F. K. Lam

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.

PERFORMANCE ANALYSIS FOR A CLASS OF ITERATIVE IMAGE THRESHOLDING ALGORITHMS

A performance analysis procedure that analyses the properties of a class of iterative image thresholding algorithms is described. The image under consideration is modeled as consisting of two maximum-entropy primary images, each of which has a quasi-Gaussian probability density function. Three iterative thresholding algorithms identified to share a common iteration architecture are employed for thresholding 4595 synthetic images and 24 practical images. The average performance characteristics including accuracy, stability, speed and consistency are analysed and compared among the algorithms. Both analysis and practical thresholding results are presented.

Recognition of Occluded Objects

Pattern completion is a ubiquitous and critical component of visual recognition under natural conditions whereby we need to make inferences from partial information. Pattern recognition from sparse information is essential when objects are rendered under poor illumination or when they are significantly occluded. Here we provide an overview of the behavioral, physiological and computational studies of pattern completion. We argue that the computational mechanisms subserving recognition of heavily occluded objects rely on neural circuits with recurrent connectivity that are capable of interpreting incoming inputs in the context of prior knowledge.

Classification of partially occluded objects using 3-point matching and distance transformation

Abstract Recognition of partially occluded objects has been recently made possible by describing an object boundary with a sequence of local feature segments localized with control or dominant points, and then matching the representations with those of a set of known reference shapes. When part of the object boundary is being occluded, contiguous intact feature segments will be used for identifying the object and distorted elements will be rejected in the matching process. The main problem associated with this method is that the nature of occlusion may separate the object contour into discrete regions, resulting in the degeneration of an individual representation into small isolated clusters reflecting only weak evidence on the object identity. In this paper, a new scheme capable of combining the isolated clusters for object classification is presented. In essence, Curvature Guided Polygonal Approximation is employed for detecting the dominant points of the boundaries. A 3-point matching technique is developed for extracting cluster(s) of dominant points on an unknown boundary which matched against reference objects, and to link up the scattered clusters to form a complete representation. The chamfer 3 4 Distance Transform, together with a quantitative measurement on the boundary covered by the matched dominant points is employed for calculating the similarity between the unknown and the reference contours. The scheme has been successfully applied for recognizing multiple overlapped handtools in different sizes and orientations.

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.

Maximum segmented image information thresholding

Abstract Utilizing information theory and considering image segmentation from a communication perspective, the image segmentation process is interpreted as a data processing step that operates on a gray-scale image and produces a segmented image. It is shown that the segmented image contains a certain amount of information about the scene, which is defined assegmented image information(SII). It is proposed that the SII should be maximized when an image is thresholded, and this is known as themaximum segmented image information(MSII) thresholding criterion. The MSII thresholding criterion possesses better properties as compared with theminimum error(MINE) and theuniform error(UNFE) thresholding criteria. Based on the MSII thresholding criterion, an MSII thresholding algorithm is proposed for the thresholding of real images. The MSII thresholding algorithm is evaluated against several well-known thresholding algorithms. The good thresholding results of both synthetic and real images confirm the capabilities of the proposed MSII thresholding algorithm.

Evoked potential estimation using modified time-sequenced adaptive filter

A method called modified time-sequenced adaptive filtering (MTSAF) is applied to estimate evoked potential (EP) signals and track the temporal variations of EPs. The MTSAF consists of a set of adaptive filters (AFs), with each processing a time segment of EP data. After convergence, each AF reaches the best estimation of EP signals over its own time segment in terms of minimum mean squared error (MMSE). Numerical results of simulated and human EP data show that the MTSAF reaches better estimation of EPs than a conventional adaptive signal enhancer (ASE). With the MTSAF, the temporal variations of EPS across trials can be estimated to reveal more subtle variations of EPs, which may be of clinical value.

Maximum segmented-scene spatial entropy thresholding

The segmented-scene spatial entropy (SSE) is defined as the amount of information contained in the spatial structure of a segmented scene resulting from segmenting an image. An automatic, nonparametric, unsupervised thresholding algorithm that maximizes the SSE of an image is described, and this algorithm is known as the maximum segmented-scene spatial entropy (MSSE) thresholding algorithm. It is shown that the MSSE-thresholded image contains the maximum amount of information about the original scene and hence good thresholding results are warranted. Simulation and practical results are presented to illustrate the improvement in performance as compared to some other histogram-based thresholding algorithms.

Robust matching process: a dominant point approach

Abstract A robust local feature matching scheme for recognition of occluded objects is developed and reported in this letter. The proposed scheme is computation efficient and can handle the nonlinear distorted local features. A set of handtools is selected to evaluate the performance of the algorithm and the results are encouraging.

Wavelet analysis of embolic heart sound detected by precordial Doppler ultrasound during continuous venous air embolism in dogs

The spectrum of the embolic heart sounds (EHS) detected by precordial Doppler ultrasound has been previously characterized, but only on small volumes of venous air embolism (VAE).We sought to determine whether real-time wavelet analysis is useful in analyzing the signals of EHS and whether the embolic power of the EHS for larger volumes of air is proportionate to the volume of VAE that has been reported for small volumes of VAE. A series of small air boli (0.01, 0.02, 0.05, 0.07, 0.1, 0.15, 0.2, 0.3, 0.4, and 0.8 mL), followed by continuous infusion of larger volumes of air (0.8, 1.6, 2.4, 4.8, and 9.6 mL), was injected into the external jugular vein through a central catheter in seven pentobarbital-anesthetized dogs. We measured the spectrum of the Doppler heart sound (DHS) in a real-time manner by using wavelet analysis at different scales. Wavelet analysis at scale = 1 yielded satisfactory results in distinguishing abnormal EHS from normal DHS with high sensitivity (100%) and good positive predictive value (100%) compared with the conventional method, which requires an anesthesiologist to listen to the audio DHS signals in a real-time manner. There was a linear relationship (y = 1.08x + 7.89, r = 0.75, P < 0.001) between the cumulative embolic power of the EHS and the air volume introduced in the form of either bolus or continuous infusion. The 95% confidence intervals for slope and intercept were 0.89-1.27 and 7.65-8.13, respectively. Our results suggest that wavelet analysis is effective as a real-time monitor and that it is possible to distinguish larger volumes of air emboli based on previous injections of small volumes of air. Implications: The real-time wavelet analysis of the heart sound detected by precordial Doppler ultrasound may be useful in estimating larger volumes of air emboli based on previous injections of small volumes of air in anesthetized dogs. (Anesth Analg 1998;86:325-31)