Md. Aynal Haque

Parameter estimation of multichannel autoregressive processes in noise

We present a new multichannel autoregressive parameter estimation method using a finite set of noisy observations without a priori knowledge of additive noise power. The proposed method is based on soving alternatively a set of nonlinear and a set of linear equations. The Newton-Raphson iteration algorithm is used to estimate the unknown noise variances solving the nonlinear equations while the unknown AR parameter matrices are estimated solving the noise-compensated Yule-Walker equations linearly. Computer simulation results are presented to evaluate the performance of the proposed method.

Developmental patterns of O2 consumption, heart rate and O2 pulse in unturned eggs.

The effects of failure to turn eggs on the developmental patterns of oxygen consumption (MO2), heart rate (fH) and O2 pulse during the second half of incubation of individual chicken eggs were examined. The MO2 of unturned eggs increased at a significantly lower rate than the control toward the end of prenatal incubation, and the plateau MO2 between day 17 and 19 was significantly lower than the control. Lack of turning also resulted in significant changes in the developmental patterns of fH and O2 pulse. It is suggested that the effects of lack of egg-turning on the developmental patterns of MO2 may be attributable to lower embryonic growth rate in addition to impairment of gas exchange through the chorioallantoic gas exchanger.

Comparisons between invasive and noninvasive determinations of embryonic heart rate in chickens

The embryonic heart rate (HR) of chicken eggs was determined by ballistocardiography, electrocardiography, impedance-cardiography and acoustocardiography. The effects of these different methods on the HR were examined for young (12-day-old) and late (16- and 18-day-old) embryos. Ballistocardiography is a non-invasive method. Nevertheless, the mean HRs determined using this method at 5-hr intervals were significantly different in many embryos. This difference was caused by the natural variability of embryonic HR. The statistical analyses indicated that any of these four methods did not influence the HR measurement and can be used as a simple means for determination of the embryonic HR.

An expert system for automated identification of obstructive sleep apnea from single-lead ECG using random under sampling boosting

Computerized obstructive sleep apnea detection is necessary to speed-up sleep apnea diagnosis and research and for assisting medical professionals. Moreover, the development of a device to monitor sleep apnea that is low-power and portable, requires a reliable and successful sleep apnea detection scheme. In this article, the problem of automated sleep apnea detection using singe-lead electrocardiogram (ECG) signals has been addressed. At first, segments of ECG signals are decomposed using a data-adaptive signal decomposition scheme, namely- tunable-Q factor wavelet transform (TQWT). Three statistical features are extracted from the TQWT sub-bands and train and test matrices are formed afterwards. These matrices are fed into the classifier to identify non-apneic and apneic ECG signal segments. In this work, a new machine learning algorithm, namely- random under sampling boosting (RUSBoost) is implemented to perform classification. This is for the first time TQWT along with RUSBoost is employed for automatic sleep apnea detection to our knowledge. The overall algorithmic performance of our method is inspected for various values of TQWT parameters. Optimal values of these parameters are investigated and determined. The efficacy and appropriateness of RUSBoost are demonstrated as opposed to the commonly used classification models. The algorithmic performance of our sleep apnea identification scheme is also evaluated against existing detection algorithms in the literature. Experimental outcomes manifest that our sleep apnea identification scheme performs better than the existing works in sensitivity, specificity, and accuracy. It can be anticipated that owing to its use of only one channel of ECG signal, the proposed method will be ideal for device implementation, eliminate the onus of clinicians of analyzing a large bulk of data manually, and expedite sleep apnea diagnosis. HighlightsA single lead ECG based automated sleep apnea screening method is proposed.A novel signal processing technique, namely TQWT is employed.We introduce RUSBoost to classify sleep apnea for the first time.Efficacy of the method is confirmed by statistical and graphical analyses.The performance of the proposed scheme, compared to the existing ones is promising.

Effects of pre-incubation egg storage on embryonic functions and growth

The effect of pre-incubation storage on physiological functions of chick embryos during the last half of incubation and the relationship to embryonic growth were studied. In the first experiment, eggs were stored for 20 or 30 days, respectively, and the developmental patterns of oxygen consumption (MO2), heart rate (fH) and O2 pulse of individual embryos were examined. The MO2 of stored eggs increased at significantly lower rate than the control between day 12 and 17 of incubation, and the stored eggs had a significantly lower plateau MO2 between day 17 and 19. The decrease in MO2 of some stored eggs was correlated with late incubation mortality. Pre-incubation storage also resulted in significant changes in the developmental patterns of fH and O2 pulse. In the second experiment, the eggs were stored for 10 and 20 days, respectively, to also examine the effect of short-term storage on the development of MO2 in relation to embryo growth. Pre-incubation storage for 10 days had no significant effect, but 20 days storage shifted developmental patterns of MO2, wet mass, dry mass and embryo water fraction to the right of the control. Furthermore, MO2 was significantly lower than expected on the basis of embryo mass after day 17 of incubation. Prolonged pre-incubation storage caused not only a rightward shift in MO2 pattern due to retarded growth, but also severe depression of MO2 during the last stages of prenatal development.

Two-dimensional cardiogenic ballistic movements of avian eggs

INCUBATED AVIAN eggs move minutely due to cardiac contractions and blood ejection from the heart of the embryo. The ballistic movement imparted by the embryonic heartbeat is designated as the ballistocardiogram (BCG) of the egg. Cain et at. (1967) reported measurements of BCG of chicken eggs using an ultra-sensitive piezoelectric transducer which was a component of a micrometeoroid momentum transducer (CArN et al., 1967). We have also developed several methods and techniques to measure the eggs BCG, with an interest in the detection of such minute biological movements and comparisons between measurements (SuzuIr et aL, t989; TAZAWA et aL, 1989a; 1993; HASHIMOTO et al., 1991). During the measurement of the BCG waves, there were frequent large vibrations of the egg attributed to embryonic posture changes, and the patterns changed. In addition, changes of BCG waves occurred frequently without large postttre changes, indicating small somatic movements of the embryo. It is of interest to us to see how eggs move two-dimensionally due to the heartbeats of the embryos wi~ the lapse of time. The present study reports the two-dimensional display of eardiogenic ballistic movements in chicken eggs and shows examples of measurements taken from eggs which were placed in either a horizontal or uptight position (blunt end up). A laser is focused by an illuminating lens on a region of the eggshell at a distance of 40 ram. Reflected fight is then picked up through a detecting lens placed at an angle of 34 ~ from the illuminating light axis, producing a spot on the LPDE. Depending on the displacement changes of the egg, the spot moves on the LPDE, producing electrical signals that are processed by the controller (LC-200t), which quantifies the displacement. The resolving power is 0.5 ~ and the output is adjusted to produce 1 mV for 1/~m displacement. The outputs from the LDM are amplified by 60 dB and bandpass-filtered to eliminate low and high frequency noise and power-line interference. The bandpass filter has a 3 dB bandwidth of 0.1-25 Hz with attenuation slope of 24 dB per octave. The signal is digitised by a 12-bit A/D convertor at a sampling rate of 200 Hz mad recorded on a microcomputer.

Identification of Sleep Apnea from Single-Lead Electrocardiogram

An automatic sleep apnea detection algorithm is essential not only for alleviating the onus of physicians of analyzing a high volume of data but also for making a portable sleep quality evaluation device feasible. Most prior studies are either multi-lead based or yield poor accuracy which hinder the aforementioned goals. In this work, we propound a statistical and spectral feature based method for automated apnea detection from singlelead electrocardiogram. The efficacy of the selected features is demonstrated by intuitive, graphical and statistical validation. RUSBoost is introduced for sleep apnea classification. Again, most of the existing works focus on the feature extraction part. The effect of various classification models is poorly studied. Besides propounding an automated sleep apnea screening method, we study the performance of eight well-know classifiers for our feature extraction scheme. The optimal choices of parameters for RUSBoost are also inspected. The results of our experiments manifest that the proposed method outperforms the state-of-theart ones in terms of accuracy.

Investigation of the nonlinearity in the heart rate dynamics

This paper presents the application of a neural network to predict human heart rate. Electrocardiograms were measured from 5 healthy adult human subjects and 5 data sets were constructed calculating instantaneous heart rate from the measured signal. The nonlinear radial basis function neural network was applied to have a one step ahead prediction of the 1000 point heart rate. The results of the prediction are compared to that obtained by a linear autoregressive model. The results show that the neural network performs better than the autoregressive model in predicting heart rate for 2 data sets while for the other 3 data sets the performance of the two models is statistically similar. This indicates that the heart rate may be controlled nonlinearly by the autonomic nervous system.

Introduction to a novel wavelet

Wavelet is a basis function used in wavelet transformation. The wavelet transformation is a strong tool, and widely used in the signal processing purposes. However, the universality of application of a particular wavelet is still restricted. Here, we have presented a novel wavelet that can be used with better performance than many of the existing wavelet family members. The new wavelet satisfies the basic properties of wavelet, and a summary of the proposed wavelet shows that it is symmetric and mesokurtic with zero mean, contains large number of vanishing moments, and can be efficiently used both for continuous and discrete type wavelet transformations with exact reconstruction of signals. The new wavelet demonstrates better performance in the analysis of biological data sets, and a similar improvement is expected when analysing many other statistical data in different sectors.

A robust method for still image compression using dynamically constructive neural network

A dynamically constructive neural network (DCNN) is proposed for still image compression. The main feature of the proposed dynamical construction is its robustness to input-to-hidden and hidden-to-output link failure. A wavelet transform based sub-image block classification technique is also proposed for partitioning training images into image clusters. Each cluster is used as a training set for training a particular DCNN. This ensures the generalization capability of DCNNs. Computer simulation results demonstrate superiority of the proposed scheme in terms of peak signal to noise ratio and robustness as compared to that of other recent methods.