Method, arrangement, and computer program product for efficient production of tomographic images

1. A method for constructing a tomographic reconstruction through the application of statistical inversion, comprising:

representing unknowns that are associated with points of a reconstruction grid with components of a vector-valued variable,

repeatedly generating a next version of said vector-valued variable by drawing for each component a value from a conditional distribution, and

using the components of a so generated version of said vector-valued variable as said tomographic reconstruction;

wherein:

said components of the vector-valued variable are divided into sets, wherein components in a set represent unknowns associated with mutually independent points of said reconstruction grid, and

for at least two components of a set, said drawing of values from a conditional distribution is performed in parallel in a computer-implemented process.

2. A method for constructing a tomographic reconstruction through the application of statistical inversion, comprising:

representing unknowns that are associated with points of a reconstruction grid with components of a vector-valued variable,

repeatedly generating a next version of said vector-valued variable by drawing for each component a value from a conditional distribution, and

using the components of a so generated version of said vector-valued variable as said tomographic reconstruction;

wherein:

said components of the vector-valued variable are divided into collections, wherein components in a first collection represent unknowns associated with points of said reconstruction grid that are mutually independent with all points, the unknowns associated with which are represented by components in a second collection, and

for components of different collections, said drawing of values from a conditional distribution is performed in parallel in a computer-implemented process.

3. A method according to claim 1 , wherein:

said drawing of values from a conditional distribution is performed in at least one of the following:

parallel threads of a computer-implemented process, parallel cores of a multicore processor, parallel processors of a multiprocessor computing architecture, parallelly operating computers, and

the method comprises, in each of said parallel threads, parallel cores, parallel processors, or parallelly operating computers, allocating and initializing a random number generator independently of other threads, cores, processors, or computers for the drawing of a value from the corresponding conditional distribution.

4. A method according to claim 1 , wherein mutually independent points of said reconstruction grid are points for which:

there is no common neighbourhood in which a prior probability density was defined in constructing said tomographic reconstruction, and

there is no common measurement ray that would pass through them.

5. A method according to claim 1 , wherein:

tomographic reconstruction is performed from different exposures with multiple views of an examined object, which different exposures have been obtained by rotating a radiation source and a detector facing each other around said object about a vertically oriented rotational axis,

mutually independent points of said reconstruction grid are obtained by selecting, from at least one vertical column of said reconstruction grid, every n:th point, where n is an integer larger than or equal to two.

6. A method according to claim 1 , wherein:

tomographic reconstruction is performed from different exposures with multiple views of an examined object, which different exposures have been obtained by rotating a radiation source and a detector facing each other around said object about a vertically oriented rotational axis,

mutually independent collections are obtained by selecting such vertical columns of said reconstruction grid that pass through mutually independent points in a horizontal cross-section of the reconstruction grid.

7. An arrangement for constructing a tomographic reconstruction through the application of statistical inversion, comprising:

a data collection interface configured to obtain a set of measurement data representative of attenuation experienced by electromagnetic waves that propagated through parts of an object to be examined, and

a processor configured to represent unknowns that are associated with points of a reconstruction grid with components of a vector-valued variable, to repeatedly generate a next version of said vector-valued variable by drawing for each component a value from a conditional distribution, and to use the components of a so generated version of said vector-valued variable as said tomographic reconstruction;

wherein:

the processor is configured to divide components of the vector-valued variable into sets, wherein components in a set represent unknowns associated with mutually independent points of said reconstruction grid, and to perform, for at least two components of a set, said drawing of values from a conditional distribution in parallel in a computer-implemented process.

8. An arrangement for constructing a tomographic reconstruction through the application of statistical inversion, comprising:

a data collection interface configured to obtain a set of measurement data representative of attenuation experienced by electromagnetic waves that propagated through parts of an object to be examined, and

a processor configured to represent unknowns that are associated with points of a reconstruction grid with components of a vector-valued variable, to repeatedly generate a next version of said vector-valued variable by drawing for each component a value from a conditional distribution, and to use the components of a so generated version of said vector-valued variable as said tomographic reconstruction;

wherein:

the processor is configured to divide components of the vector-valued variable into collections, wherein components in a first collection represent unknowns associated with points of said reconstruction grid that are mutually independent with all points, the unknowns associated with which are represented by components in a second collection, and to perform, for at least two different collections, said drawing of values from a conditional distribution in parallel in a computer-implemented process.

9. An apparatus according to claim 7 , comprising an X-ray imaging section configured to controllably irradiate the object to be examined with X-rays and to detect intensities of X-rays that passed through parts of the object to be examined in order to produce said set of measurement data.

10. An apparatus according to claim 7 , comprising user interface means configured to present a graphical image of said tomographic reconstruction to a human user.

11. A non-transitory computer readable storage medium having stored thereon a computer program product for constructing a tomographic reconstruction through the application of statistical inversion, comprising machine-readable instructions that, when executed by a processor, are configured to cause the implementation of:

representing unknowns that are associated with points of a reconstruction grid with components of a vector-valued variable,

repeatedly generating a next version of said vector-valued variable by drawing for each component a value from a conditional distribution, and

using the components of a so generated version of said vector-valued variable as said tomographic reconstruction;

wherein:

said components of the vector-valued variable are divided into sets, wherein components in a set represent unknowns associated with mutually independent points of said reconstruction grid, and

for at least two components of a set, said drawing of values from a conditional distribution is performed in parallel in a computer-implemented process.

12. A non-transitory computer readable storage medium having stored thereon a computer program product for constructing a tomographic reconstruction through the application of statistical inversion, comprising machine-readable instructions that, when executed by a processor, are configured to cause the implementation of:

representing unknowns that are associated with points of a reconstruction grid with components of a vector-valued variable,

repeatedly generating a next version of said vector-valued variable by drawing for each component a value from a conditional distribution, and

using the components of a so generated version of said vector-valued variable as said tomographic reconstruction;

wherein:

said components of the vector-valued variable are divided into collections, wherein components in a first collection represent unknowns associated with points of said reconstruction grid that are mutually independent with all points, the unknowns associated with which are represented by components in a second collection, and

for components of different collections, said drawing of values from a conditional distribution is performed in parallel in a computer-implemented process.

13. A non-transitory computer readable storage medium having stored thereon a computer program product according to claim 11 , comprising machine-readable instructions that, when executed by a processor, are configured to cause the implementation of:

performing said drawing of values from a conditional distribution in at least one of the following: parallel threads of a computer-implemented process, parallel cores of a multicore processor, parallel processors of a multiprocessor computing architecture, parallelly operating computers, and

in each of said parallel threads, parallel cores, parallel processors, or parallel computers, allocating and initializing a random number generator independently of other threads, cores, processors, or computers for the drawing of a value from the corresponding conditional distribution.

14. A non-transitory computer readable storage medium having stored thereon a computer program product according to claim 12 , comprising machine-readable instructions that, when executed by a processor, are configured to cause the implementation of:

performing said drawing of values from a conditional distribution in at least one of the following: parallel threads of a computer-implemented process, parallel cores of a multicore processor, parallel processors of a multiprocessor computing architecture, parallelly operating computers, and

in each of said parallel threads, parallel cores, parallel processors, or parallel computers, allocating and initializing a random number generator independently of other threads, cores, processors, or computers for the drawing of a value from the corresponding conditional distribution.

15. A method according to claim 2 , wherein:

said drawing of values from a conditional distribution is performed in at least one of the following:

parallel threads of a computer-implemented process, parallel cores of a multicore processor, parallel processors of a multiprocessor computing architecture, parallelly operating computers, and

the method comprises, in each of said parallel threads, parallel cores, parallel processors, or parallelly operating computers, allocating and initializing a random number generator independently of other threads, cores, processors, or computers for the drawing of a value from the corresponding conditional distribution.

16. A method according to claim 2 , wherein mutually independent points of said reconstruction grid are points for which:

there is no common neighbourhood in which a prior probability density was defined in constructing said tomographic reconstruction, and

there is no common measurement ray that would pass through them.

17. A method according to claim 2 , wherein:

tomographic reconstruction is performed from different exposures with multiple views of an examined object, which different exposures have been obtained by rotating a radiation source and a detector facing each other around said object about a vertically oriented rotational axis,

mutually independent points of said reconstruction grid are obtained by selecting, from at least one vertical column of said reconstruction grid, every n:th point, where n is an integer larger than or equal to two.

18. A method according to claim 2 , wherein:

tomographic reconstruction is performed from different exposures with multiple views of an examined object, which different exposures have been obtained by rotating a radiation source and a detector facing each other around said object about a vertically oriented rotational axis,

mutually independent collections are obtained by selecting such vertical columns of said reconstruction grid that pass through mutually independent points in a horizontal cross-section of the reconstruction grid.

19. An apparatus according to claim 8 , comprising an X-ray imaging section configured to controllably irradiate the object to be examined with X-rays and to detect intensities of X-rays that passed through parts of the object to be examined in order to produce said set of measurement data.

20. An apparatus according to claim 8 , comprising user interface means configured to present a graphical image of said tomographic reconstruction to a human user.