Robert D. Adams

A novel shape feature to classify microcalcifications

The cited references claim that microcalcifications from many benign regions are all round or oval. The detection of at least one roughly shaped microcalcification in a suspicious region could be an early sign of potentially developing malignant cancer. This paper proposes a shape analysis method to aid radiologists in classifying regions of interest that are difficult to diagnosis. A region growing and a gradient vector flow methods are used to obtain an ordered set of contour points of each microcalcification. A three level wavelet transform frequency analysis provides a band pass approximation of the normalized distance signature. A novel metric derived from the normalized distance signature is proposed to quantify the roughness of a microcalcification. An experiment using a large dataset is used to evaluate the robustness of the proposed roughness metric against several published shape features.

Analysis of Stress in speech using adaptive Empirical Mode Decomposition

Stress in human speech can be detected by various methods know as Voice Stress Analysis (VSA). The detection is accomplished by measuring the frequency shift of a microtremor normally residing in the frequency range of 8 to 12 Hz when not stressed. Conventional detection methods include Fast Fourier Transform (FFT) or McQuiston-Ford algorithm. This paper presents a new method called Adaptive Empirical Mode Decomposition (AEMD) applied to voice stress detection. Because AEMD in essence is a time-frequency analysis method, it is possible to use this method for real-time voice stress detection.

A novel image edge detection method using Linear Prediction

Traditionally, Linear Prediction is used to predict future values of a signal using past values. The goal is to minimize prediction errors. In this paper, we propose a novel method of utilizing prediction errors to extract edges of images. In this method, smooth prediction errors are minimized while steep changes (larger errors) are amplified. Therefore, when applied to image edge detection, edge information can be accurately extracted. The proposed method is compared with predominant methods such as Sobel and Canny methods. While there is no mathematical proof that the proposed method outperforms predominant methods, however, examples presented in this paper may suggest that the proposed method may perform better for certain applications.

Detection of involuntary human hand motions using Empirical Mode Decomposition and Hilbert-Huang Transform

Involuntary human hand motions, or tremors, are normally regarded as a non-stationary process. Traditional analysis methods approximate tremor signals as stationary processes. In this paper, we present a novel tremor detection method using Empirical Mode Decomposition (EMD) and Hilbert-Huang Transform (HHT). The results are expected to be helpful for real-time tremor suppression.

A novel facial expression recognition method using fast BEMD based edge detection

Traditionally, human facial expressions are recognized using standard images. These methods of recognition require subjective expertise and high computational costs. This article presents a novel facial expression recognition method using edge maps generated from standard images that can significantly improve computational efficiency. The edge maps are generated from a novel Bi-dimensional Empirical Mode Decomposition (BEMD) based edge detection method. In this paper, the BEMD edge detection algorithm is discussed, the facial expression decision metrics are developed, and detection results of facial expression databases are presented. The success rates of recognition suggest that the proposed method be a potentially efficient method for human facial expression detection and other similar recognition applications.

A microcalcification enhancement method for mammogram images

This paper proposed a novel method to enhance mammogram images used in detecting early signs of breast cancer. The proposed method uses a three level pyramid decomposition scheme that applies the squeeze box filter (SBF) instead of low-pass filtering. A previously proposed non-linear local enhancement technique is applied to the difference images to contrast enhance the structural details of a mammogram image. The enhanced mammogram image is reconstructed by adding the enhanced difference images to the original SBF processed image. The experiment and preliminary results reported in this paper provide evidence of beneficial enhancements provided by the proposed method on mammographic images.

Simulating micro-robots to find a point of interest under noise and with limited communication using Particle Swarm Optimization

This paper presents the simulation results of a swarm of micro-robots collaborating to find a point of interest in 2D space. Guided by a fitness function, the Particle Swarm Optimization (PSO) algorithm is highly efficient to explore the solution space and find such an optimum. However, in real-world scenarios in which the particles are micro-robots, there are practical constraints. The two most significant constraints are: (1) given communication and measurement noise, the fitness function evaluation will be noisy, (2) given the limited communication range of micro-robots, broadcasting the global best solution is too expensive. A neighborhood PSO (NPSO) algorithm is proposed to replace the global best by the neighborhood best. Different applications call for different fitness functions, and three benchmark functions, representing three typical scenarios, are examined: (1) a unimodal and symmetric scenario with only one global optimum, (2) a multi-modal scenario with one global optimum but many local optima, and (3) a uni-model but asymmetric scenario. For each fitness function, simulations on the effects of the two aforementioned constraints, individually or combined, are carried out. The results demonstrate that PSO is tolerant to noise up to certain level and NPSO is a practical adaptation to implement swarm intelligence in swarm robotics.

Green's dyadics for a bounded lossy medium

Dyadic Greens function for an arbitrarily shaped lossy dielectric object is presented. It is derived based on surface integral equations. Application of the presented dyadic Greens function allows a detailed study of radiation pattern and power loss of an antenna in the presence of a lossy medium. Numerical computations are implemented for a dipole in the vicinity of a lossy dielectric sphere, the radiation pattern and power loss as a function of loss tangent are presented in graphs.