Di Ge

Enhanced sampling schemes for MCMC based blind Bernoulli-Gaussian deconvolution

This paper proposes and compares two new sampling schemes for sparse deconvolution using a Bernoulli-Gaussian model. To tackle such a deconvolution problem in a blind and unsupervised context, the Markov Chain Monte Carlo (MCMC) framework is usually adopted, and the chosen sampling scheme is most often the Gibbs sampler. However, such a sampling scheme fails to explore the state space efficiently. Our first alternative, the K-tuple Gibbs sampler, is simply a grouped Gibbs sampler. The second one, called partially marginalized sampler, is obtained by integrating the Gaussian amplitudes out of the target distribution. While the mathematical validity of the first scheme is obvious as a particular instance of the Gibbs sampler, a more detailed analysis is provided to prove the validity of the second scheme. For both methods, optimized implementations are proposed in terms of computation and storage cost. Finally, simulation results validate both schemes as more efficient in terms of convergence time compared with the plain Gibbs sampler. Benchmark sequence simulations show that the partially marginalized sampler takes fewer iterations to converge than the K-tuple Gibbs sampler. However, its computation load per iteration grows almost quadratically with respect to the data length, while it only grows linearly for the K-tuple Gibbs sampler.

Spike Sorting by Stochastic Simulation

The decomposition of multiunit signals consists of the restoration of spike trains and action potentials in neural or muscular recordings. Because of the complexity of automatic decomposition, semiautomatic procedures are sometimes chosen. The main difficulty in automatic decomposition is the resolution of temporally overlapped potentials. In a previous study , we proposed a Bayesian model coupled with a maximum a posteriori (MAP) estimator for fully automatic decomposition of multiunit recordings and we showed applications to intramuscular EMG signals. In this study, we propose a more complex signal model that includes the variability in amplitude of each unit potential. Moreover, we propose the Markov Chain Monte Carlo (MCMC) simulation and a Bayesian minimum mean square error (MMSE) estimator by averaging on samples that converge in distribution to the joint posterior law. We prove the convergence property of this approach mathematically and we test the method representatively on intramuscular multiunit recordings. The results showed that its average accuracy in spike identification is greater than 90% for intramuscular signals with up to 8 concurrently active units. In addition to intramuscular signals, the method can be applied for spike sorting of other types of multiunit recordings.

Channelized model observer for the detection and estimation of signals with unknown amplitude, orientation, and size

As a task-based approach for medical image quality assessment, model observers (MOs) have been proposed as surrogates for human observers. While most MOs treat only signal-known-exactly tasks, there are few studies on signal-known-statistically (SKS) MOs, which are clinically more relevant. In this paper, we present a new SKS MO named channelized joint detection and estimation observer (CJO), capable of detecting and estimating signals with unknown amplitude, orientation, and size. We evaluate its estimation and detection performance using both synthesized (correlated Gaussian) backgrounds and real clinical (magnetic resonance) backgrounds. The results suggest that the CJO has good performance in the SKS detection-estimation task.

A multi-slice model observer for medical image quality assessment

Model observers (MOs) have been developed for the medical image quality assessment. Nowadays, numerous modern medical instruments are capable of producing 3D images, while few researchers have conducted MO studies on 3D data. In this paper, we propose a multi-slice MO when considering a relatively more realistic diagnostic task: the detection-localization of simulated multiple-sclerosis (MS) lesions on 3D magnetic resonance (MR) images. The jackknife free-response receiver operating characteristic (JAFROC) method was used to quantitatively analyse its performances and compare them with those of human observers. Our preliminary results showed that the proposed framework has the potential to approach human detection-localization task performance.

Coupled Hidden Markov Model-Based Method for Apnea Bradycardia Detection

In this paper, we present a novel framework for the coupled hidden Markov model (CHMM), based on the forward and backward recursions and conditional probabilities, given a multidimensional observation. In the proposed framework, the interdependencies of states networks are modeled with Markovian-like transition laws that influence the evolution of hidden states in all channels. Moreover, an offline inference approach by maximum likelihood estimation is proposed for the learning procedure of model parameters. To evaluate its performance, we first apply the CHMM model to classify and detect disturbances using synthetic data generated by the FitzHugh-Nagumo model. The average sensitivity and specificity of the classification are above 93.98% and 95.38% and those of the detection reach 94.49% and 99.34%, respectively. The method is also evaluated using a clinical database composed of annotated physiological signal recordings of neonates suffering from apnea-bradycardia. Different combinations of beat-to-beat features extracted from electrocardiographic signals constitute the multidimensional observations for which the proposed CHMM model is applied, to detect each apnea bradycardia episode. The proposed approach is finally compared to other previously proposed HMM-based detection methods. Our CHMM provides the best performance on this clinical database, presenting an average sensitivity of 95.74% and specificity of 91.88% while it reduces the detection delay by -0.59 s.

Switching Kalman filter based methods for apnea bradycardia detection from ECG signals

Apnea bradycardia (AB) is an outcome of apnea occurrence in preterm infants and is an observable phenomenon in cardiovascular signals. Early detection of apnea in infants under monitoring is a critical challenge for the early intervention of nurses. In this paper, we introduce two switching Kalman filter (SKF) based methods for AB detection using electrocardiogram (ECG) signal.The first SKF model uses McSharrys ECG dynamical model integrated in two Kalman filter (KF) models trained for normal and AB intervals. Whereas the second SKF model is established by using only the RR sequence extracted from ECG and two AR models to be fitted in normal and AB intervals. In both SKF approaches, a discrete state variable called a switch is considered that chooses one of the models (corresponding to normal and AB) during the inference phase. According to the probability of each model indicated by this switch, the model with larger probability determines the observation label at each time instant.It is shown that the method based on ECG dynamical model can be effectively used for AB detection. The detection performance is evaluated by comparing statistical metrics and the amount of time taken to detect AB compared with the annotated onset. The results demonstrate the superiority of this method, with sensitivity and specificity 94.74[Formula: see text] and 94.17[Formula: see text], respectively. The presented approaches may therefore serve as an effective algorithm for monitoring neonates suffering from AB.

Early detection of apnea-bradycardia episodes in preterm infants based on coupled hidden Markov model

The incidence of apnea-bradycardia episodes in preterm infants may lead to neurological disorders. Prediction and detection of these episodes are an important task in healthcare systems. In this paper, a coupled hidden Markov model (CHMM) based method is applied to detect apnea-bradycardia episodes. This model is evaluated and compared with two other methods based on hidden Markov model (HMM) and hidden semi-Markov model (HSMM). Evaluation and comparison are performed on a dataset of 233 apnea-bradycardia episodes which have been manually annotated. Observations are composed of RR-interval time series and QRS duration time series. The performance of each method was evaluated in terms of sensitivity, specificity and time detection delay. Results show that CHMM has the sensitivity of 84.92%, specificity of 94.17% and time detection delay of 2.32±4.82 seconds, which are better than the reference methods.

Numerical Stability issues on Channelized Hotelling Observer under different background assumptions

This paper addresses the numerical stability issue on the channelized Hotelling observer (CHO). The CHO is a well-known approach in the medical image quality assessment domain. Many researchers have found that the detection performance of the CHO does not increase with the number of channels, contrary to expectation. And to our knowledge, nobody in this domain has found the reason. We illustrated that this is due to the ill-posed problem of the scatter matrix and proposed a solution based on Tikhonov regularization. Although Tikhonov regularization has been used in many other domains, we show in this paper another important application of Tikhonov regularization. This is very important for researchers to continue the CHO (and other channelized model observer) investigation with a reliable detection performance calculation.

Advanced classification of ambulatory activities using spectral density distances and heart rate

Abstract As motion sensors are getting light-weighted and low-priced, there is a growing appetite for the accelerometer-based approaches for efficiently monitoring human activities. This paper proposes an original feature selection approach based on the spectral distances between a given signal and an activity model. This new technique is evaluated and compared to existing techniques in literature. This study also investigates the improvement of classification performances brought by the heart rate (HR) data in addition to the accelerometer data. The experimental dataset is composed of both acceleration and HR recordings from eight volunteers performing five ambulation activities. Four wearable sensor units, including an ECG node are employed. The response of the system to three widely used classifiers, the K-nearest neighbors K-NN, the Naive Bayes NB and the decision Tree C4.5 is reported along with the classification rates. The results reached up to 99% of overall recognition accuracy and higher than 98% using a single-sensor acceleration data and the HR data. These results demonstrate that the spectral distances approach can be adopted to accurately classify activities and that the joint processing of acceleration signals together with the HR signals can increase the classification accuracy compared to the case when processing the acceleration signals alone.