156 results about "Tomographic reconstruction" patented technology

A tomographic reconstruction method and system incorporating Bayesian estimation techniques to inspect and classify regions of imaged objects, especially objects of the type typically found in linear, areal, or 3-dimensional arrays. The method and system requires a highly constrained model M that incorporates prior information about the object or objects to be imaged, a set of prior probabilities P(M) of possible instances of the object; a forward map that calculates the probability density P(D|M), and a set of projections D of the object. Using Bayesian estimation, the posterior probability p(M|D) is calculated and an estimated model M_EST of the imaged object is generated. Classification of the imaged object into one of a plurality of classifications may be performed based on the estimated model M_EST, the posterior probability p(M|D) or MAP function, or calculated expectation values of features of interest of the object.

The present discussion relates to the use of deep learning techniques to accelerate iterative reconstruction of images, such as CT, PET, and MR images. The present approach utilizes deep learning techniques so as to provide a better initialization to one or more steps of the numerical iterative reconstruction algorithm by learning a trajectory of convergence from estimates at different convergence status so that it can reach the maximum or minimum of a cost function faster.

A dynamic SPECT camera is provided, comprising, a plurality of single-pixel detectors, a timing mechanism, in communication with each single-pixel detector, configured for enabling time-binning of the radioactive emissions impinging upon each single-pixel detector to time periods not greater than substantially 30 seconds, and a position-tracker, configured for providing information on the position and orientation of each detecting unit, with respect to the overall structure, substantially at all times, during the individual motion, the dynamic SPECT camera being configured for acquiring a tomographic reconstruction image of a region of interest of about 15×15×15 cubic centimeters, during an acquisition time of 30 seconds, at a spatial resolution of at least 10×10×10 cubic millimeter. The dynamic camera is configured for very short damping time, and may further acquire images in a stationary mode, with no motion. It is further configured for time binning at dynamically varying time-bin lengths, dynamically determining a spectral energy bin for each detecting unit, and employing an anatomic construction of voxels in the imaging and reconstruction.

A touch-sensitive apparatus comprises a panel for conducting signals from incoupling points to outcoupling points. Actual detection lines are defined between pairs of incoupling and outcoupling points to extend across the panel. A signal generator is coupled to the incoupling points to generate the signals, e.g. light, and a signal detector is coupled to the outcoupling points to generate an output signal. A data processor processes the output signal to generate a set of data samples, which are non-uniformly arranged in a two-dimensional sample space. Each data sample is indicative of detected energy on a detection line and is defined by a signal value and first and second dimension values in the sample space. The first and second dimension values define the location of the detection line on the surface portion. Adjustment factors are obtained for the data samples, each adjustment factor being representative of the local density of data samples in the sample space. The data samples are processed by tomographic reconstruction, while applying the adjustment factors, to generate a reconstructed distribution of an energy-related parameter

The present approach relates to the use of machine learning and deep learning systems suitable for solving large-scale, space-variant tomographic reconstruction and/or correction problems. In certain embodiments, a tomographic transform of measured data obtained from a tomography scanner is used as an input to a neural network. In accordance with certain aspects of the present approach, the tomographic transform operation(s) is performed separate from or outside the neural network such that the result of the tomographic transform operation is instead provided as an input to the neural network. In addition, in certain embodiments, one or more layers of the neural network may be provided as wavelet filter banks.

An interactive touchscreen system comprising a conductive panel having a surface portion adapted to receive touch interactions, the conductive panel operatively coupled to a tomograph configured for determining conductivity of the conductive panel and further adapted for incrementally improving a reconstructed tomogram estimation by integrating the output of a tomographic reconstruction algorithm having as input a unreconstructed tomogram delta, the reconstructed tomogram estimation being representative of the panel conductivity properties at locations spatially correlated to a plurality of regions notionally defined over the surface portion and thereform indicative of the touch interactions, the tomograph further processing and outputting said reconstructed tomogram estimation and an indication thereof.

Motion of an object of interest, such as a cell, has a variable velocity that can be varied on a cell-by-cell basis. Cell velocity is controlled in one example by packing cells into a capillary tube, or any other linear substrate that provides optically equivalent 360 degree viewing access, so that the cells are stationary within the capillary tube, but the capillary tube is translated and rotated mechanically through a variable motion optical tomography reconstruction cylinder. The capillary tube motion may advantageously be controlled in a start-and-stop fashion and translated and rotated at any velocity for any motion interval, under the control of a computer program. As such, there are several configurations of the optical tomography system that take advantage of this controlled motion capability. Additionally, the use of polarization filters and phase plates to reduce light scatter and diffraction background noise is described.

Methods and systems for obtaining improved computed tomographic reconstructions are provided. A camera obtains one or more images of a patient while a tomographic scan is performed. A CT scanner obtains one or more tomographic projections. The total number of the one or more tomographic projections obtained during the CT scan is determined. For each of the one or more patient images obtained, a processor correlates the patient image with the one or more tomographic projections, calculates a position of the patient and determines if the calculated position of the patient is greater than one or more predetermined constants. If it is determined that the calculated position of the patient is greater than one or more predetermined constants and a total number of the one or more tomographic projections obtained is less than a predetermined number, separate projections are substituted in the place of the one or more tomographic projections that were correlated to the patient image. If it is determined that the calculated position of the patient is greater than one or more predetermined constants and a total number of the one or more tomographic projections is greater than a predetermined number, the tomographic scan is aborted or the tomographic reconstruction is reconstructed on a reduced arc.

Concealed object detection using electromagnetic and acoustic multistatic imaging systems and methods. A method of simultaneously screening plural subjects for concealed objects includes transmitting a signal into a screening area where there is at least one subject to be screened having an associated object, receiving a reflected signal from the object when the object is located within the screening area, processing the reflected signal using multistatic Fourier space sampling and tomographic reconstruction to generate a three-dimensional image of the object and displaying the three-dimensional image. The transmitting and receiving are performed using a multi-directional array including at least three sensors. An object detection system includes a screening area, a multi-directional array including at least three sensors, a processor configured to execute multistatic Fourier space sampling and tomographic reconstruction and a display.

An operator shielded X-ray imaging system has sufficiently low mass (less than 300 kg) and is compact enough to enable portability by reducing operator shielding requirements to a minimum shielded volume. The resultant shielded volume may require a relatively small mass of shielding in addition to the already integrally shielded X-ray source, intensifier, and detector. The system is suitable for portable imaging of well cores at remotely located well drilling sites. The system accommodates either small samples, or small cross-sectioned objects of unlimited length. By rotating samples relative to the imaging device, the information required for computer aided tomographic reconstruction may be obtained. By further translating the samples relative to the imaging system, fully three dimensional (3D) tomographic reconstructions may be obtained of samples having arbitrary length.

A method and apparatus for automated reconstruction of a 3D computed tomography (CT) volume from a small number of X-ray images is disclosed. A sparse 3D volume is generated from a small number of x-ray images using a tomographic reconstruction algorithm. A final reconstructed 3D CT volume is generated from the sparse 3D volume using a trained deep neural network. A 3D segmentation mask can also be generated from the sparse 3D volume using the trained deep neural network.

The invention relates to systems and methods for tomographic imaging in diffuse media employing a fast reconstruction technique. A hybrid Fourier approach is presented that enables the fast tomographic reconstruction of large datasets. In certain embodiments, the invention features methods of in vivo fluorescence molecular tomographic (FMT) reconstruction of signals, reporters and/or agents (i.e., contrast agents or probes) in a diffusive medium (e.g., a mammalian subject). The method preserves the three-dimensional fluorophore distribution and quantitative nature of the FMT approach while substantially accelerating its computation speed, allowing FMT imaging of larger anatomies.