# US 11,432,787 B2 · Method and apparatus for image noise reduction > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,432,787 B2 - Document ID: US11432787B2 - Status: Granted - Kind: B2 - Application number: 17/480,243 - Filed: 2021-09-21 - Published: 2022-03-31 - Granted: 2022-09-06 - Earliest claimed: 2020-09-25 - Assignees: Siemens Healthcare GmbH - Applicants: Siemens Healthcare GmbH - Inventors: Magdalena Herbst; Steffen Kappler; Ludwig Ritschl - Web page: https://patentgrep.com/patent/US11432787B2 ## CPC - A61B 6/5258: {involving detection or reduction of artifacts or noise} - A61B 6/025: {Tomosynthesis} - A61B 6/032: {Transmission computed tomography [CT]} - A61B 6/487: {involving fluoroscopy} - A61B 6/502: {for diagnosis of breast, i.e. mammography} - G06T 5/10: using non-spatial domain filtering - G06T 5/75: Unsharp masking - G06T 7/40: Analysis of texture (depth or shape recovery from texture G06T7/529) - G06V 10/30: Noise filtering - G06T 2207/10081: Computed x-ray tomography [CT] - G06T 2207/20016: Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform - G06T 2207/20056: Discrete and fast Fourier transform, [DFT, FFT] - G06T 2207/20064: Wavelet transform [DWT] - G06T 2207/30004: Biomedical image processing - G06V 2201/03: Recognition of patterns in medical or anatomical images - A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D) - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract A method and a system are for image noise reduction. In an embodiment, the method includes producing a recorded image; establishing an amount of noise of the recorded image; decomposing the amount of noise into a number of N frequency-dependent noise components for N frequency bands, the number of N frequency-dependent noise components including respective data points respectively reproducing noise, of the amount of noise in the recorded image, for the respective frequency bands of the N frequency bands; examining the number of N frequency-dependent noise components for outlier data points, where an intensity lies outside a range of values, and forming moderated noise components by moderation of values of the outlier data points established in the examining of the number of N frequency-dependent noise components; and subtracting the moderated noise components from the recorded image. ## Claims ### Claim 1 A method for image noise reduction, comprising: producing a recorded image; establishing an amount of noise of the recorded image; decomposing the amount of noise into a number of N frequency-dependent noise components for N frequency bands, each of the N frequency-dependent noise components including data points, the data points reproducing the amount of noise of the recorded image for a respective N frequency band; examining the number of N frequency-dependent noise components for outlier data points, each of the outlier data points having an intensity outside of a range of values; forming moderated noise components by moderation of values of the outlier data points; and subtracting the moderated noise components from the recorded image. ### Claim 2 (depends on 1) The method of claim 1, wherein the recorded image has been produced via x-rays. ### Claim 3 (depends on 1) The method of claim 1, wherein the amount of noise of the recorded image is established, in the establishing, via an image noise reduction algorithm. ### Claim 4 (depends on 1) The method of claim 1, wherein the decomposing of the amount of noise is carried out based upon a Laplace pyramid, a Fast-Fourier transformation, or a wavelet transformation. ### Claim 5 (depends on 1) The method of claim 1, wherein the data points are image points of an image or reconstructable to image points of an image, each image point representing a respective noise component for the respective N frequency band. ### Claim 6 (depends on 1) The method of claim 1, wherein the examining of the number of N frequency-dependent noise components for outlier data points and the forming of the moderated noise components is carried out by a moderation function, the moderation function moderating the values of the outlier data points and leaving values of the data points within the range of values unchanged. ### Claim 7 (depends on 1) The method of claim 1, wherein, before the subtracting of the moderated noise components from the recorded image, the moderated noise components are combined to form one moderated amount of noise, the N frequency bands, taken together, producing a contiguous frequency range. ### Claim 8 (depends on 1) The method of claim 1, wherein producing the recorded image further includes at least one of measuring a noise behavior of equipment used for recording, wherein the moderating depends on the noise behavior measured, and calibrating a noise behavior of the equipment used for recording, wherein the calibrating takes place before a recording of an image. ### Claim 9 (depends on 1) The method of claim 1, wherein the amount of noise is decomposed during the decomposing into more than two noise components. ### Claim 10 An apparatus for image noise reduction, comprising: at least one of an imaging system designed to produce a recorded image and a data interface to receive a recorded image; a noise reduction unit to establish an amount of noise of the recorded image; a decomposition unit to decompose the amount of noise into a number of N frequency-dependent noise components for N frequency bands, each of the N frequency-dependent noise components including data points, the data points reproducing the amount of noise of the recorded image for a respective N frequency band; a moderation unit to examine the number of N frequency-dependent noise components for outlier data points each of the outlier data points having an intensity outside of a range of values, and to form moderated noise components by moderation of values of the outlier data points; and a subtraction unit to subtract the moderated noise components from the recorded image. ### Claim 11 (depends on 10) The apparatus of claim 10, further comprising: a combination unit, to combine the moderated noise components into one moderated amount of noise, the moderated noise components being combined before the moderated noise components are subtracted from the recorded image. ### Claim 12 (depends on 10) A control facility for controlling an imaging system, comprising: the apparatus of claim 10. ### Claim 13 (depends on 12) An imaging system, comprising the control facility of claim 12. ### Claim 14 (depends on 1) A non-transitory computer program product storing a computer program, directly loadable into a memory facility of a control facility, including program sections for carrying out the method of claim 1 when the computer program is executed in the control facility. ### Claim 15 (depends on 1) A non-transitory computer-readable medium, storing program sections readable in and executable by a processor, to carry out the method of claim 1 when the program sections are executed by the processor. ### Claim 16 (depends on 2) The method of claim 2, wherein the recorded image has been produced via x-rays within a framework of a recording method of radiography, fluoroscopy, mammography, tomography, or computed tomography. ### Claim 17 (depends on 2) The method of claim 2, wherein the recorded image is based on at least one of a mammography recording and a tomosynthesis recording. ### Claim 18 (depends on 2) The method of claim 2, wherein the recorded image is a synthetic 2D-image computed from a tomosynthesis recording or an image of an intermediate step of a computation. ### Claim 19 (depends on 3) The method of claim 3, wherein image noise reduction algorithm is based on at least one of BM3D, an iterative CT image filter for noise reduction, and artificial intelligence. ### Claim 20 (depends on 2) The method of claim 2, wherein the decomposing of the amount of noise is carried out based upon a Laplace pyramid, a Fast-Fourier transformation, or a wavelet transformation. ### Claim 21 (depends on 2) The method of claim 2, wherein the data points are image points of an image or reconstructable to image points of an image, each image point representing a respective noise component for the respective N frequency band. ### Claim 22 (depends on 3) The method of claim 3, wherein the recorded image is subdivided into the amount of noise and a component of image information. ### Claim 23 (depends on 6) The method of claim 6, wherein the values of the outlier data points are changed by the moderation function, the changed values being inside the range of values. ### Claim 24 (depends on 23) The method of claim 23, wherein the values of the outlier data points are set to limits of the range of values, or, as a distance of the values of the outlier data points from the range of values increases, the values of the outlier data points are changed to be closer to an average value of the range of values. ### Claim 25 (depends on 9) The method of claim 9, wherein the amount of noise is decomposed during the decomposing into three noise components. ### Claim 26 (depends on 9) The method of claim 9, wherein noise components with relatively highest frequency bands are chosen. ### Claim 27 (depends on 11) A control facility for controlling an imaging system, comprising: the apparatus of claim 11. ### Claim 28 (depends on 13) The imaging system of claim 13, wherein the imaging system is at least one of a radiography system, a fluoroscopy system, a mammography system, a tomography system or a computed tomography system. ### Claim 29 (depends on 1) The method of claim 1, wherein the range of values is defined by values within a predetermined number of standard deviations away from an average value of the noise for a respective N frequency-dependent noise component, the intensity of each outlier data point having a deviation greater than the predetermined number of standard deviations.