Abstract: A computer assisted diagnosis method using change detection of salient features in a first and a second digital image includes the step of converting the first and the second digital image into a first and a second relational attribute graph, respectively. Each graph comprises nodes and arcs. Each node corresponds to an identified salient feature and associated with information comprising a type and characteristics of the corresponding identified salient feature. The arcs correspond to a topological arrangement of the identified salient features. An optimal inexact structural match is determined between the nodes of the first graph and the nodes of the second graph so as to form matched sets of nodes comprising one node from each graph. A node lacking a determined match is matched with a null node. The characteristics of each of the nodes in a matched set are compared to one another to identify variances among the compared characteristics.

Abstract: A method for automatic CR image composition includes the following steps for each image in a CR image pair. A vertical gradient image is projected onto a y-axis to obtain a y-axis projection (420). Edge positions (candidates) are identified from white band edge positions based on an absolute maximum value of the y-axis projection (430). An intensity change constancy identifies candidates having orientation angles less than a threshold angle with respect to a horizontal (440). An intensity change value verifies intensity value differences on two candidate sides (450). A cross-correlation score is computed for the candidates having an error value below a threshold for the intensity change constancy and the intensity change value, by comparing a consistency of the candidates against image data of the CR image pair (460, 610). A final overlap is identified for the CR image pair, based upon the cross-correlation score (620).

Abstract: A computer-assisted diagnosis method for assisting diagnosis of anatomical structures in a digital volumetric medical image of at least one lung includes identifying an anatomical structure of interest in the volumetric digital medical image. The anatomical structure of interest is automatically segmented, in real-time, in a predefined volume of interest (VOI). Quantitative measurements of the anatomical structure of interest are automatically computed, real-time. A result of the segmenting step and a result of the computing step are displayed, in real-time. A likelihood that the anatomical structure of interest corresponds to a disease or an area warranting further investigation is estimating, in real-time, based on predefined criteria and the quantitative measurements. A warning is generated, in realtime, when the likelihood is above a predefined threshold.

Abstract: A computer-assisted diagnosis method and system are provided for automatically determining diagnostic saliency of digital images. The method includes the step of providing filters for evaluating the image. Each filter is designed to identify a specific type of diagnostic finding, and is associated with the following: a virtual window for defining regions in the image at which the filter is applied; a set of training image patches corresponding to typical appearances of the specific type of diagnostic finding; a distance measure between the training image patches and the regions in the image defined by the virtual window; and a feature set corresponding to the distance measure. The filters are applied to the image to compute distances between the regions in the image defined by the virtual window and the training image patches based on the distance measure and the feature set, for each of the plurality of filters.

Abstract: Controlling the stepping of an x-ray angiography imaging device (and/or table), and/or providing a visualization aid for stepping an x-ray angiography imaging device (and/or table), to maximize the diagnostic usefulness of images acquired by the x-ray angiography imaging device. The stepping and visualization aid use features, extracted from captured images, of the state of contrast media injected into a patient.

Abstract: A system for automatically composing target objects from a sequence of x-ray images for the purpose of diagnostic examination or measurement detects and matches semantically meaningful visual event hierarchy. The system can automatically composite the target object explicitly ignoring image regions that are not of interest for diagnosis. Diagnostic examination and measurement are helped significantly by the use of compound x-ray images. Prior to compound image generation, each constituent image undergoes two automatic pre-processing steps through a background detector and an emphasis field extractor. The system further includes a pair-wise aligner, a compound image generator and an examiner/measurer. The system increases the efficiency of compound image generation and improves its accuracy by de-emphasizing alignment of irrelevant structures in the images.

Abstract: An automatic selection system utilizes a computer which automatically localizes the heart and automatically selects the most suitable representative slice among a set of transverse slices, which are reconstructed from the myocardial SPECT projection data, for determination of the left ventricular oblique axis in 3-D space. An intensity profile, which consists of maximum counts of each slice and correlates to the radioactivity of the organs in the reconstructed volume, is analyzed. The heart's left ventricle is localized and the representative slice is selected based upon the local and global extrema of such profile.

Abstract: A method for automatically setting a collimator of an x-ray imaging system during image acquisition includes receiving rapid scout images at an imaging station. The location of the body regions in one of said images is then automatically detected. The detected location of the body regions is used to generate settings for the collimator. The settings are used for automatically adjusting the collimator to substantially cover the non-body regions and substantially expose the body regions.

Abstract: Controlling the stepping of an x-ray angiography imaging device (and/or table), and/or providing a visualization aid for stepping an x-ray angiography imaging device (and/or table), to maximize the diagnostic usefulness of images acquired by the x-ray angiography imaging device. The stepping and visualization aid use features, extracted from captured images, of the state of contrast media injected into a patient.

Abstract: A contexture dependent dynamic range remapping system utilizes background estimation, mask generation, parameter estimation and dynamic range remapping of an original image to provide an optimum output image. The dynamic range remapping includes an additive algorithm or a multiplicative algorithm or a combination of both algorithms. The output from the algorithms could be filtered by an edge-preserving filter to provide a filtered output image. The system adaptively compresses the dynamic range of the DC and slow-varying signals in the original image significantly while preserving and enhancing fine structures.

Abstract: A computer system automatically takes a sequence of stepping images from peripheral angiography and mosaics them into a single full-leg display by globally matching the bone and the vessel by measuring an overlapping ratio and by locally refining detailed anatomical features using deformation. The system combines multiple evidence in the entire overlapped image/candidate rows and maximizes the ratio of overlap. This approach takes evidence from the multiple rows and provides a more reliable result. In addition, local refinement is applied to compensate for the possible mismatch and nonlinear patient movement.

Abstract: An adaptive edge-preserving smoothing filter effectively reduces noise levels while preserving fine structures in data. The behavior of the filter is controlled easily by two control parameters. To adaptively control the behavior of the filter, the control parameter, .alpha., can be set as a function of the local directional variances. The control parameter, .beta., can be set as a function of all of the directional variances and directional means. The filter includes; a compute means which receives the filter window size, the number of filter directions and input data, adaptive weighting parameter map means which receives the control parameter, .alpha., adaptive weighting process means, adaptive combination parameter map means which receives the control parameter, .beta., and final compute means which provides the filtered output data.

Abstract: A SPECT study is carried out on a patient's body organ, such as the heart, and frames of image data are thereby acquired. The image data in these frames are subjected to a series of mappings and computations, from which frames containing a significant quantity of organ motion can be identified. Quantification of the motion occurs by shifting some of the mapped data within a predetermined range, and selecting that data shift which minimizes the magniture of a motion-sensitive mathematical function.

Abstract: A method is disclosed for automatically finding the true axis and the radius of cross sections of three dimensional tube-like shaped objects such as the left head ventricle, lung airways and blood vessels from 3-D transverse images constructed from known imaging techniques. The problem of finding the center axis of a 3-D tube-like shape is reduced to searches for the center points of its cross sections. The true axis of the tube-like shaped object is defined as the 3-D curve which connects all detected center points of all cross sections. Additionally, a method is disclosed for automatic determination of the center and the radius of each cross section of the tube-like shape, which may be irregular or incomplete, by analyzing the average intensity of the image intensity patterns in transverse image slices of the cross section.

Abstract: This invention provides a method for determining the position range of the heart from a sequence of projection images which are acquired for 3-D volume reconstruction, such as SPECT myocardial projection data. The essence of this invention is the use of 1-D pseudo motion analysis so that the detection is insensitive to the image intensity distribution. Heart position is determined by comparing the sampled heart motion against a standard heart motion and determining similarities between the two as an indication of the position range of the sampled heart including the proximate center, upper and lower limits of motion.

Abstract: A multi-head SPECT-capable scintillation camera has one head collimated by a parallel hole collimator and one or more focussing collimators. The focussing collimators are advantageously cone beam collimators or (for scintillation cameras which have rectangular detectors) astigmatic collimators. The scintillation camera is used to acquire information about a relatively small body organ of interest, such as the heart or the brain. From this information, the spatial location of the organ is determined. Then, during rotation of the camera heads about the patient during a SPECT study, the positions of the focussing collimators are controlled so as to maintain the organ of interest in a predetermined position with respect to their field(s) of view. Thereafter, the data acquired during the SPECT study is used to form a SPECT image of the organ of interest. The SPECT image may then be displayed or outputted.

Abstract: Two planar nuclear medicine images of a target organ are acquired using a focussing collimator at two different heights. An anatomic landmark associated with the target organ is computer-identified in each of the images, and the depth of the target organ is determined geometrically using the differences in size between the images of the identified landmark and the differences in height.

Abstract: Anatomical landmarks relating to a target organ which is to be studied in a nuclear medicine study are automatically identified by a computer. The landmarks are superimposed upon a nuclear medicine persistence image of the target organ. This facilitates technician identification of the target organ and also facilitates repeatability of, e.g., myocardial perfusion studies, which require that two studies be performed on a single patient at two different times.

Abstract: The long axis of the left ventricle is automatically identified by identifying, and correlating, local minima and maxima in images of slices of the left ventricle. Initially, the left ventricle is identified within a representative transverse slice of the left ventricle. The centerline of this slice is automatically computed and used as a reorientation axis, along which another (sagittal) slice of the left ventricle is reconstructed. The centerline of this sagittal slice is automatically computed, and is the long axis of the left ventricle. The accuracy of this identification is confirmed by reconstructing oblique transverse slices of the left ventricle and verifying that their centers are coincident.

Abstract: A nuclear medicine image is scanned and pixels of maximum and minimum intensity are identified and correlated with each other using constraints which are empirically determined to relate to the feature of interest (such as the heart). The information thus obtained is used to define a region of interest in which an anatomical feature of interest may be located, and to position a scintillation camera detector to carry out a nuclear medicine study at optimal positions.