Image analysis – Applications – Biomedical applications
Reexamination Certificate
1999-10-15
2001-07-17
Mancuso, Joseph (Department: 2621)
Image analysis
Applications
Biomedical applications
C378S065000
Reexamination Certificate
active
06263096
ABSTRACT:
FIELD OF THE INVENTION
The present invention generally concerns imaging. More specifically, the present invention concerns a method of high speed reprojection of tomographic images.
BACKGROUND OF THE INVENTION
Diagnostic imaging is an important tool for observing everything from subcellular structures to the environment, and tomographic reconstruction is an imaging technique that underlies nearly all key diagnostic imaging modalities. Such tomographic modes include X-ray Computed Tomography (CT), Positron Emission Tomography (PET), and Single Photon Emission Count Tomography (SPECT). In addition, certain acquisition methods for Magnetic Resonance Imaging (MRI) use tomography, and new techniques are emerging such as electrical impedance tomography (EIT) and optical tomography. Tomographic reconstruction is also fundamental to numerous other applications in science and engineering, including electron microscopy to determine subcellular structure, nondestructive evaluation (NDE) in manufacturing, geophysical exploration, environmental monitoring, and remote sensing.
An important component in the various tomographic reconstruction procedures is the operation of computing a set of projections of a given image, called forward projection or reprojection. Reprojection is a process by which projections are produced from an image, such that, if the projections are filtered and back projected, they yield the original image. While reprojection is widely used for various purposes, known algorithms that employ the reprojection processes incur a high computational expense.
Reprojection is of interest in several applications. Reprojection is used in X-ray CT, see, C. R. Crawford, J. G. Colsher, N. J. Pelc, and A. H. R. Lonn, “High Speed Reprojection and its Application,”
Proc. SPIE—Int. Soc. Opt. Eng. Conf. Medical Imaging II,
Newport Beach, Calif., vol 914, pt A, pp. 311-18, 1988; in iterative beam-hardening correction algorithms, see, P. M. Joseph and R. D. Spital, “A Method for Correcting Bone Induced Artifacts in Computed Tomography Scanners,”
JCAT
vol 2 No. 1, pp. 100-108, 1978, and U.S. Pat. Nos. 4,217,641 to Naparstek and 5,438,602 to Crawford, et al.); in streak suppression algorithms, see, G. Henrich, “A Simple Computational Method for Reducing Streak Artifacts in CT Images,”
Computerized Tomography,
vol. 4, pp. 67-71, 1980 and U.S. Pat. No. 5,229,934 to Mattson et al.; in algorithms for the removal of artifacts caused by the presence of metallic implants in a patient, see, G. Glover and N. J. Pelc, “An Algorithm for the Reduction of Metal Clip Artifacts in CT Reconstructions.”
Medical Physics,
vol. 8 No. 6, pp. 799-807, 1981 and U.S. Pat. No. 5,243,664 to Tuy; or other high density objects, see, U.S. Pat. No. 4,709,333 to Crawford; and in correcting for missing data, see, U.S. Pat. No. 5,396,528 to Hu et al., and partial volume effects, see, U.S. Pat No. 5,727,041 to Hsieh.
In PET and SPECT imaging, reprojection has been used to compensate for attenuation, see, J. M. Ollinger, “Reconstruction-Reprojection Processing of Transmission Scans and the Variance of PET Images,”
IEEE Trans. Nuc. Sci.
Vol. 39 No. 4, p 1122 August 1992. Reprojection is also used in detection and compensation for various acquisition errors, see, S. C. Huang and D. C. Yu, “Capability Evaluation of a Sinogram Error Detection and Correction Method in Computed Tomography,”
IEEE Trans. Nuc. Sci.
Vol. 39 No. 4, pp. 1106-1110 August 1992, and E. C. Frey, Z. W. Ju, and B. M. W. Tsui, “A Fast Projector-Backprojector Pair Modeling the Asymmetric, Spatially Varying Scatter Response Function for Scatter Compensation in SPECT Imaging,”
IEEE Trans. Nic. Sci,
vol. 40, No. 4, pp. 1192-7, August 1993, including acquisition errors caused by patient motion, see, U.S. Pat. Nos. 4,858,128 to Nowak, 5,552,605 to Arata, 5,579,358 to Lin, and 5,848,114 to Kawai et al.; and in accounting for Poisson noise statistics, see, J. M. Ollinger, “Iterative Reconstruction-Reprojection and the Expectation-Maximization Algorithm,”
IEEE Trans. Med. Imag.
Vol. 9 No. 1, p. 94 March 1990. Reprojection is also a critical component in iterative tomographic reconstruction algorithms, which are the preferred method in imaging modalities such as PET, SPECT, and nondestructive testing. Iterative tomographic reconstruction algorithms are also critical in X-ray CT algorithms for handling a variety of the aforementioned artifacts caused by the presence of metal clips in CT images.
There is a need for fast reconstruction of such tomographic images, especially with the advent of new technologies that are capable of collecting large quantities of data in real time, for example, multi-line spiral CT, cardiac imaging using X-ray CT, and an upcoming CT fluoroscopy. In addition, there exists an increasing demand for real-time interventional imaging, e.g., to monitor and guide surgery. Very often with known techniques the reconstruction of images from such data becomes a bottleneck. While it has been a subject of continuing efforts in industry and academia since the introduction of CT, known efforts to speed up tomographic reconstruction have been unsuccessful. In addition, the development of methods to speed up tomographic reconstruction has been recently identified as a major thrust area by groups such as the National Institute of Health (NIH).
Current iterative methods are far more expensive than the conventional direct filtered back projection (FBP) reconstruction, and typically require orders of magnitude more computation. This has motivated intensive work on the acceleration of iterative methods dating back to the first CT scanner around 1971, which used iterative reconstruction. More recently, the need to develop fast iterative reconstruction methods has been identified as a pressing need, see, Imaging Sciences Working Group, “Matching Clinical and Biological Needs with Emerging Imaging Technologies,” tech. rep., Diagnostic Imaging Program, National Cancer Institute, 1998. Because of the key role played by the process of reprojection in the various recontruction algorithms, the importance of accelerating it has been recognized as early as 1978.
Prior methods for reprojection are problematic. Direct methods work directly on the data, and involve direct computation of weighted sums, and interpolation. Direct A) methods can be made as accurate as needed at the cost of increased computation, and serve as a benchmark in terms of accuracy. Typical examples of direct methods are described in P. M. Joseph, “An Improved Algorithm for Reprojecting Rays Through Pixel Images,”
IEEE Trans. Med. Imag.
vol. 1, No. 3, pp. 192-198, 1982, D. C. Yu and S. C. Huang, “Study of Reprojection Methods in Terms of Their Resolution Loss and Sampling Errors,”
IEEE Trans. Nuc. Sci.,
vol. 40, No. 4, p 1174, August 1993, and G. L. Zeng, Y. L. Hsieh, and G. T. Gullberg, “A Rotating and Warping Projector/Backprojector for Fan-beam and Cone-beam Iterative Algorithm,”
IEEE Trans. Nuc. Sci.,
vol. 41, No. 6, pp. 2807-11, December 1994. However, in spite of various direct methods improvements, e.g., U.S. Pat. Nos. 5,559,335 to Zeng et al. and 5,625,190 to Crandall, direct methods require a high computational cost, proportional to N
3
(denoted O(N
3
)), to generate N projections for an image with N×N pixels. For example, using one of these methods to compute the reprojection of a typical 4096×4096-pixel image requires 16
3
=4096 times the computation needed for a 256×256 pixel image. Thus, various efforts have increased in academia and industry in an attempt to develop fast reprojection methods.
Another typical approach to accelerating the reprojection process in commercial products is to develop special purpose hardware that attempts to increase the rate at which the reprojection is performed. Such methods are described in E. B. Hinkle, J. L. C. Sanz, A. K. Jain, and D. Petkovic, “P/sup 3/E: New Life for Projection-based Image Processing,”
J. Parallel
&
Distrib. Comput.,
Vol. 4, No. 1, pp. 45-78, February 1987, T. M. Peters, “Algorithms for Fast
Boag Amir
Bresler Yoram
Michielssen Eric
Greer Burns & Crain Ltd
Mancuso Joseph
Tabatabai Abolfazl
The Board of Trustees of the University of Illinois
LandOfFree
Multilevel domain decomposition method for fast reprojection... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Multilevel domain decomposition method for fast reprojection..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Multilevel domain decomposition method for fast reprojection... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2447196