Early detection of lung cancer is crucial as it significantly improves survival rates by facilitating timely and effective treatment. Lung cancer often begins as bronchial lesions developing along the airway walls. Bronchoscopy is the minimally invasive method of choice for detecting such lesions. Currently, three complementary bronchoscopic video modalities have been utilized for this purpose: white-light bronchoscopy (WLB), narrow-band imaging (NBI), and autofluorescence bronchoscopy (AFB). Unfortunately, current practice forces the clinician to manually examine each video source and later interactively correlate the results of these exams to make final lesion decisions. Because of the lack of effective tools for multimodal endoscopic video analysis, this proves to be an extremely time-consuming, error-prone process, making it impractical for common clinical use. To address this problem, we propose a multimodal video analysis and synchronization system that enables efficient analysis of multimodal bronchoscopic videos for early cancer lesion detection. The system provides methods for planning and guiding a straightforward multimodal airway exam through the major airways. Subsequent video processing methods then draw on deep-learning-based techniques to identify candidate single-mode bronchial lesions. Next, a synchronization/registration pipeline registers all bronchoscopic video data to a reference 3D airway tree model derived from a patient's X-ray computed tomography (CT) scan. This finally facilitates interactive graphical visualization and interaction with all processed multimodal data. Results with lung cancer patient studies indicate the system's promise for efficient, effective video analysis and lesion detection.

Bronchoscopy is a major step in lung cancer staging. To perform bronchoscopy, the physician uses a procedure plan, derived from a patient's 3D computed-tomography (CT) chest scan, to navigate the bronchoscope through the lung airways. Unfortunately, physicians vary greatly in their ability to perform bronchoscopy. As a result, image-guided bronchoscopy systems, drawing upon the concept of CT-based virtual bronchoscopy (VB), have been proposed. These systems attempt to register the bronchoscope's live position within the chest to a CT-based virtual chest space. Recent methods, which register the bronchoscopic video to CT-based endoluminal airway renderings, show promise but do not enable continuous real-time guidance. We present a CT-video registration method inspired by computer-vision innovations in the fields of image alignment and image-based rendering. In particular, motivated by the Lucas-Kanade algorithm, we propose an inverse-compositional framework built around a gradient-based optimization procedure. We next propose an implementation of the framework suitable for image-guided bronchoscopy. Laboratory tests, involving both single frames and continuous video sequences, demonstrate the robustness and accuracy of the method. Benchmark timing tests indicate that the method can run continuously at 300 frames/s, well beyond the real-time bronchoscopic video rate of 30 frames/s. This compares extremely favorably to the ≥ 1 s/frame speeds of other methods and indicates the method's potential for real-time continuous registration. A human phantom study confirms the method's efficacy for real-time guidance in a controlled setting, and, hence, points the way toward the first interactive CT-video registration approach for image-guided bronchoscopy. Along this line, we demonstrate the method's efficacy in a complete guidance system by presenting a clinical study involving lung cancer patients.

A vital task in the planning of peripheral bronchoscopy is the segmentation of the airway tree from a 3-D multidetector computed tomography chest scan. Unfortunately, existing methods typically do not sufficiently extract the necessary peripheral airways needed to plan a procedure. We present a robust method that draws upon both local and global information. The method begins with a conservative segmentation of the major airways. Follow-on stages then exhaustively search for additional candidate airway locations. Finally, a graph-based optimization method counterbalances both the benefit and cost of retaining candidate airway locations for the final segmentation. Results demonstrate that the proposed method typically extracts 2-3 more generations of airways than several other methods, and that the extracted airway trees enable image-guided bronchoscopy deeper into the human lung periphery than past studies.

The assessment of lung cancer involves off-line threedimensional (3D) computed-tomography (CT) image assessment followed by live bronchoscopy. While standard bronchoscopy provides live video of information inside the lung airways, it does not give any information outside the airways. This results in a low procedure success rate. The success rate can be improved if the physician receives 3D CT-based image guidance. We describe a fast robust method that provides 3D CT-based image guidance during live bronchoscopy. The method enables a continuous registration between the 3D CT image space and the real bronchoscopic video. At a top level, it involves an interleaving of continuous video tracking and CT-video fine registration. During video tracking, the bronchoscope’s 3D motion is estimated using a point-based 3D-2D pose estimation method. Registration is performed via a warpingbased Gauss-Newton method that uses a normalized crosscorrelation- based cost for measuring similarity between the bronchoscopic video and CT image data. The method operates at a real-time rate of 10 frames per second, which is over an order of magnitude faster than past approaches. This real-time performance enables live guidance during a procedure. The method is incorporated into a computerbased system for image-guided bronchoscopy and has been applied to human lung-cancer patients. Results are presented for a phantom and lung-cancer patients. Keywords: 3D medical imaging, video tracking, image registration, virtual endoscopy, image-guided surgery

Multidetector computed-tomography (MDCT) scanners provide large high-resolution three-dimensional (3-D) images of the chest. MDCT scanning, when used in tandem with bronchoscopy, provides a state-of-the-art approach for lung-cancer assessment. We have been building and validating a lung-cancer assessment system, which enables virtual-bronchoscopic 3-D MDCT image analysis and follow-on image-guided bronchoscopy. A suitable path planning method is needed, however, for using this system. We describe a rapid, robust method for computing a set of 3-D airway-tree paths from MDCT images. The method first defines the skeleton of a given segmented 3-D chest image and then performs a multistage refinement of the skeleton to arrive at a final tree structure. The tree consists of a series of paths and branch structural data, suitable for quantitative airway analysis and smooth virtual navigation. A comparison of the method to a previously devised path-planning approach, using a set of human MDCT images, illustrates the efficacy of the method. Results are also presented for human lung-cancer assessment and the guidance of bronchoscopy.

Virtual endoscopy (VE) is an itegration of medical imaging and virtual reality. VE gives a computer-based alternative to standard radiological image viewing and to traditional fiberoptic endoscopy for examining the interior structures of human organs. It has evolved rapidly in the past decade, as high-resolution helical spiral computed tomography (CT) has evolved for clinical use. A total of 22 submissions were received, covering modalities of virtual colonoscopy, virtual bronchoscopy, and virtual angioscopy, with technical focus on image segmentation, registration, dynamical modeling, flight path planning, fast volumetric rendering, interactive and active navigation, and CAD.

Of the many proposed image segmentation methods, region growing has been one of the most popular. Research on region growing, however, has focused primarily on the design of feature measures and on growing and merging criteria. Most of these methods have an inherent dependence on the order in which the points and regions are examined. This weakness implies that a desired segmented result is sensitive to the selection of the initial growing points. We define a set of theoretical criteria for a subclass of region-growing algorithms that are insensitive to the selection of the initial growing points. This class of algorithms, referred to as symmetric region growing algorithms, leads to a single-pass region-growing algorithm applicable to any dimensionality of images. Furthermore, they lead to region-growing algorithms that are both memory- and computation-efficient. Results illustrate the method's efficiency and its application to 3D medical image segmentation.

Bronchoscopic needle biopsy is a common step for early lung-cancer detection. This procedure uses two steps: (1) 3D computed-tomography (CT) chest image analysis, to choose a biopsy site; (2) live bronchoscopy, to perform the biopsy. CT-based virtual endoscopic analysis can improve the results of biopsies, yet errors can still occur. We describe a procedure to combine the endoscopic video tracking (the "real" world) and CT-based virtual endoscopic registration (the "virtual" world). By bringing both sources of information together, a more robust surgical guidance system is realizable. Both the endoscope's video and the thoracic CT scan are used as data sources in the tracking. An optical flow algorithm estimates the endoscope motion between successive video frames. The virtual CT rendering creates a range map for the optical flow equation. This simplifies the endoscope movement calculation into a straightforward linear system. We demonstrate this method for a phantom human airway-tree example.

Accurate definition of a three-dimensional (3D) branching tree structures from large 3D medical images, has become a common problem over several imaging modalities. For example, 3D computed-tomography (CT) images of the airways as well as micro-CT images of vasculature employ new paradigms that require detailed definitions of tree structures. Manual definition is impossible for images of this scale. Existing techniques lack features necessary for navigational or quantitative analysis or fail to recover paths. We present a generic tree analysis technique that is applicable to arbitrary segmented images of branching structures for both of these purposes. The method has been tested on over 30 human, animal, phantom, and micro-CT images. An initial comparison with a previous method demonstrates its ability to capture more paths, especially in peripheral regions.

A system for virtual-endoscopic 3D medical image assessment and follow-on live endoscopy is described. The development and results focus on 3D CT images of the chest and lung-cancer assessment. Two stages are involved in the assessment. In Stage-1, the physician uses a series of tools for visualization and 3D central-axes analysis to build a guidance plan (a "case study"). Next, during Stage-2 endoscopy, the physician links the computer system and case study to the endoscope. The endoscopic video is registered to 3D virtual image data to give the physician "augmented reality" information to better perform the endoscopy. Results illustrate the system's utility for animal and human cases.

We present a multi-modal image registration technique that matches virtual surface renderings, derived from 3D medical data, to corresponding endoscopic video. Our efforts focus on the chest. The views of the two image sources arise from inside the major bronchial airways, as shown in 3D CT chest images and live endoscopic video. Our technique is part of a large system for two-stage assessment of lung cancer. Stage-1 involves 3D CT assessment, which gives planning data for endoscopy. Stage-2 involves live endoscopy, supplemented by CT-based guidance. The registration technique is a critical component of this assessment. The technique draws upon the concept of normalized mutual information (NMI) and works in near real time on standard PCs. An optimization procedure iteratively adjusts the virtual viewpoint rendering until it matches a given video frame's viewpoint. The optimization technique ends when the NMI is maximized. Results are given for phantom and human cases.

Modern video-based endoscopes offer physicians a wide-angle field of view (FOV) for minimally invasive procedures. Unfortunately, inherent barrel distortion prevents accurate perception of range. This makes measurement and distance judgment difficult and causes difficulties in emerging applications, such as virtual guidance of endoscopic procedures. Such distortion also arises in other wide FOV camera circumstances. This paper presents a distortion-correction technique that can automatically calculate correction parameters, without precise knowledge of horizontal and vertical orientation. The method is applicable to any camera-distortion correction situation. Based on a least-squares estimation, our proposed algorithm considers line fits in both FOV directions and gives a globally consistent set of expansion coefficients and an optimal image center. The method is insensitive to the initial orientation of the endoscope and provides more exhaustive FOV correction than previously proposed algorithms. The distortion-correction procedure is demonstrated for endoscopic video images of a calibration test pattern, a rubber bronchial training device, and real human circumstances. The distortion correction is also shown as a necessary component of an image-guided virtual-endoscopy system that matches endoscope images to corresponding rendered three-dimensional computed tomography views.

High-resolution three-dimensional (3-D) volumetric images obtained by today's radiologic imaging scanners are rich in detailed diagnostic information. Despite the many visualization techniques available to assess such images, there remains information that is challenging to uncover, such as the location of small structures (e.g., mediastinal lymph nodes, narrowed-airway regions). Recently, sliding thin-slab (STS) visualization was proposed to improve the visualization of interior structures. These STS techniques sometimes depend on user opacity specifications or extra preprocessing, and other rendering approaches that use the general STS mechanism are conceivable. We introduce two techniques for STS volume visualization. The first, a depth (perspective) rendering process, produces an unobstructed, high-contrast 3-D view of the information within a thin volume of image data. Results are a function of relative planar locations. Thus, rendered views accurately depict the internal properties that were initially captured as position and intensity. The second method produces a gradient-like view of the intensity changes in a thin volume. Results can effectively detect the occurrence and location of dramatic tissue variations, often not visually recognized otherwise. Both STS techniques exploit the concept of temporal coherence to form sequences of consecutive slabs, using information from previously computed slabs. This permits efficient real-time computation on a general-purpose computer. Further, these techniques require no preprocessing, and results are not dependent on user knowledge. Results using 3-D computed tomography chest images show the computational efficiency and visual efficacy of the new STS techniques.

The goal of image segmentation is to partition a digital image into disjoint regions of interest. Of the many proposed image segmentation methods, region growing has been one of the most popular. Research on region growing, however, has focused primarily on the design of feature measures and on growing and merging criteria. Most of these methods have an inherent dependence on the order in which the points and regions are examined. This weakness implies that a desired segmented result is sensitive to the selection of the initial growing points. We define a set of theoretical criteria for a subclass of region-growing algorithms that are insensitive to the selection of the initial growing points. This class of algorithms, referred to as symmetric region growing (SymRG), leads to a single-pass region-growing approach applicable to any dimensionality of images. Furthermore, they lead to region-growing algorithms that are both memory- and computation-efficient. Finally, by-products of this general paradigm are algorithms for fast connected-component labeling and cavity deletion. The paper gives theoretical results and 3-D image examples.

Modern video-based endoscopes offer physicians a wide-angle held of view for minimally-invasive procedures. Unfortunately, inherent barrel distortion prevents accurate perception of range. This makes measurement and distance judgment difficult and causes difficulties in emerging applications, such as 3D medical-image registration. Such distortion also arises in other wide field-of-view camera circumstances. This paper presents a distortion-correction technique that can automatically calculate correction parameters, without precise knowledge of horizontal and vertical-orientation. The method is applicable to any camera-distortion correction situation. Based on a least-squares estimation, our proposed algorithm considers line fits in both field-of-view directions and global consistency that gives the optimal image center and expansion coefficients. The method is insensitive to the initial orientation of the endoscope and provides more exhaustive field-of-view correction than previously proposed algorithms. The distortion-correction procedure is demonstrated for endoscopic video images of a calibration test pattern, a rubber bronchial training device, and real human circumstances. The distortion correction is also shown as a necessary component of an image-guided virtual-endoscopy system that matches endoscope images to corresponding rendered 3D CT views.

High-resolution micro-computed tomography (CT) scanners now exist for imaging small animals. In particular, such a scanner can generate very large three-dimensional (3-D) digital images of the rat's hepatic vasculature. These images provide data on the overall structure and function of such complex vascular trees. Unfortunately, human operators have extreme difficulty in extracting the extensive vasculature contained in the images. Also, no suitable tree representation exists that permits straight-forward structural analysis and information retrieval. This work proposes an automatic procedure for extracting and representing such a vascular tree. The procedure is both computation and memory efficient and runs on current PCs. As the results demonstrate, the procedure faithfully follows human-defined measurements and provides far more information than can be defined interactively.

We introduce an image enhancement method referred to as the watershed-based maximum-homogeneity filter. This method first uses watershed analysis to subdivide the image into homogeneous pixel clusters called catchment basins. Next, using an adaptive, local, catchment-basin selection scheme, similar neighboring catchment basins are combined together to produce an enhanced image. Because the method starts with watershed analysis, it can preserve edge information and run with high computational efficiency. Illustrative results show that the method performs well relative to other popular nonlinear filters.

Three-Dimensional (3D) radiologic images are widely used to assess the condition of thin tubular structures, such as the pulmonary airways, coronary arteries, and colon. Precise 3D central axes of these structures are needed, however, for accurate quantization. Commonly employed manual-axes identification techniques are time-consuming and error-prone. Recently proposed automated techniques do not adequately exploit the available gray-scale or anatomic structural information and they are also prone to errors. The authors propose a method for computing the precise central axes of branching structures contained in 3D images. The method is robust to data anisotropy and uses true gray-scale information. These axes can then be used for automated navigation and assessment in a virtual-endoscopic system. The authors present the application of their method to a human lung-cancer case.

This paper presents a method for the design of multiple Gabor filters for segmenting multi-textured images. Although design methods for a single Gabor filter have been presented previously, the development of general multi-filter multi-texture design methods largely remains an open problem. Previous multi-filter design approaches required one filter per texture or were constrained to pairs of textures. Other approaches employed ad hoc banks of Gabor filters for texture segmentation, where the parameters of the constituent filters were restricted to fixed values and were not necessarily tuned for a specific texture-segmentation problem. The proposed method removes these restrictions on the number of filters and the number of textures. This offers the potential to improve the segmentation performance or to reduce the number of filters. Further, the development of the design method and mathematical models provide new insight into the design of multiple Gabor filters for texture segmentation. Results are presented that confirm the efficacy of our filter-design method and support underlying mathematical models.

Separating halftones from text is an important step in document analysis. We present an algorithm that accurately extracts halftones from other information in printed documents. We treat halftone extraction as a texture-segmentation problem. We show that commonly used halftones, consisting of a pattern of dots, can be viewed as a texture. This texture exhibits a distinct spectral component that can be detected using a properly-tuned Gabor filter. The Gabor (1946) filter essentially transforms halftones into high-contrast regions that can be isolated by thresholding. We propose a filter-design procedure and provide experimental results.

This paper presents an integrated approach using multiple Gabor filters for the segmentation of multi-textured images. The approach includes both the design of the constituent Gabor filters and the design of the classifier and postprocessing. The classifier uses a mixture density to reduce localization error at texture boundaries, and the postprocessing uses morphological operators to remove spurious misclassifications at texture boundaries. Results are presented that confirm the efficacy of the postprocessing methods and the overall integrated approach.

The morphological skeleton and morphological shape decomposition (MSD) are two popular approaches for morphological shape representation. Each method represents an object as an algebraic combination of a number of components, where each component is given by a locus of points dilated by a specified structuring-element homothetic. This correspondence develops a theoretical comparison between the two methods. Combining the theoretical results with several representation cost measures, we make a concrete comparison of the efficiency of the two methods. The results indicate that for complex objects-i.e., objects requiring a full range of homothetic sizes in the morphological skeleton representation-the MSD represents objects more efficiently than the morphological skeleton for three of four suggested cost measures.

Mathematical morphology is well suited to capturing geometric information. Hence, morphology-based approaches have been popular for object shape representation. The two primary morphology-based approaches-the morphological skeleton and the morphological shape decomposition (MSD)-each represent an object as a collection of disjoint sets. A practical shape representation scheme, though, should give a representation that is computationally efficient to use. Unfortunately, little work has been done for the morphological skeleton and the MSD to address efficiency. We propose a flexible search-based shape representation scheme that typically gives more efficient representations than the morphological skeleton and MSD. Our method decomposes an object into a number of simple components based on homothetics of a set of structuring elements. To form the representation, the components are combined using set union and set difference operations. We use three constituent component types and a thorough cost-based search strategy to find efficient representations. We also consider allowing object representation error, which may yield even more efficient representations.

Three-dimensional (3-D) images are now common in radiology. A 3-D image is formed by stacking a contiguous sequence of two-dimensional cross-sectional images, or slices. Typically, the spacing between known slices is greater than the spacing between known points on a slice. Many visualization and image-analysis tasks, however, require the 3-D image to have equal sample spacing in all directions. To meet this requirement, one applies an interpolation technique to the known 3-D image to generate a new uniformly sampled 3-D image. The authors propose a nonlinear-filter-based approach to gray-scale interpolation of 3-D images. The method, referred to as column-fitting interpolation, is reminiscent of the maximum-homogeneity filter used for image enhancement. The authors also draw upon the paradigm of relaxation labeling to devise an improved column-fitting interpolator. Both methods are typically more effective than traditional gray-scale interpolation techniques.

Three-dimensional (3-D) high-resolution coronary angiograms offer a means for visualizing the entire coronary arterial tree from any orientation and for detecting and quantitating coronary arterial stenoses. Previously, a skilled operator had to perform several hours of tedious manual analysis using an interactive graphical user-interface (GUI) system (Tree Trace) to analyze a 3-D angiogram. The authors have devised an improved GUI system, consisting of three tools for analyzing 3-D angiograms. The Artery Extractor first performs automatic image-analysis operations to extract the central axes of the arterial tree. Next, using the Artery Display tool and results from the Artery Extractor, the operator can visualize structures in the angiogram and compute various measurements. Finally, the aforementioned Tree Trace tool can be used to manually correct irregularities in the automatically generated results of the Artery Extractor. The system greatly reduces operator analysis time, gives exactly reproducible results, uses true 3-D image-processing operations, and provides a comprehensive interface for visualizing and quantifying features of the 3-D coronary arteries.

We introduce an image enhancement method referred to as watershed-based maximum homogeneity filtering. This enhancement technique uses watershed analysis to initially cluster groups of similar connected pixels into catchment basins. Similar neighboring catchment basins are then averaged together to compute the filtered image. As illustrative results show, this technique compares well to other popular nonlinear filters.

Traditional image segmentation methods typically divide an image into separate regions based on the grayscale characteristics of the image. For most real-world image-segmentation problems, however, these methods tend to produce imperfectly shaped regions that require some degree of shape modification to yield acceptable results. Choosing an appropriate sequence of operators and associated operator parameters, though, is a tedious procedure and requires much image-processing expertise. We describe a strategy for easily selecting shape-based operations. Shape information on regions in an image is provided by the user in the form of easily-specified cues. The user is not required to be an image-processing expert to apply the strategy-he need only be able to specify the desired shape properties of the regions in the image.

Texture segmentation involves subdividing an image into differently textured regions. Many texture segmentation schemes are based on a filter-bank model, where the filters, called Gabor filters, are derived from Gabor elementary functions. The goal is to transform texture differences into detectable filter-output discontinuities at texture boundaries. By locating these discontinuities, one can segment the image into differently textured regions. Distinct discontinuities occur, however, only if the Gabor filter parameters are suitably chosen. Some previous analysis has shown how to design filters for discriminating simple textures. Designing filters for more general natural textures, though, has largely been done ad hoc. We have devised a more rigorously based method for designing Gabor filters. It assumes that an image contains two different textures and that prototype samples of the textures are given a priori. We argue that Gabor filter outputs can be modeled as Rician random variables (often approximated well as Gaussian rv's) and develop a decision-theoretic algorithm for selecting optimal filter parameters. To improve segmentations for difficult texture pairs, we also propose a multiple-filter segmentation scheme, motivated by the Rician model. Experimental results indicate that our method is superior to previous methods in providing useful Gabor filters for a wide range of texture pairs.

Introduces an image segmentation method referred to as watershed-driven relaxation labeling. The method is a hybrid segmentation process utilizing both watershed analysis and relaxation labeling. Initially, watershed analysis is used to subdivide an image into catchment basins, effectively clustering pixels together based on their spatial proximity and intensity homogeneity. Classification estimates in the form of probabilities are set for each of these catchment basins. Relaxation labeling is then used to iteratively refine and update the classifications of the catchment basins through propagating constraints and utilizing local information. The relaxation updating process is continued until a large majority of the catchment basins are unambiguously classified. The method provides fast, accurate segmentation results and exploits the individual strengths of watershed analysis and relaxation labeling. The robustness of the method is illustrated through comparisons to other popular segmentation techniques.>

Solutions to real-world image-segmentation problems typically require many image-processing steps. Unfortunately, the user must decide how to construct this sequence of steps for a given problem. So far, no system has been proposed to make the construction of these processes "easy" for the user. As a result, segmentation processes are often laboriously developed by an image-processing expert. We describe a method for automatically generating image-segmentation processes for arbitrary images. Our method uses cue-based image analysis. The user provides problem-specific information via easily defined cues. Two types of cues can be defined: (1) iconic cues, which are image-based and constructed by drawing directly onto the image data; and (2) symbolic cues, which are verbally specified facts. The cues are interpreted to help select image-processing functions. The user need not be an image-processing expert-he must only understand the significance of the specified cues for a particular problem.>

Gabor filters have been applied successfully to the segmentation of textured images. Previous investigators have used banks of Gabor filters, where the filter parameters were predetermined ad hoc and not necessarily optimized for a particular task. Other investigators have proposed using filters tuned to dominant components in the FFT of constituent textures. More recent work presented a Gabor filter design method using a Rician distribution to characterize the filtered textures. The article addresses the design of a single Gabor filter to segment multiple textures and is based on using the Rician distribution at two different scales of the Gabor-filter envelope. Furthermore, variable degrees of postfiltering and the accompanying effect on postfilter output statistics are considered.>

Gabor filters have been applied successfully to a broad range of multidimensional signal processing and image processing tasks. The present paper considers the design of a single filter to segment a two-texture image. A new efficient algorithm for Gabor-filter design is presented. The algorithm draws upon previous results that showed that the output of a Gabor-filtered texture is well represented by a Rician distribution. The new algorithm uses the Rician model to estimate the output statistics of a pair of sample textures from their windowed autocorrelation functions. A measure of the total output power is used to select the center frequency of the filter and estimate the output statistics. The method is further generalized to include the statistics of post-filtered outputs. Experimental results are presented that demonstrate the efficacy of the algorithms.>

Many texture-segmentation schemes use an elaborate bank of filters to decompose a textured image into a joint space/spatial-frequency representation. Although these schemes show promise, and although some analytical work has been done, the relationship between texture differences and the filter configurations required to distinguish them remain largely unknown. This paper examines the issue of designing individual filters. Using a 2-D texture model, we show analytically that applying a properly configured bandpass filter to a textured image produces distinct output discontinuities at texture boundaries; the analysis is based on Gabor elementary functions, but it is the bandpass nature of the filter that is essential. Depending on the type of texture difference, these discontinuities form one of four characteristic signatures: a step, ridge, valley, or a step change in average local output variation. Accompanying experimental evidence indicates that these signatures are useful for segmenting an image. The analysis indicates those texture characteristics that are responsible for each signature type. Detailed criteria are provided for designing filters that can produce quality output signatures. We also illustrate occasions when asymmetric filters are beneficial, an issue not previously addressed.>

Mathematical morphology is well-suited to capturing geometric information. Hence, morphology-based approaches have been popular for object shape representation. The two primary morphology-based approaches, the morphological skeleton and the morphological shape decomposition (MSD), each represent an object as a collection of disjoint sets. A practical shape representation scheme, though, should give a representation that is computationally efficient to use. Unfortunately, little work has been done for the morphological skeleton and the MSD to address efficiency. We propose a flexible search- based shape decomposition scheme that typically gives more efficient representations than the morphological skeleton or MSD. Our method decomposes an object into a number of simple components based on homothetics of a set of structuring elements. To form the representation, the components are combined using set union and set difference operations. We use three constituent component types and a thorough cost-based search strategy to find efficient representations. We also consider allowing some object representation error, which may yield even more efficient representations.

Many three-dimensional (3-D) medical images have lower resolution in the z direction than in the x or y directions. Before extracting and displaying objects in such images, an interpolated 3-D gray-scale image is usually generated via a technique such as linear interpolation to fill in the missing slices. Unfortunately, when objects are extracted and displayed from the interpolated image, they often exhibit a blocky and generally unsatisfactory appearance, a problem that is particularly acute for thin treelike structures such as the coronary arteries. Two methods for shape-based interpolation that offer an improvement to linear interpolation are presented. In shape-based interpolation, the object of interest is first segmented (extracted) from the initial 3-D image to produce a low-z-resolution binary-valued image, and the segmented image is interpolated to produce a high-resolution binary-valued 3-D image. The first method incorporates geometrical constraints and takes as input a segmented version of the original 3-D image. The second method builds on the first in that it also uses the original gray-scale image as a second input. Tests with 3-D images of the coronary arterial tree demonstrate the efficacy of the methods.

Describes a graphical user interface system for constructing automatic image-segmentation processes for 3D medical-imaging applications. The system provides a user-friendly environment for prototyping, testing, and executing complex image-analysis processes and can interact with a graphics-based 3D medical image-analysis package. The interface links to a library of 1D, 2D, and 3D image-processing functions and can easily link to other functions created by the user. Demonstrations of the system are given for 3D left-ventricular chamber extraction and 3D upper-airway analysis.>

Many texture-segmentation schemes use Gabor elementary functions as filters to decompose a textured image into a joint space/spatial-frequency domain. Dunn et al. (1991, 1992) showed for a 1-D model that these filters (when properly tuned) can produce certain characteristic discontinuities (step, valley and ridge) at texture boundaries, aiding image segmentation. The authors generalize the work of Dunn et al. by providing a more general 2-D analytical development. They also show when vertical and horizontal texture-phase discontinuities produce ridge or valley discontinuities, and when asymmetric filters are beneficial.>

A description is presented of three related methods for detecting 2-D region discontinuities. One method extends Leclerc's (1985) 2-D work. The other two arise from the concept behind the maximum-homogeneity filter. All of the methods use some form of 2-D neighborhood-partitioning mechanism, like that of Leclerc, and hence use 2-D local structural information. Also, all the methods allow detection of multiple edges about an image point. The methods are described and illustrative results presented.>

The problem of interpolating a high-resolution binary-valued 3-D image given a lower-resolution grayscale 3-D image and a corresponding segmented 3-D image is addressed. Three algorithms for this problem are described. Algorithms 1 and 2 perform shape- and located-based binary interpolation. Algorithm 3, which subsumes algorithms 1 and 2, incorporates grayscale information as well as object shape and slice-to-slice object spatial relationships. Results show that algorithm 3 generates better interpolated objects than the segmented volumes arising from standard grayscale-based interpolation techniques. Algorithm 2, which subsumes algorithm 1, uses only the known presegmented volume to interpolate a new binary-valued volume. the methods show particular merit in cases where the objects to interpolate are thin tube-like structures, such as the coronary arteries.>