Search PubMed⌕ Search

Biomedical subjects

C E Floyd

Publications and source records attributed to C E Floyd.

At least 73 records · Page 4Linked to original sources
PubMed1989

Scatter compensation in digital chest radiography using Fourier deconvolution.

The authors present a numerical deconvolution technique to compensate for image degrading effects caused by scattered photons in radiographic chest images. Fourier transform techniques are used to deconvolve a shift invariant model of the two dimensional point spread response functions of the scattered radiation. This approach uses a digitized radiograph acquired with a standard chest imaging protocol, so no specialized imaging equipment is required. While the shift variant shape of the scatter model is optimized for the lung field, effective compensation is provided when this model shape is applied to other chest regions. Preliminary evaluation suggests that this technique can provide improved image contrast over the entire chest region.

Computer Simulation↗
PubMed1988

Noise characteristics for cone beam collimators: a comparison with parallel hole collimator.

In order to evaluate the properties of a cone beam (CB) collimator and three-dimensional filtered backprojection algorithm, the noise characteristics of this collimator configuration were determined and comparisons with a parallel hole (PH) collimator were made. Noise characteristics were evaluated using two approaches: the first consisted of assessing the magnitude of local random fluctuations in the reconstructed images, and the second consisted of assessing the noise texture in these images in the frequency domain by evaluating the noise power spectrum. Data used for these measurements were simulated using Monte Carlo models of SPECT systems equipped with cone beam and parallel hole collimators. Finally, to compare experimentally a specially designed high resolution CB collimator with a high resolution (HRES) PH collimator, measurements of a physical phantom were performed. Results of our studies show better noise magnitude for CB collimators; however, for CB collimators with short focal lengths (40-60 cm) the shape of %RMS noise distributions differs from slice to slice.

Computer Simulation↗
PubMed1987

Experimentally measured scatter fractions and energy spectra as a test of Monte Carlo simulations.

A method for the validation of Monte Carlo photon transport calculations is presented, with particular emphasis on the scatter component of such calculations. The method is based on a quantitative comparison of calculated and experimental scatter fractions. In addition, the method includes a qualitative comparison of point spread functions and energy spectra. An application of the method is demonstrated by comparing the results of an existing Monte Carlo code with experimental results obtained with a gamma camera viewing a point source of 99Tcm (140 keV gamma rays) centred within a water-filled cylinder. The results of the comparisons show good agreement between experiment and calculation. These results allow the code to be used with increased confidence in a variety of situations, and they define more precisely the region of applicability of the code. In addition, the determination of scatter fractions and energy spectra is useful for other applications. For example, scatter fractions can be a useful parameter for evaluating possible techniques for scatter compensation.

Computer Simulation↗
PubMed1987

Convergence of the maximum likelihood reconstruction algorithm for emission computed tomography.

Convergence properties of the maximum likelihood estimator (MLE) for emission computed tomographic (ECT) image reconstruction are evaluated as a function of Poisson noise, precision of the assumed system resolution model and iteration number up to 10,000 iterations. In the ECT reconstruction problem, the photon-emitting source distribution is to be estimated from measurements of projections of the emitted photon flux. The MLE algorithm seeks a source distribution which will maximise the maximum likelihood function relating the estimated and the measured projections. A Monte Carlo model of the system transfer function of a single photon emission computed tomographic (SPECT) system allowed realistic projection data to be simulated from a known source distribution. Poisson noise was added to the Monte Carlo simulations. By using projection data from a known source distribution generated through a known system transfer function, we were able to simultaneously evaluate the convergence of both the projection estimations as well as the source distribution estimations. As predicted by theory, the estimates of the projections did continue to improve (or remain the same) for all combinations of Poisson noise (up to 10% RMS) and system resolution (+/- 10% of true value) tested. Convergence of source distribution estimates to the true value was found for up to 10,000 iterations only for low noise (0.1% RMS) with the correct resolution function. For all other combinations, there was some optimum iteration (between 30 and 400) after which the source estimate was degraded even though the estimate of the projections was improved.

Algorithms↗
PubMed1987

Brain phantom: high-resolution imaging with SPECT and I-123.

Inverse Monte Carlo (IMOC) is a unified reconstruction algorithm for single photon emission computed tomography (SPECT) that provides simultaneous compensation for attenuation and collimator divergence. IMOC was applied to the reconstruction of SPECT images of a brain phantom with iodine-123 and high-resolution collimation. Projection sets containing 80,000, 540,000, and 5.2 million counts were reconstructed. Comparison with filtered back-projection reconstructions showed that the IMOC reconstructions provided superior noise and resolution characteristics at all three photon densities. Results of this study indicate that IMOC may allow the use of high-resolution, low-sensitivity collimation for SPECT studies, which have traditionally provided photon yields too low for useful imaging.

Algorithms↗
PubMed1987

Nonisotropic attenuation in SPECT: phantom tests of quantitative effects and compensation techniques.

A quantitative study of nonisotropic attenuation in SPECT imaging is presented. The study includes a case where the spatial distribution of the attenuation coefficient is nonuniform, as well as a case where the photon path length in the attenuating medium is variable as a function of direction. The effects are studied using phantoms with known source activity and density distributions. Reconstructed images of the phantoms with and without attenuation compensation are compared with the source distribution. Three methods are used to provide partial attenuation compensation, using effective attenuation coefficients. These coefficients include some of the effects of photon scatter, but scatter is not explicitly treated. One attenuation compensation method involves a multiplicative postprocessing correction using an assumed constant attenuation coefficient. A modification of this technique is implemented using the correct nonuniform attenuation map to determine the multiplication factors. A single-iteration technique is used to provide a more complete compensation. The results indicate that nonuniform attenuation can produce significant distortion in line spread functions and in larger distributed sources. This distortion can alter volume determinations, quantitation measurements, and the shape of small objects, and can cause misplacement of counts into regions of low density. The distortion cannot be eliminated by the multiplicative postprocessing correction, but the single-iteration technique can significantly decrease the distortion.

Models, Structural↗
PubMed1986

Inverse Monte Carlo as a unified reconstruction algorithm for ECT.

Tomographic reconstruction for single photon emission computed tomography (SPECT) with simultaneous compensation for attenuation, scatter, and distance dependent collimator resolution is provided by an Inverse Monte Carlo (IMOC) reconstruction algorithm. A detection probability matrix is formed by Monte Carlo solution to the photon transport equation for SPECT acquisition from a unit source activity in each reconstruction source voxel. The measured projection vector will equal the product of this detection probability matrix with the unknown source distribution vector. The resulting large, nonsparse system of equations is solved for the source distribution using an iterative Maximum Likelihood EM estimator. Reconstruction of experimentally acquired projections from phantoms shows quantitative compensation for scatter and attenuation. Comparison with filtered backprojection (FBP) reconstruction shows an improvement in resolution recovery, contrast, and signal-to-noise for the IMOC algorithm. Reconstruction of clinical studies shows improved contrast, structural resolution, and noise characteristics.

Humans↗
PubMed1985

Deconvolution of Compton scatter in SPECT.

A deconvolution algorithm has been developed which compensates for Compton scattering in SPECT images. Compton scatter is modeled as a convolution of the nonscattered projection data with an exponential function. Deconvolution of the total (scatter + nonscatter) projection data yields compensated true projection. Using Monte Carlo methods, the scattered and nonscattered components of a SPECT image are simulated thus allowing a comparison of scatter compensated results with direct nonscatter results. The quality of the compensation is evaluated by comparing the ratio of total to direct counts with the ratio of compensated to direct counts. This deconvolution technique has been developed and evaluated for experimentally acquired SPECT data as well as for simulated data.

Filtration↗
PubMed1984

Energy and spatial distribution of multiple order Compton scatter in SPECT: a Monte Carlo investigation.

Energy and spatial projection distributions were simulated for gamma camera imaging of multiple order Compton scattered photons. SPECT imaging of a line source of radioactivity located in a water filled cylindrical phantom was modelled using Monte Carlo techniques. Photon trajectories were followed from emission to detection including the effects of all physical interactions and the resulting energy spectra and spatial projections were sorted as a function of the number of times the photon underwent Compton scattering before detection. Analysis of energy spectra demonstrates that Compton events up to second order overlap with the non-scattered events and distributions are peaked at lower energies as the scattering order increases. Analysis of spatial projections shows that, with increasing order, Compton events produce tails on the line spread function which progress from roughly exponential to nearly flat distributions. The use of Monte Carlo modelling thus allows a detailed investigation of the spatial and energy distribution of Compton scatter which could not be performed using present experimental techniques.

Models, Theoretical↗
PubMed1984

Improved SPECT quantification using compensation for scattered photons.

SPECT images are degraded by the inclusion of Compton-scattered photons within the pulse-height window. Phantom and patient studies with Tc-99m were used to evaluate a compensation method that consists of subtracting a fraction of the image reconstructed using events recorded within a secondary pulse-height window (92-125 keV) from that derived from the photopeak pulse-height window (127-153 keV). Images of line sources in air and in a water-filled phantom were stored. The compensated line spread functions (LSFs) were evaluated. In water, the absolute counting rates for the SPECT LSFs were within 10% of the rates measured in air. The phantom consisted of six solid acrylic spheres (diameters 10, 13, 16, 19, 25, 32 mm) placed within a cylindrical (22 cm diam) distribution of Tc-99m. For sphere diameters greater than 25 mm, the measured image contrasts were within 8% of the true uptake ratios. Our results have shown that high-quality, camera-based SPECT systems can reconstruct artifact-free images by making additional use of projection data acquired in a pulse-height window other than that over the primary photopeak. This compensation method results in qualitative and quantitative improvements for the limited source geometries investigated. Further studies are required to optimize this heuristic approach for other source geometries.

Humans↗