Medical image data processing and feature identification system
An anatomical feature boundary identification system for use in processing medical images including X-ray images having a substantial noise content, employs at least one repository. The at least one repository stores data representing multiple different candidate template boundary shapes for individual particular anatomical features of multiple different types of anatomical features. A computation processor coupled to the at least one repository, determines a converged boundary shape of a particular anatomical feature by iteratively substantially minimizing a first difference between, data representing a weighted combination of multiple different candidate template boundary shapes of a particular anatomical feature and data representing a boundary shape of the particular anatomical feature derived from image data of the particular anatomical feature. The computation processor iteratively substantially maximizes a second difference between, data representing the weighted combination and data representing background non-anatomical features in an image. An output processor coupled to the computation processor, provides data representing the converged boundary shape of the particular anatomical feature for presentation in a display image of the particular anatomical feature.
1. An anatomical feature boundary identification system for use in processing medical images including X-ray images having a substantial noise content, comprising:
at least one repository for storing data representing a plurality of different candidate template boundary shapes for individual particular anatomical features of a plurality of different types of anatomical features;
a computation processor coupled to said at least one repository, for,
determining a converged boundary shape of a particular anatomical feature by iteratively substantially minimizing a first difference between,
data representing a weighted combination of a plurality of different candidate template boundary shapes of a particular anatomical feature and
data representing a boundary shape of said particular anatomical feature derived from image data of said particular anatomical feature and
iteratively substantially maximizing a second difference between,
data representing the weighted combination and
data representing background non-anatomical features in an image; and
an output processor coupled to said computation processor, for providing data representing said converged boundary shape of said particular anatomical feature for presentation in a display image of said particular anatomical feature.
2. A system according to claim 1 , wherein
said computation processor determines said converged boundary shape by iteratively applying a probability density function in substantially minimizing the first difference and substantially maximizing the second difference.
3. A system according to claim 2 , wherein
said computation processor determines said converged boundary shape by iteratively applying a probability density function of the form,
p(z t |x t )p(z′ t |x′ t )p(x′ t |x t )
where p(z t |x t ) and p(z′ t |x′ t ) are object observation likelihood and background observation likelihood, respectively, which can be estimated by the image observation data, p(x′ t |x t ) is an “object-background relation density”.
4. A system according to claim 1 , wherein
said plurality of different types of anatomical features comprise different anatomical organs.
5. A system according to claim 4 , wherein
said different anatomical organs include at least two of, (a) a heart, (b) a lung, (c) a liver, (d) a kidney.
6. A system according to claim 1 , wherein
said plurality of different types of anatomical features comprise tumors.
7. A system according to claim 1 , wherein
said computation processor iteratively minimizes said difference by applying a weighted combination of vectors individually representing individual shapes of said plurality of different candidate template boundary shapes to data representing an average shape derived by taking an average of a plurality of different candidate template boundary shapes.
8. A system according to claim 7 , wherein
said average of said plurality of different candidate template boundary shapes is an arithmetic mean of said plurality of different candidate template boundary shapes.
9. A system according to claim 1 , wherein
said plurality of different candidate template boundary shapes of said particular anatomical feature are a plurality of different averaged candidate template boundary shapes of said particular anatomical feature derived by taking an average of subsets of said plurality of different candidate template boundary shapes of said particular anatomical feature.
10. An anatomical feature boundary identification system for use in processing medical images including X-ray images having a substantial noise content, comprising:
at least one repository for storing data representing a plurality of different candidate template boundary shapes for individual particular anatomical features of a plurality of different types of anatomical features;
a computation processor coupled to said at least one repository, for determining a converged boundary shape of a particular anatomical feature by iteratively substantially minimizing difference between,
data representing a weighted combination of a plurality of different candidate template boundary shapes of a particular anatomical feature and
data representing a boundary shape of said particular anatomical feature derived from image data of said particular anatomical feature, by applying a weighted combination of vectors individually representing individual shapes of said plurality of different candidate template boundary shapes to data representing an initial shape; and
an output processor coupled to said computation processor, for providing data representing said converged boundary shape of said particular anatomical feature for presentation in a display image of said particular anatomical feature.
11. A system according to claim 10 , wherein
said data representing said initial shape is derived by taking an average of a plurality of different candidate template boundary shapes.
12. A system according to claim 10 , wherein
said computation processor determines said converged boundary shape by iteratively applying a probability density function in minimizing said difference.
13. A system according to claim 10 , including
an adaptation processor for adaptively determining said different candidate template boundary shapes in response to at least one of, (a) predetermined user preference information and (b) frequency of usage of individual template boundary shapes of said different candidate template boundary shapes.
14. A system according to claim 13 , wherein
said adaptation processor adaptively determines said different candidate template boundary shapes in response to substantially immediate feedback updating said frequency of usage of said individual template boundary shapes.
15. An anatomical feature boundary identification system for use in processing medical images including X-ray images having a substantial noise content, comprising:
at least one repository for storing data representing a plurality of different candidate template boundary shapes for individual particular anatomical features of a plurality of different types of anatomical features;
a computation processor coupled to said at least one repository, for determining a converged boundary shape of a particular anatomical feature by iteratively substantially minimizing difference between,
data representing a weighted combination of a plurality of different candidate template boundary shapes of a particular anatomical feature and
data representing a boundary shape of said particular anatomical feature derived from image data of said particular anatomical feature, by iteratively applying a probability density function in minimizing said difference; and
an output processor coupled to said computation processor, for providing data representing said converged boundary shape of said particular anatomical feature for presentation in a display image of said particular anatomical feature.