Spectral X-ray imaging

Patent metadata

Field	Value
Publication number	US-10139354-B2
Application number	US-201415302397-A
Country	US
Kind code	B2
Filing date	Apr 7, 2014
Priority date	Apr 7, 2014
Publication date	Nov 27, 2018
Grant date	Nov 27, 2018

Abstract

There is provided a method for processing a radiographic image acquired with at least two energy levels. A first step (S 1 ) involves providing energy-resolved image data representative of the radiographic image with at least two energy levels, e.g. from a detector or from an intermediate storage. A second step (S 2 ) involves decomposing the provided image data into at least one basis image representation, based on a model where a combination of at least two basis functions is used to express a representation of at least one linear attenuation coefficient, and where at least one basis function models a physical material and at least one other basis function models the Non-Linear Partial Volume, NLPV, effect.

First claim

The invention claimed is:

1. An image processing method for medical imaging including processing a radiographic image acquired with at least two energy levels, said method comprising the steps of:
providing (S 1 ), to a processing circuitry and memory, energy-resolved image data, either from a detector or from an intermediate data storage, the energy-resolved image data representative of said radiographic image with at least two energy levels; and
generating, by way of the processing circuitry and memory, at least one image representation for medical imaging by decomposing (S 2 ) said provided image data into at least one basis image representation, based on a model where a combination of at least two basis functions is used to express a representation of at least one linear attenuation coefficient, where at least one ordinary basis function models a physical material, and an NLPV basis function models the Non-Linear Partial Volume (NLPV) effect,
wherein said at least one other basis function that models the NLPV effect is constructed by:
selecting (S 11 ) a set of at least one ordinary basis function, such that the energy dependent linear attenuation coefficients of homogeneous materials in the image volume can be described by a linear combination of the basis functions in the set,
determining (S 12 ) the fraction of transmitted photons as a function of energy for a detector element situated behind an inhomogeneous region of materials, as seen from the source, taking into account the spatially extended nature of either the detector, the source or both, and
determining (S 13 ) at least one NLPV basis function as the linear attenuation coefficient of a fictious material which, if one unit of distance of a homogeneous slab of this material were traversed by an x-ray beam, would give the same transmitted photon fraction as a function of energy as said calculated fraction which is transmitted through an inhomogeneous region.

2. A computer program product comprising a non-transitory computer-readable storage medium having recorded thereon a computer program, wherein the computer program comprises instructions which, when executed by at least one processor, cause the processor(s) to carry out the steps of claim 1 .

3. A system ( 120 ) for medical imaging based on processing a radiographic image acquired with at least two energy levels, said system comprises:
processing circuitry and memory, the memory comprising instruction code executable by the processing circuitry, whereby upon execution of the instruction code, the processing circuitry is operative to:
obtain energy-resolved image data representative of said radiographic image with at least two energy levels by receiving image data of an energy-resolved sinogram as input, and
generate at least one image representation for medical imaging by performing basis decomposition of said image data into at least one basis image representation, based on a model where a combination of at least two basis functions is used to express a representation of at least one linear attenuation coefficient, where at least one basis function, also referred to as an ordinary basis function, models a physical material and at least one other basis function, also referred to as an NLPV basis function, models the Non-Linear Partial Volume, NLPV, effect to produce at least one basis image corresponding to an NLPV basis function, also referred to as an NLPV basis image, or at least one subsequently reconstructed tomographic NLPV basis image, enabling the identification of interface.

4. An image processing method for medical imaging including processing a radiographic image acquired with at least two energy levels, said method comprising the steps of:
providing (S 1 ), to a processing circuitry and memory, energy-resolved image data, either from a detector or from an intermediate data storage, the energy-resolved image data representative of said radiographic image with at least two energy levels; and
generating, by way of the processing circuitry and memory, at least one image representation for medical imaging by decomposing (S 2 ) said provided image data into at least one basis image representation, based on a model where a combination of at least two basis functions is used to express a representation of at least one linear attenuation coefficient, where at least one ordinary basis function models a physical material, and an NLPV basis function models the Non-Linear Partial Volume (NLPV) effect,
wherein the provided image data is an energy-resolved sonogram, and the output is either of an NLPV basis image corresponding to an NLPV basis function or a subsequently reconstructed tomographic NLPV basis image, said output enabling the identification of interfaces.

5. A computer program product comprising a non-transitory computer-readable storage medium having recorded thereon a computer program, wherein the computer program comprises instructions which, when executed by at least one processor, cause the processor(s) to carry out the steps of claim 4 .

6. An image processing method for medical imaging including processing a radiographic image acquired with at least two energy levels, said method comprising the steps of:
providing (S 1 ), to a processing circuitry and memory, energy-resolved image data, either from a detector or from an intermediate data storage, the energy-resolved image data representative of said radiographic image with at least two energy levels; and
generating, by way of the processing circuitry and memory, at least one image representation for medical imaging by decomposing (S 2 ) said provided image data into at least one basis image representation, based on a model where a combination of at least two basis functions is used to express a representation of at least one linear attenuation coefficient, where at least one ordinary basis function models a physical material, and an NLPV basis function models the Non-Linear Partial Volume (NLPV) effect,
wherein the provided image data is an energy-resolved sonogram, and the output is at least one ordinary basis image corresponding to an ordinary basis function, which is subsequently reconstructed to yield at least one tomographic ordinary basis image, and
wherein said at least one tomographic ordinary basis image has a reduced level of NLPV artifacts compared to a tomographic basis image reconstructed from a basis image resulting from a basis material decomposition without NLPV basis.

7. A computer program product comprising a non-transitory computer-readable storage medium having recorded thereon a computer program, wherein the computer program comprises instructions which, when executed by at least one processor, cause the processor(s) to carry out the steps of claim 6 .

8. An image processing method for medical imaging including processing a radiographic image acquired with at least two energy levels, said method comprising the steps of:
providing (S 1 ), to a processing circuitry and memory, energy-resolved image data, either from a detector or from an intermediate data storage, the energy-resolved image data representative of said radiographic image with at least two energy levels; and
generating, by way of the processing circuitry and memory, at least one image representation for medical imaging by decomposing (S 2 ) said provided image data into at least one basis image representation, based on a model where a combination of at least two basis functions is used to express a representation of at least one linear attenuation coefficient, where at least one ordinary basis function models a physical material, and an NLPV basis function models the Non-Linear Partial Volume (NLPV) effect,
wherein the output basis image representation(s) include at least one sinogram of

Assignees

• Prismatic Sensors Ab
• Prismatic Sensors

Inventors

• Persson Mats

Classifications

• G06T2211/408

  Dual energy · CPC title

• G01N23/087

  using polyenergetic X-rays · CPC title

• A61B6/52

  Devices using data or image processing specially adapted for radiation diagnosis · CPC title

• G06T7/0012

  Biomedical image inspection · CPC title

• A61B6/5205

  involving processing of raw data to produce diagnostic data · CPC title