Abstract: A method of image reconstruction from wedge beam data includes collecting cone beam data in two-dimensional arrays. The collected data corresponds to rays of penetrating radiation diverging in two dimensions from a common vertex which travels along a curve. Each element of the data is related to line integrals of an object being reconstructed taken along each ray. A local coordinate system is defined having three mutually orthogonal axes and an origin at the vertex. The third axis of the local coordinate system extends in a direction tangential to the curve at the vertex. The collected data is rebinned into a wedge beam format wherein sets of parallel rays are grouped to define planes of radiation that angularly diverge from a common axis. A one-dimensional convolution of the rebinned data is computed in the local coordinate system along a direction parallel to the third axis. Finally, the convolved data is weighted and three-dimensionally backprojected.

Abstract: A method of image reconstruction from partial cone beam data includes collecting the partial cone beam data in two-dimensional arrays. The collected data corresponds to rays of radiation diverging in two dimensions from a common vertex as the vertex travels along a curve. Each element of the data represents a line integral of an object being reconstructed taken along each ray. The data is then pre-weighted and a local coordinate system is defined. A one-dimensional convolution of the pre-weighted data is computed in the local coordinate system along a direction which is tangential to the curve along which the vertex travels. Finally, the convolved data is weighted and three-dimensionally backprojected. In a preferred embodiment, the curve defines a helical path, and the kernel for the one-dimensional convolution of the pre-weighted data is a ramp convolver.

Abstract: An x-ray tube of CT or other radiographic scanner has a focal spot (30) from which x-rays with an energy distribution (32) are generated. As the x-ray source rotates a peak (34), the radiation energy distribution seen by each detector (24) through a physical collimator shifts. Each line of data generated by the detectors has a shifting energy distribution across its data values which creates artifacts in reconstructed images. A filter (50) includes a convolver (52) which convolves the lines of data with deconvolution functions from a memory (54) in a recursive loop (56) in accordance with a relative angular position of the x-ray tube and the corresponding detector. Filtered data lines are subtracted (56) from unconvolved data values to generate values with reduced off-focal radiation components. Filtered data is conveyed to a convolver (82) and a backprojector (84) which backprojects the data across an image memory (86).

Abstract: A two-dimensional radiation detector (30) receives divergent ray penetrating radiation along conical or pyramidal ray paths which converge at an apex. The radiation may emanate from one or both of an x-ray tube (16) or radionuclei injected into a human subject. As the radiation detector rotates around the subject, it is repeatedly sampled to generate two-dimensional data sets h.sub.t (u,v) at each of a plurality of samplings t. Each two-dimensional data array is weighted (44) and divided into two components. A first component processor (48) convolves a first component of each two-dimensional array with respect to u for each v. A second component processor (50) processes a second component of each two-dimensional array (FIG. 6) including weighting each component with a geometrically dependent weighting function W from a weighting function computer (46). The processed first and second components are combined (52) and backprojected (54).

Abstract: A CT scanner (10) includes a reconstruction processor (82) for reconstructing an image from digital signals from detector arrays (20). Each detector array includes scintillation crystals (22) arranged in an array for converting x-ray radiation into visible light. An array of photodetectors (24) is mounted beneath the scintillation crystal array for converting light emitted from the scintillation crystals into electrical charge. The combination scintillation crystal, photodetector array is mounted to and preferably integral with an integrated circuit (26) having charge storage devices (32, 34). Electrical charge generated by each photodetector is integrated and stored alternately by a corresponding pair of charge storage devices. The charge storage devices operate as a double buffer in which one charge storage device accumulates charge while the other holds an accumulated charge until read out by downstream circuitry.

Abstract: A CT scanner (10) has an x-ray source (12) which transmits a plurality (N) of x-ray fans across an examination region (14) to a plurality of rings of radiation detectors (28). An adjustable septum (80) adjusts a gap between the fan beams and outer collimators (82) adjust the width of the fan beams such that the effective spacing and width are adjusted. The x-ray source rotates around the examination region as a subject support (32) moves longitudinally through the examination region such that the x-ray fans move along interleaved spiral trajectories. Radiation attenuation data from each of the x-ray fans is combined and reconstructed into a plurality of images along parallel slices orthogonal to the longitudinal axis. The spacing between the x-ray beams is adjusted relative to the effective pitch of the spirals such that the leading edges of the x-ray fans intersect a common transverse plane at 180.degree./N angular intervals around the longitudinal axis.

Abstract: An x-ray source (14) is rotated along a non-circular path (18) to irradiate a subject in an examination region (10) with a fan beam (16) of radiation. Radiation detectors convert rays of the fan beam which have traversed the examination region into electronic data which is stored as data fans in an initial data memory (22). Each data fan is zero-filled (24) and convolved by a convolver (30). Preferably, the convolver transforms each data fan into Fourier-space (32), filters each data fan in Fourier-space with a roll-off filter (36), and converts the filtered data fan back from Fourier-space (38). Each fan is weighted (42) by a 1/cos weighting function and stored in a weighted data memory (44). Rays which are redundant in a 180.degree.+ fan reconstruction are removed (46) from the convolved data fans. A pixel driven backprojector (50) backprojects each convolved and redundant ray removed data fan into an image memory (56).

Abstract: During surgery, a physician speaks commands that are received by a microphone (10). A speech processor (12) converts audio signals from the microphone into word signals. A command interpreter (14) compares each word signal with a list of previously authorized command words. When the word signal corresponds to one of the preselected command words, a corresponding command signal is generated and sent to a volume imager (18), a video recorder (20), a hard copy, printer (28), or other system component. The volume imager generates an image representation signal indicative of a portion of image data stored therein which is displayed on a video monitor (B) or recorded on the video recorder.

Abstract: A C.T. scanner (A) or other medical diagnostic imager generates volume image data which is stored in an image memory (B). An image memory access controller (30) withdraws selected pixel values from the volume image memory which it provides a video processor (20) to formulate the appropriate video signals for display on a video monitor (C). An operator defines a volume on an operator controller (32). A plurality of interconnected polygons define the three-dimensional volume specified by the operator. The video processor converts these defined polygons into an appropriate display on the monitor. The polygons are selected to define a subregion to be viewed. The polygons are rotated about one or more of (x), (y), and (z) axis, scaled, and translated along these axis to select an orientation from which the polygon and enclosed data is to be viewed. The rotation, scaling and translation is stored (38, 40, 42).

Abstract: Radiation from an x-ray source (14) is collected by radiation detectors (16) to generate an original set of CT line integral data (18). The original CT data is subject to a first order correction and reconstructed (20) to generate a first order corrected image representation (22). A first artifact image representation (26) is generated (24) from the first order corrected image. The artifact image representation and the first order corrected image representation are subtractively combined (28) to generate a second order corrected image representation (30). A bone and high density clip image representation (142) is generated from the second order corrected image. The bone and clip image representation is forward projected (144) to generate a first high contrast line integral data set (146). High contrast model line integral data corresponding to rays which traverse the high density clip are topologically transposed (148) for corresponding rays of the original CT data.

Abstract: A CT or other radiographic scanner (A) generates data that is arranged into sets (32). Each set is convolved (40) with a convolution function (42) and backprojected (44) into an image memory (46) along a corresponding one of a plurality of rays. A corresponding gradient image (52) in which each pixel value has either a one or a zero value is forward projected (54) and compared (60) with a standard. The comparison indicates along which rays data sets including bad data were projected. To subtract the bad data contribution from the image, the image representation is forward projected (90) along the identified rays, convolved (40) with a negative of the convolution function (84), and backprojected (44) along the identified ray into the image memory (46). Further correction may be obtained by replacing the subtracted data with interpolated data.

Abstract: A system for three-dimensional diagnostic imaging generates a plurality of slice images of a specimen. A region of interest is selected from within a slice and is extrapolated to subsequent slices. A boundary indicative of a surface of interest is selected from within the region of interest to facilitate generation of an image representative of a three-dimensional surface of interest to be assembled from subsequent slices of the plurality. A viewing surface is defined in relation to a generated surface image which was selected from the boundary. A scaling means assigns a scaled gray level to the three-dimensional image to facilitate three-dimensional viewing of the object when it is projected on the viewing surface. Image information is selectably modified by data from the original slice images to add surface density visualization. Means is also provided to facilitate selective segmentation of a three-dimensional image along a plane or planes of interest.

Abstract: A diagnostic imaging system generates a three-dimensional display from a series of two-dimensional slice images. A region of interest, defined from a boundary of interest, is selected from one slice and is extrapolated to subsequent slices. Pixels representative of the boundary of interest are isolated and represented by three vectors having an equivalent entries in each. First and second vectors store data representative of first and second coordinates for pixels within each slice. Entries in the third vector corresponds to physical properties of a specimen at a location defined by corresponding entries in the first and second vectors. Areas representative of boundaries of interest falling between slices are extrapolated from vector data from slices neighboring the area. This is accomplished by a linear interpolation of elements of the set of smaller vectors to a number equivalent to the entries in the neighboring larger vectors.

Abstract: A system for three-dimensional diagnostic imaging generates a plurality of slice images of a specimen. A region of interest is selected from within a slice and is extrapolated to subsequent slices. A boundary of indicative of a surface of interest is selected from within the region of interest to facilitate generation of an image representation of a three-dimensional surface of interest to be assembled from subsequent slices of the plurality. A viewing surface is defined in relation to a generated surface image which was selected from the boundary. A scaling means assigns a scaled gray level to the three-dimensional image to facilitate three-dimensional viewing of the object when it is projected on the viewing surface. Image information is selectably modified by data from the original slice images to add surface density visualization. Means is also provided to facilitate selective segmentation of a three-dimensional image along a plane or planes of interest.

Abstract: In a fourth generation CT scanner, source views or data sets are generated for reconstruction processing. A fan beam (16) of radiation rays is rotated around an image region (12) to irradiate subsets of detectors of a detector ring (10). A data sampler (B) samples the detectors of each irradiated subset a plurality of times, each time with the radiation fan beam displaced incremently from the preceding time to generate a plurality of the source views or data sheets from the same detectors. A plurality of consecutive source views or data sets are interleaved to produce a signal interleaved view or data fan. More specifically, the data sets are stored in data set memories (20-26) and interleaved serially into a data fan memory (30). Each time the fan beam rotates sufficiently to irradiate a different detector subset, an additional plurality of data sets are generated and interleaved into another data fan.

Abstract: A method for approximating X-ray detector data from malfunctioning computed tomography detectors. A view of data is gathered for each of a plurality of detectors in a stationary detector scanner. Data for each malfunctioning detectors is approximated by interpolating data from detectors closely positioned with respect to the malfunctioning detectors after regions of high contrast within a cross-section of interest have been identified.