Image analysis – Applications – Biomedical applications
Reexamination Certificate
2000-08-14
2004-04-20
Boudreau, Leo (Department: 2621)
Image analysis
Applications
Biomedical applications
C382S190000
Reexamination Certificate
active
06724924
ABSTRACT:
BACKGROUND
1. Technical Field
This disclosure relates to radiographic imaging and detection, and more particularly, to a method for automatically detecting and measuring pedicles in spine images.
2. Description of the Related Art
In the radiographic diagnosis of spines, pedicles are an important anatomy for measuring rotational deformity. Pedicles may vary significantly in position, size, and shape, and may also appear in images with different brightness and contrast. These variations are rich in diagnostic information. Computer detection, measurement, and characterization of pedicles and their variations provide efficient and accurate ways in clinical diagnostic practice, yet pose substantial difficulties and challenges to the development of automatic computerized methods.
Currently, detection, measurement, and characterization of pedicles is performed manually by physicians, and is therefore subject to human errors and usually is not reproducible.
Therefore, a need exists to provide a method for automatic pedicle detection and measurement. A further need exists for a method, which is immune from changes in image brightness and image contrast.
SUMMARY OF THE INVENTION
A method for detecting pedicle positions in an image, preferably a radiographic image, in accordance with the present invention, includes providing a set of feature prototypes for a plurality of pedicle positions and orientations, providing an input image to be analyzed for pedicle positions and orientations, and determining intensity curvatures for a pedicle in the input image. The intensity curvatures are transformed to determine a feature vector for the pedicle in the input image. The feature vector is correlated to the feature prototypes to determine most likely positions and orientations of the pedicle.
Another method for detecting pedicle positions in an image of a spine, includes the step of training a set of feature prototypes for a plurality of pedicle positions and orientations. The training step is performed by selecting images of pedicles in different positions and orientations, determining intensity curvatures of the images, transforming the intensity curvatures to determine features for the pedicles for each training sample, and performing an eigenvalue decomposition to provide a set of feature prototypes. An input image is provided to be analyzed for pedicle positions and orientations, and pedicle locations and orientations are detected. The pedicle locations and orientations are detected by determining intensity curvatures for the input image of the pedicle in the input image, transforming the intensity curvatures to determine a feature vector for the pedicle in the input image and correlating the feature vector to the feature prototypes to determine most likely positions and orientations of the pedicle.
In other methods, the bone preferably includes a pedicle, but other specialized bones may be determined in accordance with the invention. The step of providing a set of feature prototypes may include the steps of selecting images of the pedicle in different positions and orientations, determining intensity curvatures of the images, transforming the intensity curvatures to determine features for the pedicle for each orientation and position and performing an eigenvalue decomposition to provide the set of feature prototypes. The step of transforming the intensity curvatures to determine features for the pedicle for each orientation and position may include employing one of a discrete cosine transformation and a wavelet transformation. The step of transforming the intensity curvatures to determine a feature vector for the pedicle in the input image may include employing one of a discrete cosine transformation and a wavelet transformation.
In still other methods, the step of correlating the feature vector to the feature prototypes to determine most likely positions and orientations of the pedicle may include the steps of projecting the feature vector to the feature prototypes to create a reconstructed feature set, and comparing the reconstructed feature set to the feature vector to determine the most likely positions and orientations of the pedicle. The method may include the step of repeating the comparing step for a plurality of different image resolutions. The method may further include the step of marking position points on the input image. The above methods may be implemented by a program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform these method steps.
These and other objects, features and advantages of the present invention will become apparent from the following detailed description of illustrative embodiments thereof, which is to be read in connection with the accompanying drawings.
REFERENCES:
patent: 4881270 (1989-11-01), Knecht et al.
patent: 5943435 (1999-08-01), Gaborski
patent: 6249590 (2001-06-01), Young et al.
Verdonck et al. “Computer Assisted Quantitive Analysis of Deformities of the Human Spine”, Medical Image Computing and Computer-Assisted Intervention, Springer-Verlag, Proceedings of MICCAI '98, Cambridge, MA, Oct. 1998, pp. 822-831.*
Tascini et al., “Automatic Quantitative Analysis of Lumbar Bone Radiographs”, Nuclear Science Symposium and Medical Imaging Conference, 1993 IEEE Conference Record, vol. 3 Oct. 31-Nov. 6, 1993, pp. 1722-1726.*
Kauffmann et al., “Digital Radiography Segmentation of Scoliotic Vertebral Body Using Deformable Models”, Proc. SPIE vol. 3034: Medical Imaging 1997: Image Processing, Apr. 1997, pp. 243-251.
Qian Jianzhong
Schramm Helmuth
Wei Guo-Qing
Boudreau Leo
F. Chau & Associates LLP
Miller Ryan J.
Paschburg Donald B.
Siemens Corporate Research Inc.
LandOfFree
Brightness and contrast invariant detection of vertebra... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Brightness and contrast invariant detection of vertebra..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Brightness and contrast invariant detection of vertebra... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3233940