D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects - Part I: Performance Analysis", IEEE Trans. Acoustic Speech and Signal Processing, pp. 886--897, 1984.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....[6, 7] but the large number of required excitations makes cardiac gating impractical. Ungated sequences often suffer from sensitivity to non uniform flow and vessel motion. Projection reconstruction from only a few views is an ill conditioned problem in general, but as observed by Rossi [8], the ultimate goal of processing the projection measurements is typically far more modest than obtaining high resolution cross sectional imagery. In fact, the goal is typically to obtain quantitative descriptions of arterial shape (perhaps as an intermediate step towards the goal of evaluating ....

....two orthogonal projection angles. Often the restrictions on are only made implicitly, such as when diameter is computed from a single view. To compensate for limited views and low SNR, stronger assumptions are necessary. The model based approach of this paper is rooted in the work of Rossi [8], who analyzed reconstruction of a circular object from projections, and of Shmueli [18] who developed a dynamic programming algorithm for outlining a single cylindrical artery in one view. Pappas [19] demonstrated the accuracy of using elliptical cross sections to represent arteries, and Bresler ....

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects---Parts i & II: Performance analysis and robustness analysis," IEEE Tr. Acoust. Sp. Sig. Proc., vol. 32, pp. 886--906, Aug. 1984.

....lower bound (CRB) Cramer Rao lower bounds are widely used in problems where the exact minimum mean square error of an estimator is difficult to evaluate. While CRB s are available for estimation of signal parameters such as direction of arrival (DOA) 3] and size and orientation of a scatterer [4], only recently has this type of analysis been conducted for estimation of target shapes [5, 6] In [5] the boundary of a star like target is parameterized using B splines, and CRB s for the B spline coe#cients are computed for several shapes in a magnetic resonance imaging problem. However, the ....

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects - Part I: Performance analysis," IEEE Trans. on Acoustic Speech and Signal Processing, pp. 886--897, August 1984.

....to relate analytically the projections to the parameters. It is then easy to calculate analytically the projection data p(0) and to define the so lution either as the ML or as the LS estimate: arg nn lip p(0)ll 2 (3) This approach has also been used with success in image reconstruction [2, 3], but the range of applicability of these methods is limited to cases in which the parametric models are actually appropriate. The third approach, which is more appropriate to our problem of shape reconstruction, consists of using a function to directly model the contour of the object and ....

D. Rossi and A. Wilsky, "Reconstruction from projections based on detection and estimation of objects," ieeeASSP, vol. ASSP-32, no. 4, pp. 886-906, 1984.

....8) 4. n(r, qS) 15) where h(r, b; 8) has an analytic expression in 8. The LS or the ML estimate when the noise is assumed to be zero mean, white and Gaussian, is then given by: arg nn lip(r, qS) h(r, qs; O)II ) 16) This approach has also been used with success in image reconstruction [18, 19, 20, 21]. But, the range of applicability of these methods is limited to the cases where the parametric models are actually appropriate. The fourth approach which is more appropriate to our problem of shape reconstruction, consists in modeling directly the contour of the object by a function, say g( ....

D. Rossi and A. Wilsky, "Reconstruction from projections based on detection and estimation of objects," ieeeASSP, vol. ASSP-32, no. 4, pp. 886-906, 1984.

....The number of unknown parameters is drastically reduced compared to other methods, but the relation between the data and these parameters is no longer linear. Two subclasses of this approach can be derived: the methods which use simple shapes with a very few number of parameters (see for example [27]) and those which model the object shape by more general deformable tem plates (see for example [1, 10] In the following, we will specially focus on the case of polyhedral shapes. This modelisation is actually a generalization of the polygonal modelisation in 2D, which will be studied in a ....

....where h(r, b; 0) has an analytic expression in 0. If the noise is assumed to be zero mean, white and Gaussian, the LS and the maximum of likelihood estimate are equal and given by 0 = argnn 11p(r, h(r, O)112 . 1. 16) This approach has also been used with success in image reconstruction [27]. But its range of applicability is limited to the cases where the parametric models are actually appropriate. Deformable templates: The contour of the object can be modeled using spline functions for example. In this case, the parameters are the control points of the spline. The polygonal shape ....

D. Rossi and A. Willsky, "Reconstruction from projections based on detection and estimation of objects," IEEE Transactions on Acoustics Speech and Signal Processing, vol. ASSP-32, no. 4, pp. 886-906, 1984.

.... K I 2akbk ,2[ g v w ) if (c) h(r,c) O) dpk(r,c) with p(r,c) 0 elsewhere k:l (s) The LS or the ML estimate when the noise is assumed to be Gaussian is then given by: arg nn [ p(r, qb) h(r, qb; 0) 2 (10) This approach has also been used with success in image reconstruction [11, 12, 13, 14, 15]. But, the range of applicability of these methods is limited to the cases where the parametric models are actually appropriate. In the third approach which is more appropriate to our problem of shape reconstruction, one starts by modeling directly the contour of the object by a function, say ....

D. J. Rossi and A. S. Wilsky, "Reconstruction from projections based on detection and estimation of objects," ieeeASSP, vol. ASSP-32, no. 4, pp. 886 906, 1984.

....be able to relate analytically the projections to the parameters. It is then easy to calculate analytically the projection data p(O) and to define the so lution either as the ML or as the LS estimate: arg nn lip P(0)I[ This approach has also been used with success in image reconstruction [2, 3]. but the range of applicability of these methods is limited to cases in which the parametric models are actually appropriate. The third approach, which is more appropriate to our problem of shape reconstruction, consists of using a function to directly model the contour of the object and ....

D. Rossi and A. Wilsky, "Reconstruction from projections based on detection and estimation of objects," ieceA$$P, vol, ASSP-32, no. 4, pp, 886-906, 1984.

....ratio (SNR) performance of the MLE and for bounding the accuracy of any other unbiased estimate. 0018 926X 01 10.00 2001 IEEE 772 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 49, NO. 5, MAY 2001 While CRBs are available for estimation of signal parameters such as target location [8] [11], direction of arrival (DOA) 12] 15] and size and orientation of a scatterer [16] 18] only recently has this type of analysis been conducted for estimation of target shapes [7] In [7] the boundary of a star like target is parameterized using B splines, and CRBs for the B spline ....

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects---Parts I and II," IEEE Trans. Acoust. Speech Signal Processing, vol. ASSP-32, pp. 886--906, 1984.

....the projections to the parameters. It is then easy to calculate analytically the projection data p( and to define the solution either as the ML or as the LS estimate : b = arg min Phi kp Gamma p( k 2 Psi (3) This approach has also been used with success in image reconstruction [2, 3], but the range of applicability of these methods is limited to cases in which the parametric models are actually appropriate. ffl The third approach, which is more appropriate to our problem of shape reconstruction, consists of using a function to directly model the contour of the object and ....

D. Rossi and A. Wilsky, "Reconstruction from projections based on detection and estimation of objects," ieeeASSP, vol. ASSP-32, no. 4, pp. 886--906, 1984.

....where h(r; OE; has an analytic expression in . The LS or the ML estimate when the noise is assumed to be zero mean, white and Gaussian, is then given by: b = arg min Phi kp(r; OE) Gamma h(r; OE; k 2 Psi (16) This approach has also been used with success in image reconstruction [18, 19, 20, 21]. But, the range of applicability of these methods is limited to the cases where the parametric models are actually appropriate. ffl The fourth approach which is more appropriate to our problem of shape reconstruction, consists in modeling directly the contour of the object by a function, say ....

D. Rossi and A. Wilsky, "Reconstruction from projections based on detection and estimation of objects," ieeeASSP, vol. ASSP-32, no. 4, pp. 886--906, 1984.

....approach to discrete valued reconstruction detects parameterized objects directly in the projection domain. This strategy is applicable when the objective is to detect specific objects or regions such as tumors in medical imaging or material defects in non destructive testing. Rossi and Willsky [11] introduced this approach by performing maximum likelihood (ML) estimation of the location of a single object in the imaging plane. This concept was later extended to a three dimensional parameterization supporting multiple objects per plane [12] Here, constrained objects in 3 D are formed as a ....

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects -- parts I and II: Performance analysis and robustness analysis," IEEE Trans. on Acoustics, Speech, and Signal Processing ASSP-32, 886--906 (1984).

....likelihood estimator (MLE) Hence, the CRB can also serve as a predictor of the high SNR performance of the MLE, and for assessment of the accuracy of particular estimate produced by it from measured data. While CRB s are available for estimation of signal parameters such as target location [8 11], direction of arrival (DOA) 12 15] and size and orientation of a scatterer [16 18] only recently has this type of analysis been conducted for estimation of target shapes [7] In [7] the boundary of a star like target is parameterized using B splines, and CRB s for the B spline coe#cients ....

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects - Part I & II," IEEE Trans. on Acoustic Speech and Signal Processing, pp. 886--906, 1984.

....directly [6, 7] but the large number of required excitations makes cardiac gating impractical. Ungated sequences often suffer from sensitivity to non uniform flow and vessel motion. Projection reconstruction from only a few views is an ill conditioned problem in general, but as observed by Rossi [8], the ultimate goal of processing the projection measurements is typically far more modest than obtaining high resolution cross sectional imagery. In fact, the goal is typically to obtain quantitative descriptions of arterial shape (perhaps as an intermediate step towards the goal of evaluating ....

....two orthogonal projection angles. Often the restrictions on are only made implicitly, such as when diameter is computed from a single view. To compensate for limited views and low SNR, stronger assumptions are necessary. The model based approach of this paper is rooted in the work of Rossi [8], who analyzed reconstruction of a circular object from projections, and of Shmueli [18] who developed a dynamic programming algorithm for outlining a single cylindrical artery in one view. Pappas [19] demonstrated the accuracy of using elliptical cross sections to represent arteries, and Bresler ....

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects---Parts i & II: Performance analysis and robustness analysis," IEEE Tr. Acoust. Sp. Sig. Proc., vol. 32, pp. 886--906, Aug. 1984.

.... 2 k (OE) p g 2 k (OE) Gamma r 2 ) if r g k (OE) 0 elsewhere (9) The LS or the ML estimate when the noise is assumed to be Gaussian is then given by: b = arg min Phi kp(r; OE) Gamma h(r; OE; k 2 Psi (10) This approach has also been used with success in image reconstruction [11, 12, 13, 14, 15]. But, the range of applicability of these methods is limited to the cases where the parametric models are actually appropriate. ffl In the third approach which is more appropriate to our problem of shape reconstruction, one starts by modeling directly the contour of the object by a function, say ....

D. J. Rossi and A. S. Wilsky, "Reconstruction from projections based on detection and estimation of objects," ieeeASSP, vol. ASSP-32, no. 4, pp. 886--906, 1984.

....real data. Keywords: ellipsoids, geometric reconstruction, ejection fraction,myocardialperfusion I. INTRODUCTION This research presents a medical imaging application of the estimation of dynamically evolving ellipsoids from noisy projection observations. Much work in geometric reconstruction [1 5] has focused on reconstructing objects such as ellipsoids from noisy lower dimensional projections. This work di#ers from previous applications of ellipsoid reconstruction [3] since the precise ellipsoid dynamics and projection geometries are unknown. In particular, a method to obtain the ejection ....

.... a broad range of ellipsoid dynamics through the following evolution equation: X(k 1) A(k) X(k)A(k) 4) where changes such as magnification, rotation, and eccentricity change may be included in a simple way in A(k) For example, in the two dimensional case, one convenient choice for A(k)is(see[5]) A(k) c(k) #(k)0 01 #(k) cos #(k)sin#(k) sin #(k)cos#(k) 5) where the first term represents uniform scaling by the factor c(k) 0, the second term an area preserving stretching along the coordinate axes by #(k) 0, and the last term a rotation by an angle #(k) This general ....

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects-parts I and II: Performance analysis and robustness analysis," IEEE Trans. Acoustic, Speech, and Signal Processing, vol. 32, no. 4, pp. 886--906, 1984.

.... 2 As an aside we note that for problems in which the data are in a different domain from the image (e.g. tomographic problems) the calculation of the likelihood statistic is most easily done directly in the data domain rather then by first forming an image and then calculating the statistic [13]. rifices sensitivity to each individual hypothesis in e Hn in order to achieve some sensitivity to all of the hypothesis in e Hn . In contrast, the statistics used in the computationally intractable but optimal GLRT have, as we shall see, significantly greater sensitivity to each of hypotheses ....

D. J. Rossi and A. S. Willsky. Reconstruction from projections based on detection and estimation of objects---parts I and II: Performance analysis and robustness analysis. IEEE Transactions on Acoustics, Speech, and Signal Processing, ASSP-- 32(4):886--906, 1984.

....is nonempty then the sequence of cyclic projections will converge weakly to a point in this intersection [21] In addition to these these two general approaches there are other approaches that depend upon severely restricting the class of objects to be reconstructed. For example, Rossi and Willsky [22] use hierarchical maximum likelihood methods to estimate the position, radius, and eccentricity of objects with a known unit profile such as the unit disk. Soumekh [23] Chang and Shelton [24] and Fishburn et al. 25] have investigated reconstruction of 3 binary objects from a small number of ....

D. J. Rossi and A. S. Willsky. Reconstruction from projections based on detection and estimation of objects--parts I and II: Performance analysis and robustness analysis. IEEE Trans. ASSP, ASSP-32(4):886--906, 1984.

No context found.

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects - Part I: Performance Analysis", IEEE Trans. Acoustic Speech and Signal Processing, pp. 886--897, 1984.

No context found.

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects - Part I: Performance Analysis", IEEE Trans. Acoustic Speech and Signal Processing, pp. 886--897, 1984.

No context found.

D.J. Rossi and A.S. Willsky. Reconstruction from Projections Based on Detection and Estimation of Objects - Part I and II. IEEE ASSP,vol. 3.. no. 4,pages 886-906, 1984.

No context found.

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects - Part I: Performance analysis," IEEE Trans. on Acoustic Speech and Signal Processing, pp. 886--897, August 1984.

No context found.

D. J. Rossi and A. S. Willsky, "Reconstruction from projections based on detection and estimation of objects," IEEE Trans. Acoust. Speech, Signal Processing ASSP-32(4), pp. 886-906, 1984.

No context found.

D. Rossi and A. Wilsky, "Reconstruction from projections based on detection and estimation of objects," ieeeASSP, vol. ASSP-32, no. 4, pp. 886-906, 1984.

No context found.

D. J. Rossi and A. S. Willsky, \Reconstruction from projections based on detection and estimation of objects," IEEE Trans. Acoust. Speech, Signal Processing ASSP-32(4), pp. 886-906, 1984.

No context found.

D. Rossi and A. Wilsky, "Reconstruction from projections based on detection and estimation of objects," ieeeASSP, vol. ASSP-32, no. 4, pp. 886--906, 1984.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC