Francis H. Y. Chan

Fuzzy EMG classification for prosthesis control

This paper proposes a fuzzy approach to classify single-site electromyograph (EMG) signals for multifunctional prosthesis control. While the classification problem is the focus of this paper, the ultimate goal is to improve myoelectric system control performance, and classification is an essential step in the control. Time segmented features are fed to a fuzzy system for training and classification. In order to obtain acceptable training speed and realistic fuzzy system structure, these features are clustered without supervision using the Basic Isodata algorithm at the beginning of the training phase, and the clustering results are used in initializing the fuzzy system parameters. Afterwards, fuzzy rules in the system are trained with the back-propagation algorithm. The fuzzy approach was compared with an artificial neural network (ANN) method on four subjects, and very similar classification results were obtained. It is superior to the latter in at least three points: slightly higher recognition rate; insensitivity to overtraining; and consistent outputs demonstrating higher reliability. Some potential advantages of the fuzzy approach over the ANN approach are also discussed.

Adaptive thresholding by variational method

When using thresholding method to segment an image, a fixed threshold is not suitable if the background is uneven. Here, we propose a new adaptive thresholding method using variational theory. The method requires only one parameter to be selected and the adaptive threshold surface can be found automatically from the original image.

Image Contrast Enhancement by Constrained Local Histogram Equalization

Histogram equalization is a widely used image contrast enhancement method. While global histogram equalization enhances the contrast of the whole image, local histogram equalization can enhance many image details by taking different transformation of the same gray level at different places in the original image. However, the local histogram equalization process often results in unacceptable modification of the original image appearance. In this paper, a constrained local histogram equalization method is proposed to balance the conflicting requirements: enhancement of the image details and the maintenance of the overall image appearance. Our method uses the variational form of histogram equalization so that a constraint condition, which forces the local gray level transformations to change continuously in the spatial domain, can be introduced into the equalization process. Experimental results of different kinds of images show the effect of our method.

A matrix-pencil approach to blind separation of colored nonstationary signals

For many signal sources such as speech with distinct, nonwhite power spectral densities, second-order statistics of the received signal mixture can be exploited for signal separation. Without knowledge of the noise correlation matrix, we propose a simple and yet effective signal extraction method for signal source separation under unknown temporally white noise. This new and unbiased signal extractor is derived from the matrix pencil formed between output autocorrelation matrices at different delays. Based on the matrix pencil, an ESPRIT-type algorithm is derived to get an optimal solution in the least square sense. Our method is well suited for systems with colored sensor noises and for nonstationary signals.

Adaptive filtering of evoked potentials with radial-basis-function neural network prefilter

Evoked potentials (EPs) are time-varying signals typically buried in relatively large background noise. To extract the EP more effectively from noise, we had previously developed an approach using an adaptive signal enhancer (ASE) (Chen et al., 1995). ASE requires a proper reference input signal for its optimal performance. Ensemble- and moving window-averages were formerly used with good results. In this paper, we present a new method to provide even more effective reference inputs for the ASE. Specifically, a Gaussian radial basis function neural network (RBFNN) was used to preprocess raw EP signals before serving as the reference input. Since the RBFNN has built-in nonlinear activation functions that enable it to closely fit any function mapping, the output of RBFNN can effectively track the signal variations of EP. Results confirmed the superior performance of ASE with RBFNN over the previous method.

Detection of brainstem auditory evoked potential by adaptive filtering.

A method of detecting brainstem auditory evoked potential (BAEP) using adaptive signal enhancement (ASE) is proposed and tested in humans and cats. The ASE in this system estimates the signal component of the primary input, which is correlated with the reference input to the adaptive filter. The reference input is carefully designed to make an optimal and rapid estimation of the signal corrupted with noise, such as ongoing EEG. With a good choice of reference input, it is possible to track the variability of BAEP efficiently and rapidly. Moreover, the number of repetitions required could be markedly reduced and the result of the system is superior to that of ensemble averaging (EA). To detect BAEP in cats, only 30 ensemble averages are needed to obtain a reasonable reference input to the adaptive filter, and, for humans, 350–750 ensemble averages are sufficient for a satisfactory result. Using the LMS adaptive algorithm, individual BAEP can be obtained in real-time.

Color image denoising with wavelet thresholding based on human visual system model

Recent research in transform-based image denoising has focused on the wavelet transform due to its superior performance over other transform. Performance is often measured solely in terms of PSNR and denoising algorithms are optimized for this quantitative metric. The performance in terms of subjective quality is typically not evaluated. Moreover, human visual system (HVS) is often not incorporated into denoising algorithm. This paper presents a new approach to color image denoising taking into consideration HVS model. The denoising process takes place in the wavelet transform domain. A Contrast Sensitivity Function (CSF) implementation is employed in the subband of wavelet domain based on an invariant single factor weighting and noise masking is adopted in succession. Significant improvement is reported in the experimental results in terms of perceptual error metrics and visual effect.

Real-time data-reusing adaptive learning of a radial basis function network for tracking evoked potentials

Tracking variations in both the latency and amplitude of evoked potential (EP) is important in quantifying properties of the nervous system. Adaptive filtering is a powerful tool for tracking such variations. In this paper, a data-reusing nonlinear adaptive filtering method, based on a radial basis function network (RBFN), is implemented to estimate EP. The RBFN consists of an input layer of source nodes, a single hidden layer of nonlinear processing units and an output layer of linear weights. It has built-in nonlinear activation functions that allow learning of function mappings. Moreover, it produces satisfactory estimates of signals against a background noise without a priori knowledge of the signal, provided that the signal and noise are independent. In clinical situations where EP responses change rapidly, the convergence rate of the algorithm becomes a critical factor. A carefully designed data-reusing RBFN can accelerate the convergence rate markedly and, thus, enhance its performance. Both theoretical analysis and simulation results support the improved performance of our new algorithm.

New Bayesian Discriminator for Detection of Atrial Tachyarrhythmias

Background—Accurate, rapid detection of atrial tachyarrhythmias has important implications in the use of implantable devices for treatment of cardiac arrhythmia. Currently available detection algorithms for atrial tachyarrhythmias, which use the single-index method, have limited sensitivity and specificity. Methods and Results—In this study, we evaluated the performance of a new Bayesian discriminator algorithm in the detection of atrial fibrillation (AF), atrial flutter (AFL), and sinus rhythm (SR). Bipolar recording of 364 rhythms (AF=156, AFL=88, SR=120) at the high right atrium were collected from 20 patients who underwent electrophysiological procedures. After initial signal processing, a column vector of 5 features for each rhythm were established, based on the regularity, rate, energy distribution, percent time of quiet interval, and baseline reaching of the rectified autocorrelation coefficient functions. Rhythm identification was obtained by use of Bayes decision rule and assumption of Gaussian distribution. For the new Bayesian discriminator, the overall sensitivity for detection of SR, AF, and AFL was 97%, 97%, and 94%, respectively; and the overall specificity for detection of SR, AF, and AFL was 98%, 98%, and 99%, respectively. The overall accuracy of detection of SR, AF, and AFL was 98%, 97% and 98%, respectively. Furthermore, sensitivity, specificity, and accuracy of this algorithm were not affected by a range of white Gaussian noises with different intensities. Conclusions—This new Bayesian discriminator algorithm, based on Bayes decision of multiple features of atrial electrograms, allows rapid on-line and accurate (98%) detection of AF with robust anti-noise performance.

A tracing evoked potential estimator

The paper presents an adaptive Gaussian radial basis function neural network (RBFNN) for rapid estimation of evoked potential (EP). Usually, a recorded EP is severely contaminated by background ongoing activities of the brain. Many approaches have been reported to enhance the signal-to-noise ratio (SNR) of the recorded signal. However, non-linear methods are seldom explored due to their complexity and the fact that the non-linear characteristics of the signal are generally hard to determine. An RBFNN possesses built-in non-linear activation functions that enable the neural network to learn any function mapping. An RBFNN was carefully designed to model the EP signal. It has the advantage of being linear-in-parameter, thus a conventional adaptive method can efficiently estimate its parameters. The proposed algorithm is simple so that its convergence behaviour and performance in signal-to-noise ratio (SNR) improvement can be mathematically derived. A series of experiments carried out on simulated and human test responses confirmed the superior performance of the method. In a simulation experiment, an RBFNN having 15 hidden nodes was trained to approximate human visual EP (VEP). For detecting gene rate=0.005) speeded up the estimation remarkably by using only 80 ensembles to achieve a result comparable to that obtained by averaging 1000 ensembles.