This work describes our experiences in discovering association rules in medical data to predict heart disease. We focus on two aspects of this work: mapping medical data to a transaction format suitable for mining association rules, and identifying useful constraints. Based on these aspects we introduce an improved algorithm to discover constrained association rules. We present an experimental section explaining several interesting discovered rules.

Clustering is a data mining problem that has received significant attention by the database community. Data set size, dimensionality and sparsity have been identified as aspects that make clustering more difficult. The article introduces a fast algorithm to cluster large binary data sets where data points have high dimensionality and most of their coordinates are zero. This is the case with basket data transactions containing items, that can be represented as sparse binary vectors with very high dimensionality. An experimental section shows performance, advantages and limitations of the proposed approach.

Assigning anatomic labels to coronary arteries in X-ray angiograms is an important task in medical imaging, motivated by the desire to standardize the assessment of coronary artery disease and to facilitate the three-dimensional (3-D) reconstruction and visualization of the coronary vasculature. However, automatic labeling poses a number of significant challenges, including the presence of noise, artifacts, competing structures, misleading visual cues, and other difficulties associated with a dynamic and inherently complex structure. We have developed a model-guided approach that addresses these challenges and automatically labels the vascular structure in coronary angiographic images. The approach consists of two models: 1) a symbolic model, represented through a directed acyclic graph, that captures vascular tree hierarchies and branch interrelationships and 2) a generalized 3-D model that captures spatial and geometric relationships. Importantly, the approach detects ambiguities (such as vessel overlaps) that may be found in a frame of a ciné sequence, and resolves these ambiguities by considering the information derived from other (unambiguous) frames in the temporal sequence, employing dynamic programming methods to match the image features found in the different (ambiguous and unambiguous) frames. This paper presents this model-guided labeling algorithm and discusses the experimental results obtained from implementing and applying the resulting labeling system to a variety of clinical images. The results indicate the feasibility of achieving robust and consistently accurate image labeling through this model-guided, temporal disambiguation method.

This paper describes a new program in Digital Image Systems Science and Engineering that has just been initiated with the assistance of a grant from the Hewlett-Packard Company. This new initiative seeks to provide a coherent educational program in digital image systems that is readily accessible to students in electrical and computer engineering, computer science, and other technical fields. A key feature of the program is the extensive use of computer technology to provide hands-on laboratory experience, classroom demonstrations, and on-line educational support materials.

The orientation, or pose, of an object is a fundamental property that helps to define the geometrical relationship between the object and its environment. In addition, knowledge of object orientation can also facilitate interpretive and decision-making tasks in a variety of practical domains, including industrial, meteorological, and medical applications. Determining object pose, however, remains an open research question in the fields of graphics and visualization. This article describes a novel yet intuitively simple approach, which we call topological goniometry, to directly determine the pose of a three-dimensional object from 3D data. The topology of interest is that of two-sided surfaces in a three-manifold, and includes objects whose shaped are unaffected by elastic transformations. Algorithmically, topological goniometry is composed of the following major steps. The first analyzes the global topology in order to generate a distribution of 3D coordinate triplets in the proximity of the desired pose axis. Using this set of 3D points, that second step then invokes a “3D Walk” algorithm that considers the local topology to produce a generalized curve representing an estimate of the object's axis of pose. The resultant pose axis is thus not constrained to lie along a straight line but can be generalized 3D curve. The methods are illustrated with a variety of synthetically created models that exhibit duct-like shapes, and are further tested by introducting noise as well as deformations to these models. The approach is also applied to a number of real discrete data obtained from meteorological and medical domains. The results suggest that the appproach is applicable to both real and synthetic datasets and is shown to be robust, computationally efficient, and applicable to a variety of problems. The approach can incorporate context- or application-dependent information about the object of interest by using a set of constraints that guide the process of orientation determination. This article describes the approach, its implementation, and the results obtained with numerous applications.

Presents a new method to determine the orientation or pose of the left ventricle (LV) of the heart from cardiac SPECT (single photon emission computed tomography) data. This proposed approach offers an accurate, fast, and robust delineation of the LV long-axis. The location and shape of the generated long-axis can then be utilized to define automatically the tomographic slices for enhanced visualization and quantification of the clinical data. The methodology is broadly composed of two main steps: (1) volume segmentation of cardiac SPECT data; and (2) topological goniometry, a novel approach incorporating volume visualization and computer graphics ideas to determine the overall shape of 3-D objects. The outcome of the algorithm is a 3-D curve representing the overall pose of the LV long-axis. Experimental results on both phantom and clinical data (50 technetium-99m and 74 thallium-201) are presented. An interactive graphical interface to visualize the volume (3-D) data, the left ventricle, and its pose is an integral part of the overall methodology. This technique is completely data driven and expeditious, making it viable for routine clinical use.

This paper presents numerous pattern recognition techniques which can be applied to radar data. In particular, nearest neighbor and linear discriminant function algorithms are discussed, as well as the use of different sets of features to represent the data. In addition, some of the most common applications are also discussed, including meteorological analyses and military applications.