Comparison of initial learning algorithms for long short-term memory method on real-time respiratory signal prediction.

Aim: This study aimed to examine the effect of the weight initializers on the respiratory signal prediction performance using the long short-term memory (LSTM) model. Methods: Respiratory signals collected with the CyberKnife Synchrony device during 304 breathing motion traces were used in this study. The effectiveness of four weight initializers (Glorot, He, Orthogonal, and Narrow-normal) on the prediction performance of the LSTM model was investigated. The prediction performance was evaluated by the normalized root mean square error (NRMSE) between the ground truth and predicted respiratory signal. Results: Among the four initializers, the He initializer showed the best performance. The mean NRMSE with 385-ms ahead time using the He initializer was superior by 7.5%, 8.3%, and 11.3% as compared to that using the Glorot, Orthogonal, and Narrow-normal initializer, respectively. The confidence interval of NRMSE using Glorot, He, Orthogonal, and Narrow-normal initializer were [0.099, 0.175], [0.097, 0.147], [0.101, 0.176], and [0.107, 0.178], respectively. Conclusions: The experiment results in this study indicated that He could be a valuable initializer in the LSTM model for the respiratory signal prediction.

Conditional GAN based augmentation for predictive modeling of respiratory signals.

Respiratory illness is the primary cause of mortality and impairment in the life span of an individual in the current COVID-19 pandemic scenario. The inability to inhale and exhale is one of the difficult conditions for a person suffering from respiratory disorders. Unfortunately, the diagnosis of respiratory disorders with the presently available imaging and auditory screening modalities are sub-optimal and the accuracy of diagnosis varies with different medical experts. At present, deep neural nets demand a massive amount of data suitable for precise models. In reality, the respiratory data set is quite limited, and therefore, data augmentation (DA) is employed to enlarge the data set. In this study, conditional generative adversarial networks (cGAN) based DA is utilized for synthetic generation of signals. The publicly available repository such as ICBHI 2017 challenge, RALE and Think Labs Lung Sounds Library are considered for classifying the respiratory signals. To assess the efficacy of the artificially created signals by the DA approach, similarity measures are calculated between original and augmented signals. After that, to quantify the performance of augmentation in classification, scalogram representation of generated signals are fed as input to different pre-trained deep learning architectures viz Alexnet, GoogLeNet and ResNet-50. The experimental results are computed and performance results are compared with existing classical approaches of augmentation. The research findings conclude that the proposed cGAN method of augmentation provides better accuracy of 92.50% and 92.68%, respectively for both the two data sets using ResNet 50 model.

[Extraction of respiratory signals from chest tomosynthesis].

OBJECTIVE: To analyze the respiratory motion of the scanned object during acquisition of digital chest tomosynthesis (CTS) using a linear model. OBJECTIVE: Respiratory signals were generated by extracting the motion of the diaphragm from the projection radiographs. The diaphragm trajectory obtained by dynamic programming (DP) was modeled and fitted, and according to the fitting of the data, the base motion curve and respiratory signal curve of the diaphragm were separated. Multipurpose chest phantom data, simulated digital Xcat phantom data and the datasets of 3 clinical patients were used to validate the performance of the proposed method. OBJECTIVE: The motion trajectory of the diaphragm extracted from multipurpose chest phantom simulation data was linear. The respiratory signals could be effectively extracted from the 3 datasets of clinical patients in different respiratory states. The correlation coefficient between the respiratory signal extracted in Xcat simulation experiment and the original design was 0.9797. OBJECTIVE: The linear model can effectively obtain the respiratory motion information of patients in real time, thus enabling the physicians to make clinical decisions on a rescan.

[Development of Respiratory Signal Extraction System Based on Photoplethysmography].

Oxygen saturation and respiratory signals are important physiological signals of human body, respiratory monitoring plays an important role in clinical and daily life. A system was established to extract respiratory signals from photoplethysmography in this study. Including the collection of pulse wave signal, the extraction of respiratory signal, and the calculation of respiratory rate and pulse rate transmitted from the slave computer to the host computer in real time.

Development of Respiratory Signal Extraction System Based on Photoplethysmography / 中国医疗器械杂志

Oxygen saturation and respiratory signals are important physiological signals of human body, respiratory monitoring plays an important role in clinical and daily life. A system was established to extract respiratory signals from photoplethysmography in this study. Including the collection of pulse wave signal, the extraction of respiratory signal, and the calculation of respiratory rate and pulse rate transmitted from the slave computer to the host computer in real time.

Experimental Study of Wireless Monitoring of Human Respiratory Movements Using UWB Impulse Radar Systems.

This paper analyzes and discusses the capability of human being detection using impulse ultra-wideband (UWB) radar with an improved detection algorithm. The multiple automatic gain control (AGC) technique is employed to enhance the amplitudes of human respiratory signals. Two filters with seven values averaged are used to further improve the signal-to-noise ratio (SNR) of the human respiratory signals. The maximum slope and standard deviation are used for analyzing the characteristics of the received pulses, which can provide two distance estimates for human being detection. Most importantly, based on the two distance estimates, we can accurately judge whether there are human beings in the detection environments or not. The data size can be reduced based on the defined interested region, which can improve the operation efficiency of the radar system for human being detection. The developed algorithm provides excellent performance regarding human being detection, which is validated through comparison with several well-known algorithms.

Correlation analysis of respiratory signals by using parallel coordinate plots.

BACKGROUND AND OBJECTIVES: The understanding of the bonds and the relationships between the respiratory signals, i.e. the airflow, the mouth pressure, the relative temperature and the relative humidity during breathing may provide the improvement on the measurement methods of respiratory mechanics and sensor designs or the exploration of the several possible applications in the analysis of respiratory disorders. Therefore, the main objective of this study was to propose a new combination of methods in order to determine the relationship between respiratory signals as a multidimensional data. METHODS: In order to reveal the coupling between the processes two very different methods were used: the well-known statistical correlation analysis (i.e. Pearson's correlation and cross-correlation coefficient) and parallel coordinate plots (PCPs). Curve bundling with the number intersections for the correlation analysis, Least Mean Square Time Delay Estimator (LMS-TDE) for the point delay detection and visual metrics for the recognition of the visual structures were proposed and utilized in PCP. RESULTS: The number of intersections was increased when the correlation coefficient changed from high positive to high negative correlation between the respiratory signals, especially if whole breath was processed. LMS-TDE coefficients plotted in PCP indicated well-matched point delay results to the findings in the correlation analysis. Visual inspection of PCB by visual metrics showed range, dispersions, entropy comparisons and linear and sinusoidal-like relationships between the respiratory signals. CONCLUSION: It is demonstrated that the basic correlation analysis together with the parallel coordinate plots perceptually motivates the visual metrics in the display and thus can be considered as an aid to the user analysis by providing meaningful views of the data.

Sleep apnea classification based on respiration signals by using ensemble methods.

In this study, an efficient and robust method classifying the minute based occurrence of sleep apnea is aimed. Three respiration signals obtained from abdominal, chest and nasal way extracted from polysomnography recordings. Wavelet transform based on feature extraction methods are applied on the 1 minute length respiration signals. Dimension reduction process is facilitated by using principal component analysis. The features obtained from 8 recordings are used for the classification sleep apnea by using three ensemble classifiers. According to the results, the classification accuracies have been obtained between 92.07-98.43%, 92.75-98.68% and 92.42-98.61% by using three different ensemble classifier based on abdominal, chest and nasal based analysis, respectively for AdaBoost, Random Forest and Random Subspace. However the best result is obtained analyzing nasal based respiratory signal by using Random Forest method. In this case accuracy is 98.68%.

Spectral analysis of respiratory flow and gas concentrations / Análise espectral de fluxo e concentrações de gases respiratórios

In this research we obtained samples of human respiratory flow, oxygen concentration and carbon dioxide concentration signals from 20 healthy subjects and evaluated the average power spectral density (PSD) of these signals. For each subject,the respiratory samples were acquired in four progressive levels of exercise in a cycle ergometer. Auto regressive moving average models were designed to represent the PSD found in each phase. An average PSD of the four levels was also calculated. Results have shown that the bandwidth of O2 concentration, CO2 concentration and flow signals was 8  Hz, 7 Hz, and 15  Hz, respectively, within the dynamic range of 50  dB. The PSD curves found can be used for optimal filter design for signal enhancing in fast on-line measurement of these signals.

Nesta pesquisa foram registradas amostras dos sinais respiratórios de fluxo, concentração de oxigênio e concentração de gás carbônico em 20 voluntários saudáveis. A densidade espectral de potência (DEP) média foi então calculada. Para cada voluntário, as amostras dos sinais foram registradas em quatro intensidades progressivas de esforço físico em uma bicicleta ergométrica. Para representar a DEP encontrada em cada fase foram ajustados modelos auto-regressivos de média móvel. Uma DEP média entre as quatro intensidades também é fornecida. Os resultados mostraram que as larguras de banda dos sinais de concentração de O2, concentração de CO2 e fluxo foram 8  Hz, 7  Hz e 15  Hz, respectivamente, dentro de uma faixa dinâmica de 50  dB. As curvas de DEP encontradas podem ser usadas em projetos de filtros ótimos para equalização destes sinais em medições em tempo real.