# US 2010/0166277 A1 · Method for noise reduction of CT images and image processing system

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2010/0166277 A1
- Document ID: US20100166277A1
- Status: Published application
- Kind: A1
- Application number: 12/591,871
- Filed: 2009-12-03
- Published: 2010-07-01
- Earliest claimed: 2008-12-30
- Assignees: Siemens Aktiengesellschaft
- Inventors: Rainer Raupach
- Web page: https://patentgrep.com/patent/US20100166277A1

## CPC

- A61B 6/482: {involving multiple energy imaging}
- A61B 6/032: {Transmission computed tomography [CT]}
- G06T 5/10: using non-spatial domain filtering
- G06T 5/20: using local operators
- G06T 5/50: using two or more images, e.g. averaging or subtraction
- G06T 5/70: Denoising; Smoothing
- G06T 2207/10081: Computed x-ray tomography [CT]
- G06T 2207/20221: Image fusion; Image merging
- G06T 2207/30004: Biomedical image processing
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

A method is disclosed for noise reduction of CT image data and an image processing system is disclosed. An object under examination is scanned and at least two CT image datasets are created, each being undertaken on the basis of a different x-ray generation process. Subsequently, the image datasets are split up into at least two split-up image datasets, with a lowest local frequency band and at least one high local frequency band. In at least one embodiment, this is followed by the determination of the noise in at least one of the image datasets for each x-ray spectrum and calculation of at least one new image dataset using an unchanged split-up image dataset in each case with the lowest frequency band and an image dataset created from a noise-minimized weighted combination of split-up image datasets, which originate from the scans with different x-ray energy spectrums.

## Claims

### Claim 1

A method for noise reduction of CT image data,
comprising:
scanning an object under examination;
creating at least two CT image datasets, each dataset created on the basis of a different x-ray energy spectrum;
splitting the at least two CT image datasets into at least two split-up image datasets respectively, with a lowest local frequency band and at least one high local frequency band;
determining noise in the at least one of at least two CT image datasets for each x-ray energy spectrum;
calculating at least one new image dataset using an unchanged split-up one of the two CT image datasets with the lowest local frequency band and one of the at least two CT image datasets created from a noise-minimized weighted combination of split-up image datasets which originate from the scans with different x-ray energy spectrums.

### Claim 2 (depends on 1)

The method as claimed in claim 1, wherein the scanning is executed with different x-ray energy spectrums.

### Claim 3 (depends on 1)

The method as claimed in claim 1, wherein at least one energy-selective detector is used and for image creation, only one part of the measured energy spectrum is evaluated.

### Claim 4 (depends on 1)

The method as claimed in claim 1, wherein the splitting up of the image datasets, is executed by filtering with different local frequency filters.

### Claim 5 (depends on 4)

The method as claimed in claim 4, wherein the sum of the local frequency filters used is normalized to 1.

### Claim 6 (depends on 1)

The method as claimed in claim 1, wherein the splitting up of the image datasets, is executed by a wavelet transformation and the local frequency bands are determined by the level of the wavelet transformation.

### Claim 7 (depends on 1)

The method as claimed in claim 1, wherein the splitting up of the image datasets is executed by Fourier transformation and the local frequency bands are determined by the Fourier coefficients assigned to a local frequency.

### Claim 8 (depends on 1)

The method as claimed in claim 1, wherein the noise is determined in each respective raw image dataset.

### Claim 9 (depends on 1)

The method as claimed in claim 1, wherein the noise is determined, in each case, by at least one split-up image dataset.

### Claim 10 (depends on 8)

The method as claimed in claim 8, wherein the noise is determined, in each case, over the entire image dataset.

### Claim 11 (depends on 8)

The method as claimed in claim 8, wherein the noise is determined area-by area in the image dataset and weights are determined by area for the combination.

### Claim 12 (depends on 8)

The method as claimed in claim 8, wherein the noise is calculated pixel-by-pixel from an adjacent image area in the CT image dataset and the weights are calculated pixel-by-pixel.

### Claim 13 (depends on 1)

The method as claimed in claim 1, wherein the splitting up of the image datasets into precisely one high and precisely one low frequency band takes place.

### Claim 14 (depends on 1)

The method as claimed in claim 1, wherein the at least two CT image datasets include IA and IB and the at least two split-up image datasets include (IA(k), IB(k)), the new image datasets (I′A, I′B) being calculated in accordance with the following formulae:
I
A
′
=
I
A
(
1
)
+
∑
k
=
2
N

(
w
A
,
A
(
k
)

I
A
(
k
)
+
w
A
,
B
(
k
)

I
B
(
k
)
)
and
I
B
′
=
I
B
(
1
)
+
∑
k
=
2
N

(
w
B
,
A
(
k
)

I
A
(
k
)
+
w
B
,
B
(
k
)

I
B
(
k
)
)
,
with wA,A(k),wA,B(k),wB,A(k),wB,B(k) representing the mixture weights.

### Claim 15 (depends on 14)

The method as claimed in the previous claim 14, wherein the mixture weights wA,A(k),wA,B(k),wB,A(k),wB,B(k) are determined in accordance with the following formulae
w
A
,
A
(
k
)
=
w
B
,
A
(
k
)
=
(
σ
B
(
k
)
)
2
(
(
σ
A
(
k
)
)
2
+
(
σ
B
(
k
)
)
2
)
and
w
A
,
B
(
k
)
=
w
B
,
B
(
k
)
=
(
σ
A
(
k
)
)
2
(
(
σ
A
(
k
)
)
2
+
(
σ
B
(
k
)
)
2
)
with the noise being designated σA and σB, the indices k specifying the frequency band and wX,Y(k) representing the weight of the contribution of an image IY(k) to an image I′X.

### Claim 16 (depends on 1)

An image processing system comprising:
a computer; and
a memory for storing program code, the program code being provided in the memory to execute, during operation, the method claimed in claim 1.

### Claim 17 (depends on 2)

The method as claimed in claim 2, wherein at least one energy-selective detector is used and for image creation, only one part of the measured energy spectrum is evaluated.

### Claim 18 (depends on 9)

The method as claimed in claim 9, wherein the noise is determined, in each case, over the entire image dataset.

### Claim 19 (depends on 9)

The method as claimed in claim 9, wherein the noise is determined area-by area in the image dataset and weights are determined by area for the combination.

### Claim 20 (depends on 1)

A computer readable medium including program segments for, when executed on a computer device, causing the computer device to implement the method of claim 1.