Search

In the miniature scintillation camera that is composed of a position-sensitive photomultiplier tube (PSPMT) and a scintillation crystal, a computer simulation of maximum-likelihood position estimation (MLPE) for better correction of image than that by a conventional algorithm has been studied. So far MLPE has been proposed in several papers, and mostly experiments finding look-up table (LUT) is necessary to apply this method. In this paper, the MLPE using a simulation method instead of experiments, which can consider all stochastic processes of converting the location of gamma-ray interaction with scintillator into position signals, has been proposed, implemented, and compared to conventional algorithm. From the analysis and comparison of images, the quality of image was better than that of resulting from the conventional algorithm. We verified this simulation method through comparison this simulation images with experimental images using small gamma camera of a NaI(Tl) and a PSPMT. From the result, it is proposed that this simulation method could be applied for all PSPMT with only one simulation instead of every tedious experiment to obtain the LUT for different PSPMT even with the same specification.

Many PET scanners nowadays have the possibility to record event-by-event information, known as listmode data. This has the advantage to keep the data in the highest possible resolution (both temporal and spatial). In most cases, listmode data are then binned into sinogram format before reconstruction. In this paper, we discuss at which stage normalisation factors should be introduced. It is shown that noise is greatly reduced by performing the normalisation after the binning. We also briefly discuss the consequences for strictly listmode-based algorithms. We illustrate this with acquired and simulated data for the quad HiDAC camera.

This paper addresses tomographic reconstruction of binary objects from a small number of noisy projections in applications where the global dose remains constant with an increase or a decrease of the number of projections. A region based reconstruction method using curve evolution is proposed. To manage the lack of data, the tomographic reconstruction problem is defined as a domain optimization problem. A geometrical criterion directly embedded in a dynamical scheme is defined. The minimization of this criterion drives the region determination. This scheme is implemented using a level set method including regularization via local curvature. The quality of objects reconstructed from noiseless projections is satisfying even without regularization. However when the number of projections or the signal to noise ratio is decreased, reconstructed images suffer from geometrical degradations. We show the improvement of the results brought by regularization in drastic conditions such as reconstruction from three noisy projections.

Functional images obtained PET and/or SPECT become more useful when those images are provided with detail anatomical information obtained by X ray CT or MRI. A series of image processing including image registration and segmentation is presented for quantitative analysis of functional images. A clinical application described in the paper is to evaluate the effect of radiation therapy for lung cancer on lung functions quantitatively. Though SPECT images give information on lung functions such as ventilation and perfusion, it is difficult to identify correctly the location and amount of radioactivity distribution with only those images. To overcome this difficulty, we synthesize both anatomical image (X ray CT, or shortly XCT) and functional image (SPECT) effectively. Furthermore, we have developed a method for dividing lung in XCT image into the lobes that are anatomically meaningful. The segmented lobes in XCT image can be used for quantitative evaluation in each lobe. Change in perfusion at each lobe of lung along with the therapy is presented as an effective example.

Multi-slice computed tomography (MCT) is one of the most recent technology advancements in CT. Because of the large number of detector rows placed along the patient long axis, MCT offers improved volume coverage, better contrast utilization, and increased spatial resolution. Despite many of its advantages, MCT brings new challenges to the image reconstruction. In general, the challenges of multi-slice CT reconstruction can be classified into two categories: the effects of cone beam and helical interpolation. The cone beam effect comes from the fact that the scanning planes for the outer detector rows are no longer parallel to the plane-of-reconstruction. The helical interpolation effect is mainly due to the inherent projection data inconsistency as a result of the constant table motion. In this paper, we present an algorithmic approach to address both issues. We first propose a weighted hyper-plane reconstruction (WHR) to overcome the sharp transitions in the helical weight. Using WHR, a weighting function is derived based on the scaled average of a set of weights obtained from multiple reconstruction hyper-planes. This approach effectively enables the derivation of a smooth weighting function from any non-smooth functions. We then present a cone-angle dependent weighting (CADW) scheme to overcome the cone beam effects of the outer detector rows. Extensive phantom and clinical experiments have shown that the proposed algorithm is superior to the previously proposed weighting approaches, in terms of reduction of image artifacts and preservation of slice thickness. At the same time, the algorithm utilizes only fan beam back-projections, instead of the computationally more intensive cone beam back-projection.

This paper proposes a hierarchical structure for the data transmission of medical volume data with three-dimensional (3-D) wavelet transform and a loss-less coding scheme. Technological development of CT and MRI has enabled the acquisition of volumetric medical data consisting of a large number of slice images. PACS (Picture Archiving and Communication System) have been popularized and the data are going to be increasingly used in networks. However, transmission delay and storage costs are outstanding problems for large volumes of data. The proposed structure improves them dramatically using a new 3-D wavelet transform scheme. Furthermore, the scheme improves the quality of the rendered images reconstructed at the intermediate stages in hierarchical transmission.

In this paper, we propose a reconstruction methodology where PET data are post-processed using Poisson denoising algorithms developed by Timmermann and Nowak (1999). The denoising algorithms, which follow from a Bayesian framework and multiscale signal representation, compute minimum mean square error estimates of the unknown means of independent Poisson random variables from their observations. Timmermann and Nowak estimated certain parameters needed by their algorithms using the method of moments and observed data. Instead, we estimate these parameters more accurately using a maximum likelihood method and off-line training procedure. We demonstrate the utility of the proposed image reconstruction methodology by applying it to synthetic data and real positron emission tomography data.

4D CT is a dynamic volume imaging system of moving organs with an image quality comparable to conventional CT. Dynamic cone-beam CT can realize it with several breakthroughs. They are: 1) large-area 2-dimensional (2D) detector, 2) high-speed data transfer system, 3) reconstruction algorithm, 4) ultra-high-speed reconstruction computer and 5) high-speed and continuous rotating gantry. Among them the development of the 2D detector is one of the biggest tasks because it should have as wide dynamic range and high data acquisition speed (view rate) as present CT detectors. We are now developing a 4D CT-scanner together with the key-components. It will take one volume in 0.5 sec with a 3D matrix of 512 /spl times/ 512 /spl times/ 512. This paper describes its concepts and design, as well as preliminary development of the 2D detector.

Recently a number of approaches solving the long object problem with spiral scan appeared in the literature. Unlike the algorithms which employ circular scans for solving the long object problem, with only spiral scan it is necessary to scan regions of the object adjacent to the ROI in order to reconstruct the ROI without contamination; the spiral path required beyond the ROI is referred to as overscan in the literature. Overscan exposes the regions in the long object adjacent to the ROI to radiation as well as incurs additional computation time In this paper we present a number of methods to reduce the undesirable overscan, with emphasis on the local ROI algorithm.

Recently, independent component analysis (ICA) has been introduced to solve the blind source separation problem. In the original and extended versions of ICA, nonlinearity functions are fixed to have specific forms such as supergaussian or subgaussian, limiting their performance. In this paper, we utilized ICA with mixture density model such that any assumption about the source density is not required, thus better separation is possible by matching flexible parametric nonlinearity to any kind of density of sources. Through simulation studies, the algorithm was validated and its better performance was demonstrated in comparison to other versions of ICA. Then mixture density ICA was applied to functional magnetic resonance imaging (fMRI) and electroencephalogram (EEG) data to localize the independent sources for alpha activity. We found that there is a strong spatial correlation between the sources in fMRI and EEG, proving the usefulness of our approach in its application to source separation problem in biomedical signal processing.

The non-stationary Poisson noise in single photon emission computed tomography (SPECT) sinogram is a major cause to compromise the quality of reconstructed images and challenge any compensation strategy for photon attenuation, scatter and collimator response. Research utilizing three-dimensional Wiener filter after Anscombe transformation, which converts Poisson distributed noise into Gaussian distributed one, has been conducted recently with noticeable success. However, the prerequisite of stationary random process for the Wiener approach is not exactly valid for the SPECT sinogram. In this paper, the non-stationary Poisson noise was first modulated by the Anscombe transformation and then the modulated noise model was analyzed. A Kalman filter for the modulated non-stationary noise model was designed to extract the means or signals from the Poisson noise sinogram. Monte Carlo program was used to generate projection data from the MCAT phantom, simulating SPECT data acquisition. The reconstructed results demonstrated a significant improvement with the Kalman filter, as compared to the Wiener approach. Dramatic improvement is seen, as compared to linear low-pass filters, at noisy levels of 100 thousand counts in a 128/spl times/128/spl times/128 sinogram size. The capability of Kalman filter for nonstationary noise model was theoretically proved and experimentally demonstrated.

Segmented-based attenuation correction is now a widely accepted technique to reduce noise contribution of measured attenuation correction. In this paper, we present a new method for segmenting transmission images in positron emission tomography. This reduces the noise on the correction maps while still correcting for differing attenuation coefficients of specific tissues. Based on the Fuzzy C-Means (FCM) algorithm, the method segments the PET transmission images into a given number of clusters to extract specific areas of differing attenuation such as air, the lungs and soft tissue, preceded by a median filtering procedure. The reconstructed transmission image voxels are therefore segmented into populations of uniform attenuation based on the human anatomy. The clustering procedure starts with an over-specified number of clusters followed by a merging process to group clusters with similar properties and remove some undesired substructures using anatomical knowledge.

SPECT using a rotating multi-segment slant-hole (RMSSH) collimator provides a superior alternative for imaging the breast and the heart comparing to conventional SPECT. The aim of this paper is to investigate the effects of the number of camera positions and the number of collimator rotation stops at each camera position on RMSSH SPECT reconstruction image quality, using both Monte-Carlo simulation and physical phantom experiments. Our results demonstrate the importance of sufficient camera positions for complete angle data acquisition and accurate 3D RMSSH SPECT image reconstruction; and for the same total acquisition time, tradeoff exists between the number of collimator rotation stops and the amount of acquisition time at each rotation stop. We conclude that complete angle data acquisition is crucial for accurate 3D RMSSH SPECT image reconstructions, but using inadequate number of camera positions may also result in usable diagnostic information in shorter acquisition time. An optimal trade-off between image resolution, noise and image artifacts is needed to determine the number of collimator rotation stops at each camera position.