventos en Sonidos Cardiopulmonares de Sujetos Saludables

“…La forma clásica de identificación de sonidos pulmonares es la auscultación con estetoscopio simple, pero este método cuenta con la desventaja de que el oído humano tiene un rango limitado y hay frecuencias fuera de la percepción humana, lo que le dificulta al médico diagnosticar con certeza la existencia de alguna enfermedad [5] . Además, con el paso del tiempo el oído humano va perdiendo eficiencia en la percepción de los sonidos, por lo cual es necesario un sistema que no dependa del oído humano que pueda segmentar y clasificar sonidos cardiacos y sonidos pulmonares por métodos automatizados y computarizados.…”

“…Algunos autores proponen métodos basados en análi-sis multi-escala de frecuencias instantáneas (IF) y envolventes instantáneas (IE) que son calculadas después de aplicar un conjunto de modos de descomposi-ción empírica (EEMD) [5] . Aquí, las señales respiratorias son descompuestas en una serie de componentes, llamadas funciones de modo intrínseco (IMFs), para los cuales la IF puede ser definida en cada punto.…”

“…Los clasificadores están basados en el hecho de que la dispersión de las IFs de las señales respiratorias disminuye marcadamente cuando los ruidos continuos adventicios (CAS) aparecen en los ciclos de respiración. Dado que los segmentos de CAS tienen menor dispersión de frecuencias instantáneas que los segmentos que contienen respiración normal, los segmentos que son considerados como CAS pueden ser detectados con una ventana deslizante, la cual se desliza sobre la totalidad de las secuencias IFs [5] . Sin embargo, sigue siendo no evidente obtener la componente correcta.…”