Abstract: A system for browsing and editing video, in accordance with the present invention, includes a video source for providing a video document which includes audio information, and an audio classifier coupled to the video source, the audio classifier being adapted to classify audio segments of the audio information into a plurality of classes. An audio spectrogram generator is coupled to the video source for generating spectrograms for the audio information to check that the audio segments have been identified correctly by the audio classifier. A browser is coupled to the audio classifier for searching the classified audio segments for editing and browsing the video document. A method for editing and browsing a video, in accordance with the invention, includes providing a video clip including audio, and segmenting and labeling the audio into music, silence and speech classes in real-time.

Abstract: A mark-free computer-assisted diagnosis method and system are provided for assisting diagnosis of abnormalities in digital medical images using diagnosis based image enhancement. The method includes the steps of: receiving indicia identifying one or more regions of interest in a digital medical image; and displaying one or more enhanced views of the regions of interest, the enhanced views being based on diagnostic parameters for the regions of interest and diagnostic parameters corresponding to a particular abnormality.

Abstract: A system and method for automatically detecting anatomical landmarks in a radiographic image, preferably for total hip replacement applications, in accordance with the present invention, provides a region of interest of the image, and determines a first landmark in the region of interest by computing an intensity ridge map. A second landmark is determined in the region of interest based on a position and orientation of the first landmark, and measurements are performed on the image based on positions of the first and second landmarks.

Abstract: A method for managing document information on an information net, such as the World Wide Web (WWW), comprises the steps of inputting a structured document; extracting selected document properties from the structured document; forming a feature vector representative of the properties; and forming a caricature derived from the feature vector.

Abstract: A diagnostic decision support system provides diagnostic decision support for auditory evaluation of anatomical features and is applicable to virtually any living creature. The system processes an acoustic signal for medical applications by acquiring acoustic data representative of an acoustic signal associated with an anatomical function. The acquired acoustic data is stored in a file associated with a patient medical record. The acquired acoustic data and medical record information is automatically analyzed to determine physiologically significant features useful in medical diagnosis. Information is generated supporting medical diagnosis based on the automatic analysis. In addition, the analysis of the acquired acoustic data may be partially automatic and involve User input of information for use in facilitating diagnostic decision making. The system also processes patient identification information and acoustic test type information (e.g.

Abstract: Systems and methods for correcting the image intensity of under-exposed regions of CR (computed radiography) images for mosaic composition. In one aspect, a method for generating a mosaic CR image comprises acquiring a set of CR images and processing each CR image to detect an under-exposed region in the CR image. If an under-exposed region is detected, the CR image is separated into an under-exposed region and normal exposure region. The image intensity of the under-exposed region is then adjusted to be substantially similar to, or as close as possible to, the image intensity of the normal exposure region of the CR image using nonlinear enhancement and/or linear transformation techniques. The intensity corrected CR images are then combined in a mosaic image.

Abstract: A method for video assistance for ultrasound guided needle biopsy of a patient, includes the steps of obtaining an ultrasound image of said patient in an ultrasound imaging plane by way of a transducer head; obtaining a body surface video image of said patient including an area wherein said ultrasound imaging plane intersects said body surface, by way of a video camera mounted on said transducer head and having an optical axis lying in said ultrasound imaging plane such that said ultrasound imaging plane corresponds to a line in said video image; and overlaying graphic markers onto said video image for indicating said ultrasound imaging plane as a line of possible entry points for needle biopsy.

Abstract: A method for automatic detection of lung nodules in high-resolution multi-slice CT images by using cartwheel projection analysis is disclosed. The projection analysis employs weighted area curves obtained from a series of 2-dimensional cutting planes rotated around a defined axis in the volume of interest. Nodules are detected if the shape analysis of the weighted area curves meets a set of test-criteria.

Abstract: A fully automatic system and method for detecting endplates of vertebra in radiographic images, in accordance with the present invention, provides an intensity curvature map of an image of a spine and filters the intensity curvature map in a direction relative to a spine axis. At a plurality of points of the spine axis, maximum and minimum scores are computed by summing curvatures of the intensity curvature map and by summing directional intensity gradients along line segments to determine if an endplate passes through the points. Endplate positions are found by comparing relative positions of plotted values corresponding to the maximum and minimum scores.

Abstract: A system and method for automatically detecting a spinal axis and spinal boundaries from an image, in accordance with the present invention, includes scanning an input image in at least two scan directions to generate a ridge map associated with each scan direction. The ridge maps are fused to enhance a projected spinal axis. A piece-wise linear approximation of the projected spine axis is provided, and a gradient map of the input image is enhanced project spine boundaries. Spine boundaries are determined by constraining the projected spine boundaries relative to each other and the spine axis.

Abstract: The present invention relates to a method and a system for extracting a 3D region of interest from a stack of medical scan slices which is still a challenging task even in this new millennium. One difficulty is that there is no robust automatic algorithm that can handle all the different situations and applications. Human intervention is usually unavoidable for extracting desired area or organs. However, without convenient tools, it is a tedious job for human being to go through hundred or thousand of slices just to extract region of interest. A system combines and modifies several advanced techniques to help a user extract volume of interest very easily and quickly. The technique of livewire for initial-delineate and modify regions of interest from 2D slices is utilized. Then variational interpolation technique is employed to derive a 3D shape from these delineated 2D contours.

Abstract: A method and system for visual servoing of a linear apparatus rotatable in a plane about a fixed point implements cross-ratios and provides alignment of the linear apparatus to the target in an image, such as a fluoroscope image in 3 iterations. The imaging device through which the scene is observed can be of any suitable type and does not need to be calibrated.

Abstract: A method and apparatus for real-time tracking of a non-rigid target. The tracking is based on visual features, such as color and/or texture, where statistical distributions of those features characterize the target. A degree of similarity (&rgr;(y0)) is computed between a given target (at y0) in a first frame and a candidate target (at y1) in a successive frame, the degree being expressed by a metric derived from the Bhattacharyya coefficient. A gradient vector corresponding to a maximization of the Bhattacharyya coefficient is used to derive the most probable location of the candidate target in the successive frame.

Abstract: A method for extracting an image representing textual information from a video sequence includes the following steps. First, receiving a sequence of video frames, each including an image of textual information. Then, locating the textual information in each frame of the video sequence to form a stack of text arrays, each array containing data representing substantially only the textual information. Finally, extracting a single textual image array representing the image of the textual information from the stack of text arrays. Apparatus for extracting an image representing textual information from a video sequence includes a source of a video sequence having a plurality of frames, each containing an image of the textual information; and a processor, coupled to the video sequence source, responsive to all of the plurality of frames, for generating a single array representing an image of the textual information.

Abstract: A method and system for computer assisted site navigation includes capturing an image of a scene for providing real-time image data, calibrating a camera and localizing the image by utilizing the image data and retrieved marker registration and floor-map data for providing position and orientation data, transmitting the data to a web server, utilizing the image data and information on the scene retrieved from a database for deriving an augmented view, and displaying the augmented view.

Abstract: A system for organizing digital videos to archive and access them at different levels of abstraction uses data available from a closed-caption text along with off-the-shelf natural language processing tools to segment the video into self-contained story sections and speaker blocks. If the subject changes are marked, the system uses these points to divide the video into distinct stories which are represented as nodes attached to the root node in a tree structure and groups speaker segments belonging to a story under the story node as its children.

Abstract: A signal processing technique can be effectively used for source separation, signal enhancement, and noise reduction when using a twin microphone system. The class of stochastic signals for which ratio-estimates can be computed from histograms is defined. This class fits real-world signals of interest such as voice signals. Theoretical computation in closed form of the optimal estimator for this class of signals is disclosed. Two practical implementation solutions are disclosed, as is a practical solution to exploit an echoic environment model. Furthermore, two novel techniques for signal demixing are presented. The application of the optimal estimator and the suboptimal estimator to the case of more than two channels is disclosed.

Abstract: A system of hyperlink authoring allows authors to establish links in multimedia documents interactively and provides generalization and specialization mechanisms for transforming the content and context of the instances of links indicated by the authors into more general link rules that when executed can automatically generate a large number of syntactically similar links. A major component of the system, the link editor, comprises specifications for the source, the intermediate destination, if required, and the final destination. Each specification comprises content pattern generalization, pattern variable instantiation, attribute modification, context specialization and attribute variable instantiation. The output of the link editor is link specifications which are forwarded to an automatic hyperlinker which outputs linked documents.

Abstract: There is provided a method for accelerating the generation and display of volume-rendered cut-away-views of three-dimensional images. The method includes the step of rendering a three-dimensional image from a set of textured image surfaces. At least one image corresponding to at least one intermediate result of the rendering step is accumulated. The at least one image includes at least one accumulation of at least two of the textured image surfaces. The at least one image is stored in a volume buffer for subsequent rendering of cut-away-views of the three-dimensional image therefrom. The accumulating step includes the step of accumulating one image for each of the plurality of textured image surfaces. Alternatively, the accumulating step includes the step of accumulating one image for only some of the plurality of textured image surfaces.

Abstract: A method for extracting features from cardiac acoustic signals includes the steps of obtaining a cardiac acoustic signal, and extracting physiologically significant features from the cardiac acoustic signal using a neural network. A method for evaluating cardiac acoustic signals includes the steps of obtaining a cardiac acoustic signal, analyzing the cardiac acoustic signal with a wavelet decomposition to extract time-frequency information, and identifying basic heart sounds using neural networks applied to the extracted time-frequency information. A method for determining cardiac event sequences from cardiac acoustic signals includes the steps of obtaining a cardiac acoustic signal, and processing a sequence of features extracted from the cardiac acoustic signal by a probabilistic finite-state automaton to determine a most probable sequence of cardiac events given the cardiac acoustic signal.