William G. Wee

Review: A comparative study of deformable contour methods on medical image segmentation

A comparative study to review eight different deformable contour methods (DCMs) of snakes and level set methods applied to the medical image segmentation is presented. These DCMs are now applied extensively in industrial and medical image applications. The segmentation task that is required for biomedical applications is usually not simple. Critical issues for any practical application of DCMs include complex procedures, multiple parameter selection, and sensitive initial contour location. Guidance on the usage of these methods will be helpful for users, especially those unfamiliar with DCMs, to select suitable approaches in different conditions. This study is to provide such guidance by addressing the critical considerations on a common image test set. The test set of selected images offers different and typical difficult problems encountered in biomedical image segmentation. The studied DCMs are compared using both qualitative and quantitative measures and the comparative results highlight both the strengths and limitations of these methods. The lessons learned from this medical segmentation comparison can also be translated to other image segmentation domains.

A Formulation of Fuzzy Automata and Its Application as a Model of Learning Systems

Based on the concept of fuzzy sets defined by Zadeh, a class of fuzzy automata is formulated similar to Mealys formulation of finite automata. A fuzzy automaton behaves in a deterministic fashion. However, it has many properties similar to that of stochastic automata. Its application as a model of learning systems is discussed. A nonsupervised learning scheme in automatic control and pattern recognition is proposed with computer simulation results presented. An advantage of employing fuzzy automaton as a learning model is its simplicity in design and computation.

Neighboring gray level dependence matrix for texture classification

Abstract A new approach, neighboring gray level dependence matrix (NGLDM), for texture classification is presented. The major properties of this approach are as follows: (a) texture features can be easily computed; (b) they are essentially invariant under spatial rotation; (c) they are invariant under linear gray level transformation and can be made insensitive to monotonic gray level transformation. These properties have enhanced the practical applications of the texture features. The accuracies of the classification are comparable with those found in the literature.

Automated Vertebra Detection and Segmentation from the Whole Spine MR Images

Our algorithm contains two major steps: the intervertebral disk localization step, and the vertebra detection and segmentation step. In the first step, we apply a model-based searching method to approximately locate all the intervertebral disk clues between adjacent vertebrae of the whole spine and the best slice selection. A new approach using an intensity profile on a polynomial function for fitting all these disk clues on the best slice is then used to refine the disk search process. Vertebra centers are detected, and initial boundaries are extracted in the second step. The initial test of the algorithm on the five sets of 7 sagittal slices locates all 23 intervertebral disk centers for the best slice of all five sets. For the evaluation of the boundary extraction of 22 vertebrae, our algorithm successfully locates 100%, 96.6%, 93.2%, 95.5%, 87.5% vertebra corners in image set No.1, 2, 3, 4, and 5, respectively

Knowledge-based image analysis for automated boundary extraction of transesophageal echocardiographic left-ventricular images

A system for automatically determining the contour of the left ventricle (LV) and its bounded area, from transesophageal echocardiographic (TEE) images is presented. It uses knowledge of both heart anatomy and echocardiographic imaging to guide the selection of image processing methodologies for thresholding, edge detection, and contour following and the center-based boundary-finding technique to extract the contour of the LV region. To speed up the processing a rectangular region of interest from a TEE picture is first isolated and then reduced to a coarse version, one-ninth original size. All processing steps, except the final contour edge extraction, are performed on this reduced image. New methods developed for automatic threshold selection, region segmentation, noise removal, and region center determination are described.

Generalized Inverse Approach to Adaptive Multiclass Pattern Classification

Abstract—In this paper a least-square approach to multiclass pattern classification is undertaken. The generalized inverse computation is used to furnish a quick solution to the problem of fixed training samples. The use of recursive on-line computation is also recommended. Experimental results are presented to illustrate the approach. Both deterministic and statistical interpretations have been given to the approach. The pattern classifier proposed by Chaplin and Levadi [1] and the adaptive pattern classifier proposed by Patterson and Womack [2] are special cases of this approach.

Deformable Contour Method: A Constrained Optimization Approach

In this paper, a class of deformable contour methods using a constrained optimization approach of minimizing a contour energy function satisfying an interior homogeneity constraint is proposed. The class is defined by any positive potential function describing the contour interior characterization. An evolutionary strategy is used to derive the algorithm. A similarity threshold Tv can be used to determine the interior size and shape of the contour. Sensitivity and significance of Tv and σ (a spreadness measure) are also discussed and shown. Experiments on noisy images and the convergence to a minimum energy gap contour are included. The developed method has been applied to a variety of medical images from CT abdominal section, MRI image slices of brain, brain tumor, a pig heart ultrasound image sequence to visual blood cell images. As the results show, the algorithm can be adapted to a broad range of medical images containing objects with vague, complex and/or irregular shape boundary, inhomogeneous and noisy interior, and contour with small gaps.

Evaluation of mammographic calcifications using a computer program.

A computer pattern recognition program was developed to evaluate calcifications seen on mammograms. When the approximate horizontal length, average internal gray level, and contrast were used as criteria, 84.3% of the lesions studied were correctly identified as malignant or benign. These results show that a computer program can help evaluate calcifications of the breast.

Wastewater Treatment Systems from Case–Based Reasoning

Case-Based Reasoning (CBR) is one of the emerging paradigms for designing intelligent systems. Preliminary studies indicate that the area is ripe for theoretical advances and innovative applications. Heuristic search is one of the most widely used techniques for obtaining optimal solutions to many real-world problems. We formulated the design of wastewater treatment systems as a heuristic search problem. In this article we identify some necessary properties of the heuristic search problems to be solved in the CBR paradigm. We designed a CBR system based on these observations and performed several experiments with the wastewater treatment problem. We compare the performance of the CBR system with the A* search algorithm.

On Methods of Three-Dimensional Reconstruction from a Set of Radioisotope Scintigrams

The three-dimensional reconstruction of the radioisotope source distribution of coordinated radioisotope scintigrams is complicated by the point spreading function created by the imaging system (conical imaging) and the attenuation of gamma rays through the body. A new reconstruction method is proposed by converting the complicated existing reconstruction problem into the classical problem of weighted projection in the Fourier complex space upon which a modified additive algebraic reconstruction technique (ART) can be employed. Mathematical derivation, reconstruction algorithms, and illustrative computer results are included.