X-ray dose determination device

Abstract

The x-ray dose determination device has an input unit, a processing unit, and an output unit, wherein the input unit is configured to provide a visible or infrared image of a patient to the processing unit, the processing unit is configured to determine a silhouette of the patient in the visible or infrared image of the patient, the processing unit is configured to perform a patient thickness determination comprising utilization of the patient's silhouette in the visible or infrared image of the patient, the processing unit is configured to perform an x-ray dose determination for an x-ray examination of the patient comprising utilization of the patient's thickness, and / or the processing unit is configured to perform a determination that an atypical x-ray dose is required for an x-ray examination of the patient comprising utilization of the patient's thickness, and the output unit is configured to output an indication of the x-ray dose for the x-ray examination of the patient, and / or the output unit is configured to output an indication that an atypical x-ray dose is required for an x-ray examination of the patient.

Description

[Technical Field]

[0001] The present invention relates to an X-ray dose determination device, an X-ray dose determination system, an X-ray system, an X-ray dose determination method, a computer program element, and a computer readable medium. [Background technology]

[0002] X-ray imaging of patients presents a problem in that images of overweight patients are consistently more complex. This is due in part to the difficulty of positioning the detector or achieving the correct field of view. However, a recurring problem is that of exposure. The amount of radiation emitted in taking an X-ray (CXR) must be carefully calibrated to achieve a compromise between image quality and patient radiation exposure.

[0003] In the case of underexposure, even if the image can be amplified and rescaled to provide a good grayscale representation, the noise in the image is amplified as well, often resulting in a noisy, grainy image. When such suboptimal radiation doses are applied, retakes are often necessary due to poor image quality. See, for example, W. Huda, J.A. Seibert, K. Ogden, E. Gingold, R. Schaetzing, Physics Teaching File for Radiology Residents, Upstate Medical University, (https: / / www.upstate.edu / radiology / education / rsna / radiography / issues.php.).

[0004] Except in extreme cases, overexposed images usually have excellent radiographic quality with high contrast and low noise. Unfortunately, however, patients in this situation are exposed to unnecessary radiation. In some cases, overexposure of 3 to 5 times or more can occur. See, for example, W. Huda, J. A. Seibert, K. Ogden, E. Gingold, R. Schaetzing, Physics Teaching File for Radiology Residents, Upstate Medical University, (https: / / www.upstate.edu / radiology / education / rsna / radiography / issues.php) and S. B. Gay, J. Olazagasti, J. W. Higginbotham, A. Gupta, A. Wurm, J. Nguyen, Introduction to Chest Radiology, University of Virginia Health Sciences Center, Department of Radiology, available online: (https: / / introductiontoradiology.net / courses / rad / cxr / index.html#).

[0005] Radiation exposure is primarily a function of two factors: the accelerating voltage of the x-ray tube, expressed in kV (kilovolts), and the exposure-time product (radiation intensity), expressed in mAs. kV controls the wavelength of the photons emitted by the tube and, therefore, how they pass through different tissues. Therefore, kV controls the contrast of the image. The kV parameter is less important in modern digital imaging systems, since contrast can be digitally enhanced. See, for example, L-C. Chen, G. Papandreou, F. Schroff, H. Adam, Rethinking Atrous Convolution for Semantic Image Segmentation, arXiv cs.CV, available online: https: / / arxiv.org / abs / 1706.05587v3.

[0006] mAs is a function of time that controls the number of photons that hit the detector. Thus, high mAs produce overexposed images, while low mAs produce images that are heavily corrupted by noise. See, for example, L-C. Chen, G. Papandreou, F. Schroff, H. Adam, Rethinking Atrous Convolution for Semantic Image Segmentation, arXiv cs.CV, available online: https: / / arxiv.org / abs / 1706.05587v3. Summary of the Invention [Problem to be solved by the invention]

[0007] Other parameters such as collimation and distance to the patient can also affect the quality of the x-rays produced, primarily affecting the field of view and, in the case of collimation, tissue contrast.

[0008] These issues need to be addressed. [Means for solving the problem]

[0009] It would be advantageous to have an improved technique that helps determine the X-ray dose for X-ray image acquisition. The object of the present invention is solved by the subject matter of the independent claims, with further embodiments being incorporated in the dependent claims.

[0010] In a first aspect, there is provided an x-ray dose determination apparatus comprising an input unit, a processing unit, and an output unit. The input unit is configured to provide a visible or infrared image of a patient to the processing unit. The processing unit is configured to determine a silhouette of the patient in the visible or infrared image of the patient. The processing unit is configured to perform a determination of the patient's thickness comprising use of the patient's silhouette in the visible or infrared image of the patient. The processing unit is configured to perform an x-ray dose determination for an x-ray examination of the patient comprising use of the patient's thickness, and / or the processing unit is configured to perform a determination that an atypical x-ray dose is required for the x-ray examination of the patient comprising use of the patient's thickness. The output unit is configured to output an indication of an x-ray dose for the x-ray examination of the patient, and / or the output unit is configured to output an indication that an atypical x-ray dose is required for the x-ray examination of the patient.

[0011] In this way, a simple 2D image of the person acquired by the camera can be used to determine the patient's thickness from which an accurate x-ray dose can be determined, and / or an indication to the operator that an atypical or non-atypical x-ray dose is required. In this way, for patients who are larger and / or have more body fat than normal, an accurate x-ray dose can be administered to obtain resultant x-rays with the required contrast, mitigating over- or under-exposure of the patient.

[0012] In other words, the 2D image of the patient is used to determine the patient's morphology, which can determine the x-ray dose for the x-ray examination.

[0013] Alternatively, a simple camera such as an RGB camera can be used to acquire an image from which the patient's silhouette can be assessed and used to determine the patient's thickness and, from there, the exact x-ray dose.

[0014] In one example, determining a silhouette of a patient in a visible or infrared image of the patient includes implementation by the processing unit of a segmentation machine learning algorithm that analyzes the visible or infrared image of the patient to determine a segmentation mask that represents the silhouette of the patient in the visible or infrared image of the patient.

[0015] In one example, a segmentation machine learning algorithm is trained on multiple visible or infrared images of one or more people and associated multiple segmentation masks obtained from annotations of the contours of one or more people, each visible or infrared image having an image of one person.

[0016] In one example, the processing unit is configured to determine a contour of the patient's silhouette in a visible or infrared image of the patient, including utilizing a silhouette of the patient in a visible or infrared image of the patient. Determining the patient's thickness can then include utilizing a contour of the patient's silhouette in the visible or infrared image of the patient.

[0017] In one example, the processing unit is configured to determine a plurality of feature points of a silhouette of the patient in the visible or infrared image of the patient, the feature points representing an outline of the silhouette of the patient in the visible or infrared image of the patient. Determining the thickness of the patient can then include utilizing the plurality of feature points of the silhouette of the patient in the visible or infrared image of the patient.

[0018] In one example, determining the plurality of feature points includes determining a plurality of turning points at one or more boundaries of a silhouette of the patient in a visible or infrared image of the patient, each turning point defining a feature point.

[0019] In one example, determining the turning point includes determining a first direction associated with a first pair of contiguous pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient and determining a second direction associated with a second pair of contiguous pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, wherein the turning point is determined when an angle between the first direction and the second direction is greater than or equal to a threshold angle.

[0020] In one example, a first pair of contiguous pixels at the boundary of a patient's silhouette in a visible or infrared image of the patient is contiguous with a second pair of contiguous pixels at the boundary of the patient's silhouette in a visible or infrared image of the patient, or in one example, a first pair of contiguous pixels at the boundary of a patient's silhouette in a visible or infrared image of the patient shares a common pixel with a second pair of contiguous pixels at the boundary of the patient's silhouette in a visible or infrared image of the patient.

[0021] In this way, by determining the feature points of the silhouette of the patient's image, features such as the arms, legs, waist, head, and chest can be distinguished from one another. If an X-ray examination of any of these body parts needs to be performed, the thickness of this body part can be determined from the silhouette of the 2D image. This may show the arms at the waist, for example. However, the waist itself can be identified in the silhouette from the feature points, and from the feature points, the thickness of the waist itself can be determined, allowing the correct X-ray dose for the waist X-ray examination to be determined.

[0022] In one example, determining the patient's thickness includes determining a width of the patient perpendicular to a viewing direction of a camera that acquired the visible or infrared image of the patient, including using a silhouette of the patient in a visible or infrared image of the patient, and the processing unit is configured to convert the patient's width to the patient's thickness.

[0023] In one example, converting patient width to patient thickness involves utilizing a look-up table or a regression model or a mathematical model.

[0024] In other words, knowledge of the width versus depth relationship for people can be used to convert, for example, the width of the chest or waist in a simple 2D image, as seen by x-rays in an x-ray examination, into the thickness of the chest or waist, and this thickness can be used to determine the correct x-ray dose for the examination.

[0025] In one example, determining the width of the patient perpendicular to the viewing direction of a camera that captured the visible or infrared image of the patient includes using the outline of the patient's silhouette in the visible or infrared image of the patient.

[0026] In a second aspect, there is provided an x-ray dose determination system comprising a visible or infrared camera, a processing unit, and an output unit. The visible or infrared camera is configured to acquire a visible or infrared image of a patient. The visible or infrared camera is configured to provide the visible or infrared image of the patient to the processing unit. The processing unit is configured to determine a silhouette of the patient in the visible or infrared image of the patient. The processing unit is configured to perform a determination of the patient's thickness comprising use of the patient's silhouette in the visible or infrared image of the patient. The processing unit is configured to perform an x-ray dose determination for an x-ray examination of the patient comprising use of the patient's thickness, and / or the processing unit is configured to perform a determination that an atypical x-ray dose is required for the x-ray examination of the patient comprising use of the patient's thickness. The output unit is configured to output an indication of an x-ray dose for the x-ray examination of the patient, and / or the output unit is configured to output an indication that an atypical x-ray dose is required for the x-ray examination of the patient.

[0027] In a third aspect, there is provided an X-ray system comprising an X-ray image acquisition unit, a visible or infrared camera, and a processing unit. The visible or infrared camera is configured to acquire a visible or infrared image of a patient prior to performing an X-ray examination using the X-ray image acquisition unit. The visible or infrared camera is configured to provide the visible or infrared image of the patient to the processing unit. The processing unit is configured to determine a silhouette of the patient in the visible or infrared image of the patient. The processing unit is configured to perform a determination of the patient's thickness comprising use of the patient's silhouette in the visible or infrared image of the patient. The processing unit is configured to perform an X-ray dose determination for the X-ray examination of the patient comprising use of the patient's thickness, and / or the processing unit is configured to perform a determination that an atypical X-ray dose is required for the X-ray examination of the patient comprising use of the patient's thickness, and to output an indication that an atypical X-ray dose is required for the X-ray examination of the patient using an output unit.

[0028] In a fourth aspect, there is provided a method for determining an x-ray dose, the method comprising the steps of providing a visible or infrared image of a patient to a processing unit; determining, by the processing unit, a silhouette of the patient in the visible or infrared image of the patient; performing, by the processing unit, a patient thickness determination comprising a step of utilizing the silhouette of the patient in the visible or infrared image of the patient; determining, by the processing unit, an x-ray dose for an x-ray examination of the patient comprising a step of utilizing the patient thickness, and / or determining, by the processing unit, that an atypical x-ray dose is required for the x-ray examination of the patient comprising a step of utilizing the patient thickness; and outputting, by an output unit, an indication of the x-ray dose for the x-ray examination of the patient and / or outputting, by the output unit, an indication that an atypical x-ray dose is required for the x-ray examination of the patient.

[0029] In one aspect there is provided a computer program element for controlling an apparatus according to the first aspect, configured to perform the method of the fourth aspect when executed by a processor.

[0030] In one aspect there is provided a computer program element for controlling a system according to the second aspect, configured to perform the method of the fourth aspect when executed by a processor.

[0031] In one aspect there is provided a computer program element for controlling a system according to the third aspect, configured to perform the method of the fourth aspect when executed by a processor.

[0032] Thus, according to an aspect, there is provided a computer program element for controlling one or more of the devices / systems as described above, which is adapted to perform the methods described above when the computer program element is executed by a processor.

[0033] According to another aspect, a computer readable medium having stored thereon the aforementioned computer elements is provided.

[0034] The computer program element may for example be a software program, but also an FPGA, a PLD or any other suitable digital means.

[0035] Advantageously, any advantages provided by any of the above aspects apply equally to all of the other aspects, and vice versa.

[0036] These aspects and examples will be apparent from and elucidated with reference to the embodiments described hereinafter.

[0037] Exemplary embodiments are described below with reference to the following drawings: [Brief explanation of the drawings]

[0038] [Figure 1] 1 shows an example of an X-ray dose determination device. [Figure 2] 1 shows an example of an X-ray dose determination system. [Figure 3] An example of an X-ray system is shown. [Figure 4] The method for determining the X-ray dose is shown. [Figure 5] Figure 5 shows an example of an image that has been annotated to provide a segmentation mask for use in training a segmentation machine learning algorithm. [Figure 6] Illustrates examples of direction encoding of different contour points, where encoding instance is equal to 1 and encoding instance is equal to 6. [Figure 7] An example of determining feature points will be shown. [Figure 8] An example of increasing radiation dose depending on patient thickness in chest radiography is shown, resulting in images of comparable quality. DETAILED DESCRIPTION OF THE INVENTION

[0039] FIG. 1 illustrates an example of an X-ray dose determination device 10. The device 10 includes an input unit 20, a processing unit 30, and an output unit 40. The input unit is configured to provide a visible or infrared image of a patient to the processing unit. The processing unit is configured to determine a silhouette of the patient in the visible or infrared image of the patient. The processing unit is configured to determine a patient thickness including utilizing the silhouette of the patient in the visible or infrared image of the patient. The processing unit is configured to determine an X-ray dose for an X-ray examination of the patient including utilizing the patient thickness. Additionally or alternatively, the processing unit is configured to determine that an atypical X-ray dose is required for an X-ray examination of the patient including utilizing the patient thickness. The output unit is configured to output an indication of an X-ray dose for an X-ray examination of the patient. Additionally or alternatively, the output unit is configured to output an indication that an atypical X-ray dose is required for an X-ray examination of the patient.

[0040] In this way, a simple 2D image of the person acquired by the camera can be used to determine the patient's thickness from which an accurate x-ray dose can be determined, and / or an indication to the operator that an atypical or non-atypical x-ray dose is required. In this way, for patients with larger than normal and / or larger bodies, the fat can be accommodated, allowing resultant x-rays with the necessary contrast to be acquired, and an accurate x-ray dose can be administered to mitigate patient over- or under-exposure.

[0041] In other words, the 2D image of the patient is used to determine the patient's morphology, which can determine the x-ray dose for the x-ray examination.

[0042] Alternatively, a simple camera such as an RGB camera can be used to acquire an image from which the patient's silhouette can be assessed and used to determine the patient's thickness and, from there, the exact x-ray dose.

[0043] In one example, the visible or infrared image is a 2D visible or infrared image acquired by a conventional 2D camera.

[0044] However, the visible or infrared image may be a 3D image acquired by a time-of-flight camera or a 3D camera using grid distortion techniques. The depth resolution of such cameras is generally lower than the lateral resolution, and therefore determining the silhouette of such a 3D image to determine patient thickness, and x-ray dose determination, may be more accurate than determining patient thickness and then x-ray dose from the 3D image data itself.

[0045] In one example, the indication of the x-ray dose and / or the indication that an atypical x-ray dose is required includes the x-ray dose in numerical form presented on the VDU.

[0046] In one example, an indication of the x-ray dose and / or an indication that an atypical x-ray dose is required includes a visual indication that a different dose than expected is suggested.

[0047] In one example, the indication of the x-ray dose and / or the indication that an atypical x-ray dose is required includes a visual representation in color, for example, having a first color indicating that the x-ray dose is as expected, a second color indicating that a higher dose than expected is required, and a third color indicating that a lower dose than expected is required.

[0048] In one example, the indication of the x-ray dose and / or the indication that an atypical x-ray dose is required includes a visual representation of a marker shown on a scale; thus, for example, a scale with annotated upper and lower limits can be presented to the technician using a slider or marker that moves to the required dose after processing of the visible or infrared image.

[0049] According to one example, determining a silhouette of a patient in a visible or infrared image of the patient includes implementing by the processing unit a segmentation machine learning algorithm that analyzes the visible or infrared image of the patient to determine a segmentation mask representing the silhouette of the patient in the visible or infrared image of the patient.

[0050] According to one example, a segmentation machine learning algorithm is trained on multiple visible or infrared images of one or more people and associated multiple segmentation masks obtained from annotations of the contours of the one or more people, each visible or infrared image having an image of one person.

[0051] In one example, machine learning algorithms utilize DeepLab.

[0052] According to one example, the processing unit is configured to determine a contour of the patient's silhouette in a visible or infrared image of the patient, including utilizing a silhouette of the patient in a visible or infrared image of the patient. Determining the patient's thickness can then include utilizing a contour of the patient's silhouette in the visible or infrared image of the patient.

[0053] According to one example, the processing unit is configured to determine a plurality of feature points of a silhouette of the patient in the visible or infrared image of the patient, the feature points representing an outline of a silhouette of the patient in the visible or infrared image of the patient. Determining the thickness of the patient can then include utilizing the plurality of feature points of the silhouette of the patient in the visible or infrared image of the patient.

[0054] According to one example, determining the plurality of feature points includes determining a plurality of turning points at one or more boundaries of a silhouette of the patient in a visible or infrared image of the patient, each turning point defining a feature point.

[0055] According to one example, determining the turning point includes determining a first direction associated with a first pair of consecutive pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, and determining a second direction associated with a second pair of consecutive pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, and the turning point is determined when an angle between the first direction and the second direction is greater than or equal to a threshold angle.

[0056] In one example, the threshold angle is 35 degrees, or 40 degrees, or 45 degrees, or 50 degrees, or 55 degrees, or 60 degrees, or 65 degrees, or 70 degrees, or 75 degrees, or 80 degrees, or 85 degrees, or 90 degrees, or 95 degrees, or 100 degrees, or 105 degrees, or 110 degrees.

[0057] According to one example, a first pair of contiguous pixels at the boundary of a patient's silhouette in a visible or infrared image of the patient is contiguous with a second pair of contiguous pixels at the boundary of the patient's silhouette in the visible or infrared image of the patient, or according to one example, a first pair of contiguous pixels at the boundary of the patient's silhouette in the visible or infrared image of the patient shares a common pixel with a second pair of contiguous pixels at the boundary of the patient's silhouette in the visible or infrared image of the patient.

[0058] In this way, by determining feature points in the silhouette of the patient's image, features such as the arms, legs, waist, head, and chest can be distinguished from one another, and if an X-ray examination of any of these body parts is required to take the thickness of this body part, for example, the arms can be shown at the waist level from the silhouette, but the waist itself can be identified within the silhouette from feature points that allow the thickness of the waist itself to be determined, and the correct X-ray dose for the X-ray examination of the waist can be determined.

[0059] According to one example, determining the patient's thickness includes determining a width of the patient perpendicular to a viewing direction of a camera that acquired the visible or infrared image of the patient, including using a silhouette of the patient in a visible or infrared image of the patient, and the processing unit is configured to convert the patient's width to the patient's thickness.

[0060] According to one example, converting patient width to patient thickness includes utilizing a look-up table or a regression model or a mathematical model.

[0061] In other words, knowledge of the width versus depth relationship for people can be used to convert, for example, the width of the chest or waist in a simple 2D image, as seen by x-rays in an x-ray examination, into the thickness of the chest or waist, and this thickness can be used to determine the correct x-ray dose for the examination.

[0062] In one example, the look-up table is associated with the body part that is to be x-rayed by the patient.

[0063] In one example, the look-up table takes into account component parts and / or organs and / or within the body part.

[0064] In one example, the regression model is associated with the body part that is the subject of an x-ray examination of the patient.

[0065] In one example, the regression takes into account component parts and / or organs and / or within the body part.

[0066] In one example, mathematics is associated with the body parts that are subject to x-ray examination of a patient.

[0067] In one example, the mathematical model considers component parts and / or organs of a body part and / or within a body part.

[0068] Thus, the thickness of a body part such as the legs, chest, stomach, etc. can be determined from visible or infrared images and from knowledge of the body part, the expected amount of other structures such as bone or organs and how much these structures absorb X-rays can be determined, and how much body fat is present in addition to the organs / bone can also be taken into account.

[0069] According to one example, determining the width of the patient perpendicular to the viewing direction of a camera that acquired the visible or infrared image of the patient includes utilizing the outline of the patient's silhouette in the visible or infrared image of the patient.

[0070] In one example, determining the x-ray dose for an x-ray examination of a patient involves utilizing a correlation between x-ray dose and patient thickness.

[0071] FIG. 2 illustrates an example of an x-ray dose determination system 100. The system 100 includes a visible or infrared camera 110, a processing unit 120, and an output unit 130. The visible or infrared camera is configured to acquire a visible or infrared image of a patient. The visible or infrared camera is configured to provide the visible or infrared image of the patient to the processing unit. The processing unit is configured to determine a silhouette of the patient in the visible or infrared image of the patient. The processing unit is configured to determine a patient thickness including utilizing the silhouette of the patient in the visible or infrared image of the patient. The processing unit is configured to determine an x-ray dose for an x-ray examination of the patient including utilizing the patient thickness. Additionally or alternatively, the processing unit is configured to determine that an atypical x-ray dose is required for an x-ray examination of the patient including utilizing the patient thickness. The output unit is configured to output an indication of an x-ray dose for an x-ray examination of the patient. Additionally or alternatively, the output unit is configured to output an indication that an atypical x-ray dose is required for an x-ray examination of the patient.

[0072] In one example, the visible or infrared image is a 2D image acquired by a 2D camera.

[0073] However, the visible or infrared image may be a 3D image acquired by a 3D camera using techniques such as time-of-flight or grid distortion.

[0074] In one example, determining a silhouette of a patient in a visible or infrared image of the patient includes implementation by the processing unit of a segmentation machine learning algorithm for analyzing the visible or infrared image of the patient to determine a segmentation mask representing the silhouette of the patient in the visible or infrared image of the patient.

[0075] In one example, a segmentation machine learning algorithm is trained on multiple visible or infrared images of one or more people and associated multiple segmentation masks obtained from annotations of the contours of one or more people, each visible or infrared image having an image of one person.

[0076] In one example, machine learning algorithms utilize DeepLab.

[0077] In one example, the processing unit is configured to determine a contour of the patient's silhouette in a visible or infrared image of the patient, including utilizing a contour of the patient's silhouette in a visible or infrared image of the patient. Determining the patient's thickness can include utilizing a contour of the patient's silhouette in a visible or infrared image of the patient.

[0078] In one example, the processing unit is configured to determine a plurality of feature points of a silhouette of the patient in the visible or infrared image of the patient, the feature points representing an outline of a silhouette of the patient in the visible or infrared image of the patient. Determining the thickness of the patient can include utilizing the plurality of feature points of the silhouette of the patient in the visible or infrared image of the patient.

[0079] In one example, determining the plurality of feature points includes determining a plurality of turning points at one or more boundaries of a silhouette of the patient in a visible or infrared image of the patient, each turning point defining a feature point.

[0080] In one example, determining the turning point includes determining a first direction associated with a first pair of consecutive pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, and determining a second direction associated with a second pair of consecutive pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, the first direction and the second.

[0081] In one example, the threshold angle is 35 degrees, or 40 degrees, or 45 degrees, or 50 degrees, or 55 degrees, or 60 degrees, or 65 degrees, or 70 degrees, or 75 degrees, or 80 degrees, or 85 degrees, or 90 degrees, or 95 degrees, or 100 degrees, or 105 degrees, or 110 degrees.

[0082] In one example, a first pair of contiguous pixels at the boundary of a patient's silhouette in a visible or infrared image of the patient is contiguous with a second pair of contiguous pixels at the boundary of the patient's silhouette in a visible or infrared image of the patient.

[0083] In one example, a first pair of contiguous pixels at the boundary of a patient's silhouette in a visible or infrared image of the patient shares a common pixel with a second pair of contiguous pixels at the boundary of the patient's silhouette in a visible or infrared image of the patient.

[0084] In one example, determining the patient's thickness includes determining a width of the patient perpendicular to a viewing direction of a camera that acquired the visible or infrared image of the patient, including using a silhouette of the patient in a visible or infrared image of the patient, and the processing unit is configured to convert the patient's width to the patient's thickness.

[0085] In one example, converting the patient width to the patient thickness includes utilizing a look-up table or a regression model or a mathematical model.

[0086] In one example, the look-up table is associated with the body part that is to be x-rayed by the patient.

[0087] In one example, the look-up table takes into account component parts and / or organs and / or within the body part.

[0088] In one example, the regression model is associated with the body part that is the subject of an x-ray examination of the patient.

[0089] In one example, the regression takes into account component parts and / or organs and / or within the body part.

[0090] In one example, mathematics is associated with the body parts that are subject to x-ray examination of a patient.

[0091] In one example, the mathematical model takes into account component parts and / or organs of a body part and / or within a body part.

[0092] In one example, determining the width of the patient perpendicular to the viewing direction of a camera that captured the visible or infrared image of the patient includes using the outline of the patient's silhouette in the visible or infrared image of the patient.

[0093] In one example, determining the x-ray dose for an x-ray examination of a patient involves utilizing a correlation between x-ray dose and patient thickness.

[0094] FIG. 3 illustrates an example of an X-ray system 200. The system includes an X-ray image acquisition unit 210, a visible or infrared camera 220, and a processing unit 230. The visible or infrared camera is configured to acquire a visible or infrared image of a patient prior to performing an X-ray examination using the X-ray image acquisition unit. The visible or infrared camera is configured to provide the visible or infrared image of the patient to the processing unit. The processing unit is configured to determine a silhouette of the patient in the visible or infrared image of the patient. The processing unit is configured to determine a patient thickness including utilizing the silhouette of the patient in the visible or infrared image of the patient. The processing unit is configured to determine an X-ray dose for the X-ray examination of the patient including utilizing the patient thickness. Additionally or alternatively, the processing unit is configured to determine that an atypical X-ray dose is required for the X-ray examination of the patient, including utilizing the patient thickness, and output an indication that an atypical X-ray dose is required for the X-ray examination of the patient using the output unit.

[0095] In one example, the visible or infrared image is a 2D image acquired by a 2D camera.

[0096] However, the visible or infrared image may be a 3D image acquired by a 3D camera using techniques such as time-of-flight or grid distortion.

[0097] In one example, determining a silhouette of a patient in a visible or infrared image of the patient includes implementation by the processing unit of a segmentation machine learning algorithm for analyzing the visible or infrared image of the patient to determine a segmentation mask representing the silhouette of the patient in the visible or infrared image of the patient.

[0098] In one example, a segmentation machine learning algorithm is trained on multiple visible or infrared images of one or more people and associated multiple segmentation masks obtained from annotations of the contours of one or more people, each visible or infrared image having an image of one person.

[0099] In one example, machine learning algorithms utilize DeepLab.

[0100] In one example, the processing unit is configured to determine a contour of the patient's silhouette in a visible or infrared image of the patient, including utilizing a contour of the patient's silhouette in a visible or infrared image of the patient. Determining the patient's thickness can include utilizing a contour of the patient's silhouette in a visible or infrared image of the patient.

[0101] In one example, the processing unit is configured to determine a plurality of feature points of a silhouette of the patient in the visible or infrared image of the patient, the feature points representing an outline of a silhouette of the patient in the visible or infrared image of the patient. Determining the thickness of the patient can include utilizing the plurality of feature points of the silhouette of the patient in the visible or infrared image of the patient.

[0102] In one example, determining the plurality of feature points includes determining a plurality of turning points at one or more boundaries of a silhouette of the patient in a visible or infrared image of the patient, each turning point defining a feature point.

[0103] In one example, determining the turning point includes determining a first direction associated with a first pair of consecutive pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, and determining a second direction associated with a second pair of consecutive pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, and the turning point is determined when an angle between the first direction and the second direction is greater than or equal to a threshold angle.

[0104] In one example, the threshold angles are 35 degrees, 40 degrees, 45 degrees, 50 degrees, 55 degrees, 60 degrees, 65 degrees, 70 degrees, 75 degrees, 80 degrees, 85 degrees, 90 degrees, 95 degrees, 100 degrees, 105 degrees, and 110 degrees.

[0105] In one example, a first pair of contiguous pixels at the boundary of a patient's silhouette in a visible or infrared image of the patient is contiguous with a second pair of contiguous pixels at the boundary of the patient's silhouette in a visible or infrared image of the patient.

[0106] In one example, a first pair of consecutive pixels at the boundary of a patient's silhouette in a visible or infrared image of the patient shares a common pixel with a second pair of consecutive pixels at the boundary of the patient's silhouette in a visible or infrared image of the patient.

[0107] In one example, determining the patient's thickness includes determining a width of the patient perpendicular to a viewing direction of a camera that acquired the visible or infrared image of the patient, including using a silhouette of the patient in a visible or infrared image of the patient, and the processing unit is configured to convert the patient's width to the patient's thickness.

[0108] In one example, converting the patient width to the patient thickness includes utilizing a look-up table or a regression model or a mathematical model.

[0109] In one example, the look-up table is associated with the body part that is to be x-rayed by the patient.

[0110] In one example, the look-up table takes into account component parts and / or organs and / or within the body part.

[0111] In one example, the regression model is associated with the body part that is the subject of an x-ray examination of the patient.

[0112] In one example, the regression takes into account component parts and / or organs and / or within the body part.

[0113] In one example, mathematics is associated with the body parts that are subject to x-ray examination of a patient.

[0114] In one example, the mathematical model takes into account component parts and / or organs of a body part and / or within a body part.

[0115] In one example, determining the width of the patient perpendicular to the viewing direction of a camera that captured the visible or infrared image of the patient includes using the outline of the patient's silhouette in the visible or infrared image of the patient.

[0116] In one example, determining the x-ray dose for an x-ray examination of a patient involves utilizing a correlation between x-ray dose and patient thickness.

[0117] 4 shows the basic steps of an x-ray dose determination method 300. Method 300 includes: providing 310 a visible or infrared image of the patient to a processing unit; determining 320, by the processing unit, a silhouette of the patient in a visible or infrared image of the patient; determining 330, by the processing unit, the thickness of the patient, including using a silhouette of the patient in a visible or infrared image of the patient; determining 340, by a processing unit, an x-ray dose for the x-ray examination of the patient, comprising utilizing the patient thickness; and / or determining 340, by a processing unit, that an atypical x-ray dose is required for the x-ray examination of the patient, comprising utilizing the patient thickness; outputting, by an output unit, an indication of an x-ray dose for the x-ray examination of the patient, and / or outputting, by the output unit, an indication that an atypical x-ray dose is required for the x-ray examination of the patient; It has.

[0118] In one example, the visible or infrared image is a 2D image acquired by a 2D camera.

[0119] However, the visible or infrared image may be a 3D image acquired by a 3D camera using techniques such as time-of-flight or grid distortion.

[0120] In one example, determining the silhouette of the patient within the visible or infrared image of the patient includes implementing, by the processing unit, a segmentation machine learning algorithm to analyze the visible or infrared image of the patient to determine a segmentation mask representing the silhouette of the patient within the visible or infrared image of the patient.

[0121] In one example, a segmentation machine learning algorithm is trained on multiple visible or infrared images of one or more people and associated multiple segmentation masks obtained from annotations of the contours of one or more people, each visible or infrared image having an image of one person.

[0122] In one example, machine learning algorithms utilize DeepLab.

[0123] In one example, the method includes determining, by the processing unit, an outline of a silhouette of the patient in a visible or infrared image of the patient, and includes utilizing the silhouette of the patient in the visible or infrared image of the patient. Determining the thickness of the patient can include utilizing the outline of the silhouette of the patient in the visible or infrared image of the patient.

[0124] In one example, the method includes determining, by the processing unit, a plurality of feature points of a silhouette of the patient in the visible or infrared image of the patient, the feature points representing an outline of a silhouette of the patient in the visible or infrared image of the patient. Determining the thickness of the patient can include utilizing the plurality of feature points of the silhouette of the patient in the visible or infrared image of the patient.

[0125] In one example, determining a plurality of feature points includes determining a plurality of turning points at one or more boundaries of a silhouette of the patient in a visible or infrared image of the patient, each turning point defining a feature point.

[0126] In one example, determining the turning point includes determining a first direction associated with a first pair of consecutive pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, and determining a second direction associated with a second pair of consecutive pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, and the turning point is determined when an angle between the first direction and the second direction is greater than or equal to a threshold angle.

[0127] In one example, the threshold angle is 35 degrees, or 40 degrees, or 45 degrees, or 50 degrees, or 55 degrees, or 60 degrees, or 65 degrees, or 70 degrees, or 75 degrees, or 80 degrees, or 85 degrees, or 90 degrees, or 95 degrees, or 100 degrees, or 105 degrees, or 110 degrees.

[0128] In one example, a first pair of contiguous pixels at the boundary of a patient's silhouette in a visible or infrared image of the patient is contiguous with a second pair of contiguous pixels at the boundary of the patient's silhouette in a visible or infrared image of the patient.

[0129] In one example, a first pair of contiguous pixels at the boundary of a patient's silhouette in a visible or infrared image of the patient shares a common pixel with a second pair of contiguous pixels at the boundary of the patient's silhouette in a visible or infrared image of the patient.

[0130] In one example, determining the patient's thickness includes utilizing a silhouette of the patient in a visible or infrared image of the patient, determining a width of the patient perpendicular to a viewing direction of a camera that acquired the visible or infrared image of the patient, and the method includes converting, by a processing unit, the patient's width to a patient's thickness.

[0131] In one example, converting the patient width to the patient thickness includes utilizing a look-up table or a regression model or a mathematical model.

[0132] In one example, the look-up table is associated with the body part that is to be x-rayed by the patient.

[0133] In one example, the look-up table considers component parts and / or organs of a body part and / or within a body part.

[0134] In one example, the regression model is associated with the body part that is the subject of an x-ray examination of the patient.

[0135] In one example, the regression takes into account component parts and / or organs and / or within the body part.

[0136] In one example, mathematics is associated with the body parts that are subject to x-ray examination of a patient.

[0137] In one example, the mathematical model takes into account component parts and / or organs of a body part and / or within a body part.

[0138] In one example, determining the width of the patient perpendicular to the viewing direction of the camera that acquired the visible or infrared image of the patient includes utilizing the outline of the patient's silhouette within the visible or infrared image of the patient.

[0139] In one example, determining the x-ray dose for an x-ray examination of a patient involves utilizing a correlation between x-ray dose and patient thickness.

[0140] Here, the X-ray dose determination device, the X-ray dose determination system, the X-ray system, and the X-ray dose determination method will be specifically described with reference to FIGS.

[0141] The following relates to the use of 2D visible or infrared images, however the visible or infrared images may also be 3D images acquired by a time-of-flight camera or a 3D camera using grid distortion techniques, and the following discussion of 2D images can also be applied to 3D images.

[0142] The X-ray dose determination system, X-ray system, and X-ray dose determination method provide for patient position and body shape to be assessed via 2D plain images to determine thickness and allow the X-ray radiation dose to be determined so that kV and mAs parameters can be determined, resulting in better quality X-ray images. Using the 2D images, the radiation dose can be determined to help technicians select appropriate voltage (kV) and radiation intensity (mAs) parameters for the optimal X-ray procedure by estimating the patient's body type and thickness for subsequent X-ray passes, and for a fixed kV, radiation intensity (mAs). Avoiding underexposed images reduces the number of retakes, ultimately reducing equipment utilization costs and patient X-ray exposure. Avoiding overexposed images reduces patient exposure, ultimately resulting in better patient outcomes.

[0143] The inventors recognized that the human body has a certain scale similarity that can be applied across body parts, for example, there is a relationship between the width at the waist and the depth of the person at the waist (perpendicular to the width), and there is a relationship between the width at the chest and the depth of the person at the waist (perpendicular to the width), and such scaling has a certain scale similarity that can be applied across body parts. This led to the recognition that for a portion of the body that is the subject of an x-ray examination to obtain an x-ray image, a simple 2D image of the patient can be used to determine the thickness of the patient as seen by the x-rays passing through the body. This thickness can then be used to determine the x-ray dose that will provide acceptable image quality for the examination without over- or under-exposing.

[0144] Thus, for example, one or more photographs of a patient acquired by an RGB camera associated with an X-ray image acquisition unit actually act as a body type assessment module, and are evaluated by a processing unit that determines the thickness of the patient as seen by X-rays based on the one or more photographs. This allows the radiation dose for the patient to be determined. Details such as the patient's height and other characteristics, such as the width of the patient's chest, waist, and other body parts, can determine not only the thickness through the patient but also the thickness of the fat layer through which the X-rays propagate, thereby enabling a more accurate determination of the radiation dose to be determined. In effect, the patient's basic body type can be determined. This can determine the body thickness and how much fat constitutes that thickness. However, determining only the thickness through the patient without determining the fat layer can itself be used to determine the patient's X-ray dose that will result in a satisfactory X-ray image. Therefore, the patient's determined dose or determined thickness at the location where the X-ray examination is to be performed can then utilize the dose level provided to them or can be notified to a technician who can recognize that a different dose level than normally used is currently required, such as a higher dose for a heavier patient. The determined dose itself can be automatically provided to the x-ray imaging unit for automatic adjustment of dose levels as needed without interaction from the technician. As is becoming more common, cameras with time-of-flight functionality or grid distortion capable of acquiring 3D images can also be utilized rather than simple 2D visible or infrared cameras.

[0145] The processing unit functioning as the body type assessment module can be calibrated to convert pixel size seen by the camera into physical measurements, such as the thickness of tissue traversed by the x-rays. This calibration can be achieved by utilizing an object of known physical dimensions (an "etalon") placed near the patient during the study. This allows the camera to be moved closer or farther away from the patient. However, if the angular extent of the pixel is known, a simple determination of the distance to the patient provides the calibration, and if the camera is utilized at a fixed distance, the calibration need only be performed once. There are many such methods for calibrating 2D camera images, so that the actual size of the imaged object can be determined. The above discussion applies to infrared cameras, RGB cameras, and indeed grayscale cameras.

[0146] Therefore, for preliminary X-rays, a photograph of the patient is taken with an RGB camera, and with the help of AI, a subsequent X-ray recommendation regarding the required dose is predicted. A segmentation network, such as one using DepLabV3, is used to obtain the patient's silhouette in the photograph (or image) acquired by the RGB camera. The segmentation network can be trained using an annotated dataset. Once the patient's silhouette is obtained, specific key points can be determined that allow for the evaluation of desired measurements, such as whether the patient is lying flat, the distance between the top of the patient's head and the end of the legs, and most importantly, the thickness through the patient and the patient's fat layer. Once the patient and, optionally, the thickness through the patient's fat layer are evaluated, if the thickness is greater than expected, a "high patient" warning can be displayed on the screen to allow the technician to adapt the dose. Alternatively or additionally, the required dose can be automatically adapted using a precalculated lookup table or other means.

[0147] A hybrid AI segmentation algorithm using the third version of DeepLab is used to extract a patient's silhouette from 2D images acquired by an RGB camera. Segmenting a patient's silhouette while lying in bed uses a classic supervised machine learning task, similar to road scene segmentation or lung segmentation in autonomous driving. See, for example, L.C. Chen, G. Papandreou, F. Schroff, and H. Adam, "Rethinking Atrous Convolution for Semantic Image Segmentation," arXiv cs.CV, available online: https: / / arxiv.org / abs / 1706.05587v3. To train the hybrid AI segmentation algorithm, an annotated dataset of 2D camera images of a patient or person is used. An example of an annotated image is shown in Figure 5, where pixels in the image are identified as either patient-related or non-patient-related. This annotated image is effectively a segmentation mask that is also the patient's silhouette. Both images, along with multiple examples of other 2D images and their annotated counterparts, are then used to train a hybrid AI segmentation algorithm. Thus, a 2D image of a patient on a bed being prepared for an X-ray examination is acquired and fed into the hybrid AI segmentation algorithm, which outputs a segmentation mask that provides the patient's silhouette.

[0148] Therefore, for this training, two pairs of images are used: a source image and a segmentation mask. The mask is obtained from annotations in the source image; at a simple level, it can be obtained by photographing the source image in paint and then drawing the outline of the patient object with a mouse. A mask is created from these annotations. It is an image of the same size as the source image. The mask is completely black and preserves only the boundary pixels and the pixels inside the boundary. The image is given as input to a fully convolutional neural network (FCN), which can be implemented using DeepLab or another algorithm, and the result is obtained. This result is compared with the mask, and the goal is to minimize the L1 loss, which is the sum of the absolute differences for each pixel. Another common option is the L2 loss, i.e., the sum of squared differences. Maximizing the Dice Score (also known as Intersection over Union, IoU) is also possible. The error is used to modify the neural network weights via backpropagation and gradient descent.

[0149] Then, during the inference phase, when a segmentation mask is needed for a new image, only the image is given at the input of the neural network, and the mask is obtained at its output.

[0150] Thus, in the particular example described herein, patient images acquired by an RGB camera are annotated to define the patient's silhouette on the image, DeepLabV3 is then trained, and finally, new patient silhouettes are obtained by feeding the images to the trained DeepLab algorithm.

[0151] We mention using DeepLab, but other networks such as UNet can also be used.

[0152] DeepLab is a family of highly efficient fully convolutional neural networks for image semantic segmentation. The first version of DeepLab appeared around the same time as UNet (2015), and they share the overall "encoder-decoder" architecture. However, there are important differences.

[0153] UNet has a symmetric encoder-decoder architecture, which means the sizes of the intermediate tensors mirror each other at different stages. Furthermore, UNet has a concatenation mechanism that helps to obtain segmentation results that closely match the boundaries of objects in the original image.

[0154] Conversely, DeepLab has a large encoder called a "projection head" and a smaller decoder. It is common to train a large classification network such as ResNet101 on the ImageNet dataset and use it as DeepLab's encoder. The third version of DeepLab utilizes Atrous Spatial Pyramid Pooling (ASPP), which deploys inverse transposition convolutions at different scales to match the features extracted by the encoder to the image. The result is a segmentation mask, where the shapes of the corresponding objects closely match, which is why DeepLab V3 was used, although other algorithms such as UNet could also be used.

[0155] Therefore, a hybrid AI algorithm is used to extract the patient's silhouette and assess the thickness of the transverse and transverse fat layers. This technique is applicable to both face and profile photographs, and can be applied to patients in prone, supine, or lateral positions.

[0156] However, rather than simply utilizing the patient's silhouette, silhouette feature points are extracted, allowing silhouettes to be defined and body parts to be distinguished from each other, such as arms from a body track. From these, accurate thicknesses at the thoracic and lumbar regions can be determined. This is done as follows, with reference to Figures 6 and 7.

[0157] Once the silhouette is extracted, a silhouette contour is obtained. A starting point on the edge of the silhouette is arbitrarily selected, and all points c of the contour are extracted. i is converted into a sequence of numbers based on the direction of the neighboring pixels. The number codes of the neighboring pixels are shown in the first diagram of Figure 6, where the directions are numbered 0 to 8 at 45 degree intervals. In the second diagram of Figure 6, the direction of the neighboring pixel is 1, and the direction of the next pixel is shown in the third diagram of Figure 6, where the direction is 6. Therefore, the encoding c of this part of the contour i are 3 and 6.

[0158] Adjacent direction c i Once the sequence of is obtained for the boundary of the silhouette, it can be calculated by the relation d i ←→c i -c0≠0 By applying i can be used to create a sequence of

[0159] That is, a turning point is defined by the fact that the direction of its neighboring pixels is different from the direction of the starting point.

[0160] Finally, the feature point F i teeth, F i ←→|d i+1 -d i |=X It can be defined using the following relationship:

[0161] That is, the feature point F iis defined when the rotation direction of the contour is 90° for a value of X equal to 2, but other values ​​can be used and the average direction can be calculated for different parts of the silhouette boundary or contour. This process is shown in Figure 7.

[0162] The feature points then define or encode the patient's body shape.

[0163] At this stage, the distances between feature points are known in terms of pixels. To convert these pixel measurements into physical quantities, an etalon can be used. An etalon is any object of known physical dimensions that is imaged along with the patient. A good candidate for an etalon is the L / R marker that technicians must already insert into each x-ray to indicate the left and right sides of the patient, respectively. By requiring that the marker always be a physical, predetermined size, we can turn this marker into an etalon.

[0164] The feature points allow the accurate width of different parts of the patient's body, e.g., the waist and chest, to be determined, and from this width, the thickness through the patient, e.g., the thickness perpendicular to the width, can be determined from known values ​​for body thickness versus perpendicular body width. The x-ray dose required as a function of body thickness to produce a satisfactory x-ray image is then used, as shown, for example, in Figure 8, which shows the relative increase in x-ray dose as a function of body thickness for a chest x-ray.

[0165] Thus, in a situation where a patient with a 40 cm thick chest is x-rayed, a 10-fold increase in x-ray dose is required to obtain an x-ray image of equivalent quality as if the patient had only a 20 cm thick chest. This new technology allows this correction to be made automatically, and if necessary, can simultaneously inform the technician who can verify or initiate the necessary dose change, or this change can be made automatically.

[0166] As detailed above, linking patient thickness with patient width is used. Here, a lookup table can be utilized, which may be a two- or three-column table, where column 1 is, for example, width at the lumbar / thoracic region and column 2 is thickness at the lumbar / thoracic region. A graph such as that shown in FIG. 8 can then be utilized to provide dose adaptation. However, the dose adaptation required for the lumbar / thoracic region may be in the third column of the lookup table. This is then repeated for shoulder width, hip width, etc. A polynomial regression model or any AI regression model can also be fitted to enable conversion between width and thickness.

[0167] In another exemplary embodiment, a computer program or a computer program element is provided, characterized in that it is configured to perform, on a suitable device or system, any of the method steps of the method according to one of the previous embodiments.

[0168] Thus, the computer program element may be stored in a computer unit that may be part of the embodiments. This computing unit may be configured to execute or trigger the execution of the steps of the above-mentioned method. Furthermore, it may be configured to operate the components of the above-mentioned system. The computing unit may be configured to operate automatically and / or to execute user commands. The computer program may be loaded into the working memory of a data processor. The data processor may thus be equipped to execute the method according to one of the above-mentioned embodiments.

[0169] This exemplary embodiment of the present invention encompasses both computer programs that use the present invention from the beginning and computer programs that, through updates, turn existing programs into programs that use the present invention.

[0170] Furthermore, the computer program element may provide all the steps necessary to fulfill the steps of the exemplary embodiments of the procedures described above.

[0171] According to a further exemplary embodiment of the present invention, a computer readable medium such as a CD-ROM, a USB stick, etc. is presented, the computer readable medium having stored thereon a computer program element, which computer program element is described by the previous section.

[0172] The computer program may be stored and / or distributed on a suitable medium, such as an optical storage medium or a solid-state medium, supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless communication systems.

[0173] However, the computer program may also be presented via a network such as the World Wide Web and downloaded into the working memory of a data processor from such a network. According to a further exemplary embodiment of the invention, a medium for making a computer program element available for downloading is provided, the computer program element being configured to perform a method according to one of the aforementioned embodiments of the invention.

[0174] It should be noted that embodiments of the present invention are described with reference to different subject matters. In particular, some embodiments are described with reference to method-type claims, and other embodiments are described with reference to apparatus-type claims. However, those skilled in the art will gather from the above and following description that, unless otherwise indicated, any combination of features belonging to one type of subject matter, as well as any combination between features relating to different subject matters, are considered to be disclosed in this application. However, all features can be combined to provide synergistic effects that are greater than the simple sum of the features.

[0175] While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered exemplary or illustrative and not restrictive. The invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the dependent claims.

[0176] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single processor or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be interpreted as limiting the scope.

Claims

1. 1. An X-ray dose determination device, comprising: An input unit; a processing unit; Output unit and and the input unit is configured to provide a visible or infrared image of a patient to the processing unit; the processing unit is configured to determine a silhouette of a patient in a visible or infrared image of the patient; the processing unit is configured to perform a determination of the patient's thickness comprising use of a silhouette of the patient in a visible or infrared image of the patient; the processing unit is configured to perform an x-ray dose determination for an x-ray examination of the patient, the x-ray dose determination comprising a utilization of the patient's thickness; and / or the processing unit is configured to perform a determination that an atypical x-ray dose is required for an x-ray examination of the patient, the atypical x-ray dose determination comprising a utilization of the patient's thickness; the output unit is configured to output an indication of an X-ray dose for the X-ray examination of the patient, and / or the output unit is configured to output an indication that an atypical X-ray dose is required for the X-ray examination of the patient. X-ray dose determination device.

2. 10. The apparatus of claim 1, wherein determining a silhouette of the patient in the visible or infrared image of the patient comprises implementation by the processing unit of a segmentation machine learning algorithm to analyze the visible or infrared image of the patient to determine a segmentation mask representing the silhouette of the patient in the visible or infrared image of the patient.

3. 3. The apparatus of claim 2, wherein the segmentation machine learning algorithm is trained on a plurality of visible or infrared images of one or more people and a plurality of associated segmentation masks obtained from annotations of contours of the one or more people, each visible or infrared image comprising an image of one person.

4. 4. The apparatus of claim 1, wherein the processing unit is configured to determine an outline of a patient's silhouette in a visible or infrared image of the patient, and wherein the processing unit is configured to determine the patient's thickness in a visible or infrared image of the patient, and wherein the processing unit is configured to determine the patient's thickness in a visible or infrared image of the patient.

5. 5. The apparatus of claim 4, wherein the processing unit is configured to determine a plurality of feature points of a silhouette of the patient in the visible or infrared image of the patient that represent an outline of a silhouette of the patient in the visible or infrared image of the patient, and wherein determining the thickness of the patient comprises utilizing a plurality of feature points of the silhouette of the patient in the visible or infrared image of the patient.

6. 6. The apparatus of claim 5, wherein determining the plurality of feature points comprises determining a plurality of turning points in one or more boundaries of a silhouette of the patient in the visible or infrared image of the patient, each turning point defining a feature point.

7. 7. The apparatus of claim 6, wherein determining the turning point comprises determining a first direction associated with a first pair of consecutive pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient and determining a second direction associated with a second pair of consecutive pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, and wherein a turning point is determined when an angle between the first direction and the second direction is greater than or equal to a threshold angle.

8. 8. The apparatus of claim 7, wherein a first pair of contiguous pixels at a boundary of a patient's silhouette in the visible or infrared image of the patient is contiguous with a second pair of contiguous pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient, or a first pair of contiguous pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient shares a common pixel with a second pair of contiguous pixels at a boundary of the patient's silhouette in the visible or infrared image of the patient.

9. 9. The apparatus of claim 1, wherein determining the patient's thickness comprises determining a patient's width perpendicular to a line of sight of a camera that acquired the visible or infrared image of the patient, the determination comprising utilizing a silhouette of the patient in a visible or infrared image of the patient, and the processing unit is configured to convert the patient's width to the patient's thickness.

10. 10. The apparatus of claim 9, wherein converting the patient width to the patient thickness comprises utilizing a look-up table or a regression model or a mathematical model.

11. 11. Apparatus according to any one of claims 9 to 10 when dependent on any one of claims 4 to 8, wherein determining the width of the patient perpendicular to the line of sight of a camera that acquired the visible or infrared image of the patient comprises utilising the outline of a silhouette of the patient in the visible or infrared image of the patient.

12. 1. An X-ray dose determination system comprising: a visible or infrared camera; a processing unit; Output unit and Equipped with the visible or infrared camera is configured to acquire a visible or infrared image of the patient; the visible or infrared camera is configured to provide a visible or infrared image of the patient to the processing unit; the processing unit is configured to determine a silhouette of a patient in a visible or infrared image of the patient; the processing unit is configured to perform a determination of the patient's thickness comprising use of a silhouette of the patient in a visible or infrared image of the patient; the processing unit is configured to perform an x-ray dose determination for an x-ray examination of the patient, the x-ray dose determination comprising a utilization of the patient's thickness; and / or the processing unit is configured to perform a determination that an atypical x-ray dose is required for an x-ray examination of the patient, the atypical x-ray dose determination comprising a utilization of the patient's thickness; the output unit is configured to output an indication of an X-ray dose for the X-ray examination of the patient, and / or the output unit is configured to output an indication that an atypical X-ray dose is required for the X-ray examination of the patient. X-ray dose determination system.

13. 1. An X-ray system comprising: an X-ray image acquisition unit; a visible or infrared camera; Processing unit and and the visible or infrared camera is configured to acquire a visible or infrared image of a patient prior to performing an X-ray examination using the X-ray image acquisition unit; the visible or infrared camera is configured to provide a visible or infrared image of the patient to the processing unit; the processing unit is configured to determine a silhouette of a patient in a visible or infrared image of the patient; the processing unit is configured to perform a determination of the patient's thickness comprising use of a silhouette of the patient in a visible or infrared image of the patient; the processing unit is configured to perform an x-ray dose determination for the x-ray examination of the patient, the determination comprising utilization of the patient thickness; and / or the processing unit is configured to perform a determination that an atypical x-ray dose is required for the x-ray examination of the patient, the determination comprising utilization of the patient thickness, and to output, using an output unit, an indication that an atypical x-ray dose is required for the x-ray examination of the patient. X-ray system.

14. 1. A method for determining an X-ray dose, said method comprising: providing a visible or infrared image of the patient to a processing unit; determining, by the processing unit, a silhouette of a patient in a visible or infrared image of the patient; performing, by the processing unit, a determination of patient thickness, comprising utilizing a silhouette of the patient in a visible or infrared image of the patient; determining, by the processing unit, an x-ray dose for an x-ray examination of the patient, the x-ray dose comprising using the patient's thickness; and / or determining, by the processing unit, that an atypical x-ray dose is required for an x-ray examination of the patient, the x-ray dose comprising using the patient's thickness; outputting, by an output unit, an indication of an x-ray dose for the x-ray examination of the patient, and / or outputting, by the output unit, an indication that an atypical x-ray dose is required for the x-ray examination of the patient. A method comprising:

15. 12. A computer program element for controlling an apparatus according to any one of claims 1 to 11, which, when executed by a processor, is configured to carry out the method of claim 14, or which, when executed by a processor, is configured to carry out the method of claim 14 for controlling a system according to claim 12, or which, when executed by a processor, is configured to carry out the method of claim 14 for controlling a system according to claim 13.

Images