Abstract: When calibrating a nuclear camera, such as a SPECT camera, point spread functions (PSF) are simulated for all possible photon energies and detection distances that the camera may experience. During manufacturer-side calibration, a point source (14) is imaged by the nuclear camera and a PSF therefor is measured. The measured PSF is compared to one or more simulated PSFs and a difference therebetween is determined. To calibrate the measured PSF, simulated PSF data is interpolated into the measured PSF to minimize the difference between the measured and simulated PSFs. The calibrated PSF is stored to memory and accessed during user-side calibration. A user then images a tracer sample during a routine camera calibration, and a PSF for the tracer sample is compared to the calibrated PSF to determine a level of contamination in the tracer sample. The tracer sample PSF is then calibrated for use in reconstructing nuclear images.

Abstract: An imaging system (10) includes at least one radiation detector unit (16) disposed adjacent a field of view (20) to detect and measure radiation from the field of view (20). The detector unit (16) includes multiple detection modules (18) which each detects radiation from a prespecified region of the field of view (20), each region being a fraction of the field of view. One or more pinholes (52) are associated with the detector unit (16). Each pinhole (52) receives radiation from the prespecified region of the field of view (20) and transmits radiation to one or more associated detection modules (18).

Abstract: When calibrating a nuclear camera, such as a SPECT camera, point spread functions (PSF) are simulated for all possible photon energies and detection distances that the camera may experience. During manufacturer-side calibration, a point source (14) is imaged by the nuclear camera and a PSF therefor is measured. The measured PSF is compared to one or more simulated PSFs and a difference therebetween is determined. To calibrate the measured PSF, simulated PSF data is interpolated into the measured PSF to minimize the difference between the measured and simulated PSFs. The calibrated PSF is stored to memory and accessed during user-side calibration. A user then images a tracer sample during a routine camera calibration, and a PSF for the tracer sample is compared to the calibrated PSF to determine a level of contamination in the tracer sample. The tracer sample PSF is then calibrated for use in reconstructing nuclear images.

Abstract: When performing a static image reconstruction of acquired single photon emission computed tomography (SPECT) data for myocardium, dynamic tracer uptake, redistribution, and washout information is generated with reduced or eliminated artifacts by back-projecting ray projections onto a reconstructed myocardial surface. A complete SPECT scan is performed after tracer injection, and a static image of the myocardial surface is reconstructed. The reconstructed image is segmented and a polar plot of it is generated. A contemporaneously acquired subset of the SPECT projection data is then back-projected onto the segmented surface of the polar plot. Contributions from emissions not originating from the myocardium (e.g., from adjacent anatomical structures) are compensated. The resultant image data, which describes tracer distributions across heart segments per projection time, are overlaid on the polar plot and presented to a user.

Abstract: In order to reduce an x-ray dose applied to a patient, it is necessary to know the dose absorbed by the patient. According to the present invention, there is provided a method of determining a local patient dose applied to a patient where after the reconstruction of the scan data into a diagnostic image, the scan data are backprojected into the patient volume, using the attenuation information of the diagnostic image to form a spatially varying photon fluence map. In parallel, the diagnostic image is segmented into anatomical structures to which dose-weighting factors are assigned. The locally absorbed dose is then calculated on the basis of the fluence map and the corresponding dose weights.

Abstract: An imaging system (10) comprises at least one radiation detector unit (16) disposed adjacent a field of view (20) to detect and measure radiation from the field of view (20). The detector unit (16) includes multiple detection modules (18) which each detects radiation from a prespecified region of the field of view (20), each region being a fraction of the field of view. One or more pinholes (52) are associated with the detector unit (16). Each pinhole (52) receives radiation from the prespecified region of the field of view (20) and transmits radiation to one or more associated detection modules (18).

Abstract: The invention relates to a method of adapting imaging parameters for a computer tomographic radiograph of a body volume, comprising the following steps: obtaining a three-dimensional pilot radiograph with a low dose of radiation (1); determining a region of interest and a desired image quality in the pilot radiograph (2) with the aid of a patient model (4) or interactively (3); determining optimal imaging parameters (5); generating an X-ray image using the determined imaging parameters (6). Optionally, the X-ray image is combined (7) with the pilot radiograph.

Abstract: In order to reduce an x-ray dose applied to a patient, it is necessary to know the dose absorbed by the patient. According to the present invention, there is provided a method of determining a local patient dose applied to a patient where after the reconstruction of the scan data into a diagnostic image, the scan data are backprojected into the patient volume, using the attenuation information of the diagnostic image to form a spatially varying photon fluence map. In parallel, the diagnostic image is segmented into anatomical structures to which dose-weighting factors are assigned. The locally absorbed dose is then calculated on the basis of the fluence map and the corresponding dose weights.

Abstract: A speech recognition method is described in which a basic set of models is adapted to a current speaker on account of the speaker's already noticed speech data. The basic set of models comprises models for different acoustic units. The models are described each by a plurality of model parameters. The basic set of models is then represented by a supervector in a high-dimensional vector space (model space), the supervector being formed by a concatenation of the plurality of the model parameters of the models of the basic set of models. The adaptation of this basic set of models to the speaker is effected in the model space by means of a MAP method in which an asymmetric distribution in the model space is selected as an a priori distribution for the MAP method.

Abstract: Described here is a method of determining an eigenspace for representing a plurality of training speakers, in which first speaker-dependent sets of models are formed for the individual training speakers while training speech data of the individual training speakers are used and the models (SD) of a set of models are described each by a plurality of model parameters. For each speaker a combined model is then displayed in a high-dimensional model space by concatenation of the model parameters of the models of the individual training speakers to a respective coherent super vector. Subsequently, a transformation is carried out while the model space dimension is reduced to obtain eigenspace basis vectors (Ee), which transformation utilizes a reduction criterion based on the variability of the vectors to be transformed. Then the high-dimensional model space is first in a first step reduced to a speaker subspace by a change of basis, in which speaker subspace all the training speakers are represented.

Abstract: Described here is a method of determining an eigenspace for representing a plurality of training speakers, in which first speaker-dependent sets of models are formed for the individual training speakers while training speech data of the individual training speakers are used and the models (SD) of a set of models are described each by a plurality of model parameters. For each speaker a combined model is then displayed in a high-dimensional model space by concatenation of the model parameters of the models of the individual training speakers to a respective coherent supervector. Subsequently, a transformation is carried out, while the dimension of the model space is reduced for recovering eigenspace basis vectors (Ee). To guarantee an unambiguous assignment of the model parameters in the supervectors, first a common speaker-independent set of models is developed for the training speakers and this set of models is adapted to the individual training speakers to develop the speaker-dependent sets of models.

Abstract: A speech recognition method is described in which a basic set of models is adapted to a current speaker on account of the speaker's already noticed speech data. The basic set of models comprises models for different acoustic units. The models are described each by a plurality of model parameters. The basic set of models is then represented by a supervector in a high-dimensional vector space (model space), the supervector being formed by a concatenation of the plurality of the model parameters of the models of the basic set of models. The adaptation of this basic set of models to the speaker is effected in the model space by means of a MAP method in which an asymmetric distribution in the model space is selected as an a priori distribution for the MAP method.

Abstract: Described here is a method of determining an eigenspace for representing a plurality of training speakers, in which first speaker-dependent sets of models are formed for the individual training speakers while training speech data of the individual training speakers are used and the models (SD) of a set of models are described each by a plurality of model parameters. For each speaker a combined model is then displayed in a high-dimensional model space by concatenation of the model parameters of the models of the individual training speakers to a respective coherent super vector. Subsequently, a transformation is carried out while the model space dimension is reduced to obtain eigenspace basis vectors (Ee), which transformation utilizes a reduction criterion based on the variability of the vectors to be transformed. Then the high-dimensional model space is first in a first step reduced to a speaker subspace by a change of basis, in which speaker subspace all the training speakers are represented.