# US 11,998,376 B2 · Method for generating training data for training an algorithm for estimating positioning variances in an imaging apparatus of an x-ray imaging system

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

- Number: US 11,998,376 B2
- Document ID: US11998376B2
- Status: Granted
- Kind: B2
- Application number: 18/212,864
- Filed: 2023-06-22
- Published: 2024-01-04
- Granted: 2024-06-04
- Earliest claimed: 2022-06-29
- Assignees: Siemens Healthineers AG
- Applicants: Siemens Healthcare GmbH
- Inventors: Josef Zeidler; Alexander Krämer; Gerhard Wurzer; Berthold Baumann
- Web page: https://patentgrep.com/patent/US11998376B2

## CPC

- A61B 6/4458: {the source unit or the detector unit being attached to robotic arms}
- A61B 6/4482: {involving power assist circuits}
- A61B 6/4441: {the rigid structure being a C-arm or U-arm}
- A61B 6/547: {involving tracking of position of the device or parts of the device}
- A61B 6/584: {determining position of components of the apparatus or device using images of the phantom}
- A61B 34/10: Computer-aided planning, simulation or modelling of surgical operations
- A61B 34/30: Surgical robots
- A61B 90/50: Supports for surgical instruments, e.g. articulated arms
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 3/09: Supervised learning
- G06T 7/0012: {Biomedical image inspection}
- G06T 12/20: Inverse problem, i.e. transformations from projection space into object space
- G16H 30/20: for handling medical images, e.g. DICOM, HL7 or PACS
- G16H 30/40: for processing medical images, e.g. editing
- G16H 40/63: for local operation
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G16H 50/50: for simulation or modelling of medical disorders
- A61B 6/405: {Source units specially adapted to modify characteristics of the beam during the data acquisition process (A61B6/4021, A61B6/4035 take precedence)}
- A61B 6/4464: {the source unit or the detector unit being mounted to ceiling}
- A61B 2090/3762: {using computed tomography systems [CT]}
- G06T 2207/20081: Training; Learning
- G06T 2211/441: AI-based methods, deep learning or artificial neural networks
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA

## Abstract

A computer-implemented method for generating training data for training an algorithm for estimating positioning variances in an imaging apparatus of an x-ray imaging system is provided. The method includes predetermining target positioning data for the imaging apparatus and calculating faulty positioning data, wherein the faulty positioning data deviates from the target positioning data by at least one predetermined positioning variance. An imaging method is simulated, in order to generate two or more simulated x-ray recordings, wherein a positioning of the imaging apparatus is simulated in accordance with the faulty positioning data. An annotated training data record is generated and stored, wherein the training data record contains two or more simulated x-ray recordings and at least one positioning variance as an annotation of the two or more simulated x-ray recordings.

## Claims

### Claim 1

A computer-implemented method for generating training data for training an algorithm for estimating positioning variances in an imaging apparatus of an x-ray imaging system, the computer-implemented method comprising:
predetermining target positioning data for the imaging apparatus;
calculating faulty positioning data as a function of the target positioning data, wherein the faulty positioning data deviates from the target positioning data by at least one predetermined positioning variance;
simulating an imaging method, wherein two or more simulated x-ray recordings are generated, wherein a positioning of the imaging apparatus is simulated in accordance with the faulty positioning data; and
generating and storing an annotated training data record that comprises the two or more simulated x-ray recordings and at least one positioning variance as an annotation of the two or more simulated x-ray recordings.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, wherein the target positioning data establishes a target trajectory for the imaging apparatus and the faulty positioning data establishes a faulty trajectory for the imaging apparatus.

### Claim 3 (depends on 2)

The computer-implemented method of claim 2, wherein the target trajectory for a target positioning of the imaging apparatus establishes a target time instant, and
wherein the faulty trajectory for the target positioning of the imaging apparatus establishes a faulty time instant that differs from the target time instant according to the at least one positioning variance.

### Claim 4 (depends on 2)

The computer-implemented method of claim 2, wherein a first target positioning for a first simulated x-ray recording differs from a second target positioning for a second simulated x-ray recording in accordance with the target positioning data according to the target relative pose,
wherein a first faulty positioning for the first simulated x-ray recording deviates from a second faulty positioning for the second simulated x-ray recording in accordance with the faulty positioning data in accordance with a faulty relative pose, and
wherein the faulty relative pose deviates from the target relative pose.

### Claim 5 (depends on 4)

The computer-implemented method of claim 4, wherein a third target positioning for a third simulated x-ray recording corresponds with a fourth target positioning for a fourth simulated x-ray recording in accordance with the target positioning data, and
wherein a third faulty positioning for the third simulated x-ray recording deviates from a fourth faulty positioning for the fourth simulated x-ray recording in accordance with the faulty positioning data.

### Claim 6 (depends on 1)

The computer-implemented method of claim 1, wherein the at least one positioning variance is determined in accordance with a predetermined model for the positioning of the imaging apparatus.

### Claim 7 (depends on 6)

The computer-implemented method of claim 6, wherein the positioning of the imaging apparatus is established by respective axis positions of kinematic axes of the x-ray imaging system,
wherein the predetermined model comprises a statistical position distribution according to a predetermined axle clearance at least for a first of the kinematic axes, and
wherein the at least one positioning variance is determined as a function of the statistical position distribution.

### Claim 8 (depends on 7)

The computer-implemented method of claim 7, wherein the calculating of the faulty positioning data, the simulating of the imaging method, and the generating and storing of the annotated training data record are repeated multiple times, and
wherein, with each repetition, the at least one positioning variance is determined again as a function of the statistical position distribution.

### Claim 9 (depends on 6)

The computer-implemented method of claim 6, wherein the predetermined model comprises a reverberation characteristic, and
wherein the at least one positioning variance is determined at least in part in accordance with the reverberation characteristic.

### Claim 10 (depends on 1)

The computer-implemented method of claim 1, wherein target dose data is predetermined,
wherein faulty dose data is calculated as a function of the target dose data,
wherein the faulty dose data deviates from the target dose data by at least one dose variance,
wherein, when the imaging method is simulated, an emission of x-ray radiation is simulated in accordance with the faulty dose data, and
wherein the annotated training data record comprises the at least one dose variance as a further annotation of the two or more simulated x-ray recordings.

### Claim 11 (depends on 1)

The computer-implemented method of claim 1, further comprising:
predicting at least one positioning variance by applying the algorithm to the two or more simulated x-ray recordings; and
adjusting the algorithm as a function of the at least one predicted positioning variance and the at least one positioning variance of the annotated training data record.

### Claim 12

A method for x-ray imaging, the method comprising:
generating two or more x-ray recordings of an object to be mapped by an x-ray imaging system;
positioning an imaging apparatus of the x-ray imaging system in accordance with predetermined target positioning data;
estimating at least one positioning variance by applying an algorithm trained in accordance with a computer-implemented method to the two or more x-ray recordings; and
correcting the two or more x-ray recordings as a function of the at least one estimated positioning variance.

### Claim 13 (depends on 12)

The method of claim 12, wherein a three-dimensional reconstruction of the object to be mapped is generated based on the corrected two or more x-ray recordings.

### Claim 14 (depends on 13)

The method of claim 13, wherein at least one subtraction angiography mapping of the object to be mapped is additionally generated based on the corrected two or more x-ray recordings.

### Claim 15 (depends on 12)

The method of claim 12, wherein at least one subtraction angiography mapping of the object to be mapped is generated based on the corrected two or more x-ray recordings.

### Claim 16

A data processing apparatus comprising:
at least one processing switching circuit,
wherein the data processing apparatus is configured to adjust the at least one processing switching circuit by:
predetermining target positioning data for an imaging apparatus;
calculating faulty positioning data as a function of the target positioning, wherein the faulty positioning data deviates from the target positioning data by at least one predetermined positioning variance;
simulating an imaging method, wherein two or more simulated x-ray recordings are generated, wherein a positioning of the imaging apparatus is simulated in accordance with the faulty positioning data;
generating an annotated training data record that comprises the two or more simulated x-ray recordings and at least one positioning variance as an annotation of the two or more simulated x-ray recordings;
predicting at least one positioning variance by applying an algorithm to the two or more simulated x-ray recordings; and
adjusting the at least one processing switching circuit as a function of the at least one predicted positioning variance and the at least one positioning variance of the annotated training data record.