Camera-based deep learning prediction and guidance for medical imaging protocols

Abstract

Systems or techniques are provided that facilitate camera-based deep learning prediction and guidance for medical imaging protocols. In various embodiments, a system may infer, via execution of a first deep learning neural network (e.g., 302), a prescribed imaging protocol (e.g., 304) to be performed on a medical patient (e.g., 106) by a medical imaging scanner (e.g., 104). In various aspects, the system may infer whether the medical patient is ready for the prescribed imaging protocol via performing a second deep learning neural network (e.g., 702) on a ready image or video (e.g., 204) of the medical patient captured by a camera (e.g., 108) associated with the medical imaging scanner. In various examples, the system may initiate an electronic guidance action (e.g., 1304) that interprets or shows how to prepare the medical patient for the prescribed imaging protocol in response to an inference (e.g., 704) that the medical patient is not preparing for the prescribed imaging protocol.

Description

Technical Field

[0001] This disclosure relates generally to medical imaging scanners, and more specifically to camera-based deep learning prediction and guidance for medical imaging protocols. Background Technology

[0002] Medical imaging scanners can perform a wide variety of imaging protocols on medical patients. When combined with the realities of clinical time constraints and the wide range of varying levels of experience among technicians, this diversity of protocols can lead to unacceptably high scan error rates.

[0003] Therefore, a system or technology that can solve one or more of these technical problems may be desirable. Summary of the Invention

[0004] The following summary is provided to offer a basic understanding of one or more embodiments. This summary is not intended to identify key or essential elements, nor is it intended to depict any scope of specific embodiments or any scope of the claims. Its sole purpose is to present the concepts in a simplified form as a prelude to the more detailed description that follows. In one or more embodiments described herein, devices, systems, computer-implemented methods, apparatuses, or computer program products are described that facilitate camera-based deep learning prediction and guidance for medical imaging protocols.

[0005] According to one or more embodiments, a system is provided. The system may include a non-transitory computer-readable storage memory that stores computer-executable components. The system may also include a processor operatively coupled to and capable of executing the computer-executable components stored in the non-transitory computer-readable storage memory. In various embodiments, the computer-executable components may include a protocol component that can infer a prescribed imaging protocol to be performed on a medical patient by a medical imaging scanner via the execution of a first deep learning neural network. In various aspects, the computer-executable components may include a preparation component that can infer whether the medical patient is ready for the prescribed imaging protocol via the execution of a second deep learning neural network on a preparation image or video of the medical patient captured by a camera associated with the medical imaging scanner. In various instances, the computer-executable components may include a guidance component that can initiate an electronic guidance action in response to an inference that the medical patient is not ready for the prescribed imaging protocol, the electronic guidance action explaining or showing how to prepare the medical patient for the prescribed imaging protocol.

[0006] According to one or more embodiments, a computer-implemented method is provided. In various embodiments, the computer-implemented method may include: inferring a predetermined imaging protocol to be performed on a medical patient by a device operatively coupled to a processor via the execution of a first deep learning neural network. In various aspects, the computer-implemented method may include: inferring whether the medical patient is ready for the predetermined imaging protocol by the device via the execution of a second deep learning neural network on prepared images or videos of the medical patient captured by a camera associated with the medical imaging scanner. In various instances, the computer-implemented method may include: initiating an electronic guidance action by the device in response to an inference that the medical patient is not ready for the predetermined imaging protocol, the electronic guidance action explaining or indicating how to prepare the medical patient for the predetermined imaging protocol.

[0007] According to one or more embodiments, a computer program product is provided for facilitating camera-based deep learning prediction and guidance for medical imaging protocols. In various embodiments, the computer program product may include a non-transitory computer-readable storage device having program instructions embodied therein. In various aspects, the program instructions are executable by a processor to cause the processor to infer a prescribed imaging protocol to be performed on the medical patient by executing a first deep learning neural network on a physician's prescription corresponding to the medical patient or a video feed depicting the medical patient. In various aspects, the program instructions are further executable to cause the processor to infer whether the MRI coil position on the medical patient fails to match a necessary MRI coil position specified in the prescribed imaging protocol by executing a second deep learning neural network on the video feed. In various instances, the program instructions are further executable to cause the processor to irradiate the body of the medical patient with an actuable light or laser associated with the MRI scanner in response to the inference that the MRI coil position does not match the necessary MRI coil position, thereby visibly illuminating the necessary MRI coil position on the medical patient. Attached Figure Description

[0008] Figure 1 A block diagram of an example non-limiting system according to one or more embodiments described herein is illustrated, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols.

[0009] Figure 2 A block diagram of an example non-limiting system comprising prescription documents and prepared images or videos according to one or more embodiments described herein is illustrated, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols.

[0010] Figure 3A block diagram of an example non-limiting system comprising a first deep learning neural network and a prescribed imaging protocol according to one or more embodiments described herein is illustrated, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols.

[0011] Figures 4 to 5 Example non-limiting block diagrams illustrating how a prescribed imaging protocol can be determined from a prescription document, according to one or more embodiments described herein.

[0012] Figure 6 Example non-limiting block diagrams illustrating a specified imaging protocol according to one or more embodiments described herein are shown.

[0013] Figure 7 A block diagram illustrating an example non-limiting system comprising a second deep learning neural network and a pre-determined exemplary system according to one or more embodiments described herein, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols.

[0014] Figures 8 to 12 Example non-limiting block diagrams illustrating how a ready determination can be obtained from an image or video are shown according to one or more embodiments described herein.

[0015] Figure 13 A block diagram of an example non-limiting system comprising one or more preparatory actions and one or more guiding actions according to one or more embodiments described herein is illustrated, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols.

[0016] Figure 14 A block diagram of an example non-limiting system excluding prescription documents, according to one or more embodiments described herein, is illustrated, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols.

[0017] Figure 15 Example non-limiting block diagrams illustrating how a prescribed imaging protocol can be determined based on an image or video are shown according to one or more embodiments described herein.

[0018] Figure 16 A flowchart illustrating an example non-limiting computer-implemented method according to one or more embodiments described herein, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols, is shown.

[0019] Figure 17 Example non-limiting block diagrams illustrating how various artificial intelligence models can be trained according to one or more embodiments described herein are presented.

[0020] Figure 18A flowchart illustrating an example non-limiting computer-implemented method according to one or more embodiments described herein, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols, is shown.

[0021] Figure 19 A block diagram illustrating an example non-limiting operating environment in which one or more embodiments described herein may be facilitated.

[0022] Figure 20 Example networking environments are illustrated that are operable to perform the various specific implementations described herein. Detailed Implementation

[0023] The following specific embodiments are merely illustrative and are not intended to limit the implementation or application / use of the embodiments. Furthermore, they are not intended to be construed as being bound by any express or implied information presented in the foregoing "Background Art" or "Summary of the Invention" or "Detailed Description" sections.

[0024] One or more embodiments will now be described with reference to the accompanying drawings, wherein the same reference numerals are always used to denote the same elements. In the following description, numerous specific details are set forth for purposes of explanation in order to provide a more thorough understanding of one or more embodiments. However, it will be apparent, in various cases, that one or more embodiments may be practiced without these specific details.

[0025] Medical imaging scanners (e.g., computed tomography (CT) scanners) can perform a wide variety of imaging protocols (e.g., imaging protocols defined by different configurations of scan parameters) on medical patients (e.g., humans, animals, or other medical patients). When combined with the reality of clinical time constraints and widely varying levels of technician experience, this protocol diversity can lead to unacceptably high scan error rates. In practice, it is often necessary for scanner operators or technicians in clinics or hospitals to scan many medical patients in a short period using different or corresponding imaging protocols. Due to this time pressure, the likelihood that scanner operators or technicians will select or utilize an imaging protocol for a given patient that does not match the protocol already prescribed by the consulting clinician for that given patient can increase. Additionally, many geographical locations lack experienced scanner operators or technicians and therefore rely more heavily on those who are less experienced. Due to the immense operational complexity of medical imaging scanners, this lack of experience can further exacerbate the possibility of erroneous or misused non-prescribed imaging protocols.

[0026] Therefore, a system or technology that can solve one or more of these technical problems may be desirable.

[0027] The various embodiments described herein address one or more of these technical problems. One or more embodiments described herein may include systems, computer-implemented methods, apparatus, or computer program products that facilitate camera-based deep learning prediction and guidance for medical imaging protocols. In other words, the various embodiments described herein may utilize the image analysis capabilities of deep learning to reduce or eliminate protocol selection errors in medical imaging scanners. Specifically, the various embodiments described herein may utilize a first deep learning model to infer or predict what imaging protocol has been prescribed for a given medical patient. In some instances, such inference may be based on text data typed or written by the attending or consulting clinician of the given medical patient. It should be noted that prescribing an imaging protocol may specify not only a particular configuration of the operating parameters of the medical imaging scanner that should be set, but also a specific body posture of the given medical patient or a specific body position of the scanner coil on the given medical patient (e.g., based on anatomy or laterality) that should be used to correctly perform the prescribed imaging protocol. In various aspects, the various embodiments described herein may involve utilizing a second deep learning neural network to infer or predict whether a given medical patient is currently or at present ready for a prescribed imaging protocol. In various scenarios, such inference may be based on a live camera feed depicting a given medical patient on or within the gantry, table, or gantry of a medical imaging scanner. More specifically, a second deep learning model may infer whether the actual body posture or the actual scanner coil position on the body of the given medical patient matches the body posture or scanner coil position specified in a prescribed imaging protocol. If so, the various embodiments described herein may involve automating or automatically prompting the operator or technician to authorize the performance of the prescribed imaging protocol on the given medical patient. If not, the various embodiments described herein may alternatively involve displaying body posture guidance or scanner coil position guidance to the operator or technician (e.g., to show or explain which body posture or scanner coil position is required by the prescribed imaging protocol). Such guidance may take any suitable form, such as on-screen instructions or diagrams, or such as a beam of light or laser directed onto the body of the given medical patient. Thus, such embodiments may be considered to automatically assist the operator or technician in correctly performing a scan of the medical patient, which may be desirable. In practice, such implementations can increase the scanning throughput of operators or technicians (e.g., with the assistance of the various implementations described herein, operators or technicians may not have to perform unnecessary rescans for any given patient and may therefore be able to scan more patients correctly in less time than is otherwise possible).Furthermore, because such implementations can automatically provide operators or technicians with body posture guidance or scanner coil position guidance in real time, they can reduce or eliminate the capture of poor, artifact-overlapping, or otherwise unusable scan images, which can help improve subsequent or downstream diagnosis or prognosis, and thus improve patient outcomes.

[0028] The various implementations described herein can be considered as computerized tools (e.g., any suitable combination of computer-executable hardware or computer-executable software) that can facilitate camera-based deep learning prediction and guidance for medical imaging protocols. In various aspects, such computerized tools may include access components, protocol components, preparation components, or guidance components.

[0029] In various implementations, a medical imaging scanner may be present. In various aspects, a medical imaging scanner may be any suitable medical modality, apparatus, or device capable of capturing or generating medical scan images (e.g., CT scan images, X-ray scan images, magnetic resonance imaging (MRI) scan images) of any suitable medical patient.

[0030] In various implementations, a medical imaging scanner may be associated with a preparatory camera. In various instances, the preparatory camera can be any suitable type of camera capable of capturing any suitable visible spectrum image or video of the medical patient. In some cases, the preparatory camera may be in the same room as the medical imaging scanner (e.g., physically built into or integrated into the medical imaging scanner), allowing the preparatory camera to view the medical patient physically occupying the scanner's rack, worktable, or stand. In other cases, the preparatory camera may be in a different room from the medical imaging scanner (e.g., in an adjacent room where the medical patient puts on or removes medical gear in preparation for their upcoming scan). In any case, the preparatory camera may capture real-time or recent images or videos of the medical patient while they prepare for or await the start of their scan.

[0031] In various instances, it may be desirable to provide automated scanning assistance regarding medical imaging scanners and medical patients. As described in this article, computerized tools can provide such automated scanning assistance.

[0032] In various implementations, the access component of the computerized tool can electronically access a medical imaging scanner or pre-reservation camera. For example, the access component can electronically interact or communicate with the medical imaging scanner or pre-reservation camera (e.g., transmit electronic commands to or read electronic signals from the medical imaging scanner or pre-reservation camera). In any case, the access component can be considered a channel through which other components of the computerized tool can electronically interact with the medical imaging scanner or pre-reservation camera (e.g., manipulate, execute, activate, deactivate, or modify the medical imaging scanner or pre-reservation camera).

[0033] Furthermore, in various aspects, the access component can electronically access prescription documents or prepared images or videos. That is, the access component can electronically receive, retrieve, or otherwise acquire prescription documents or prepared images or videos, enabling other components of the computerized tool to electronically interact with the prescription documents or prepared images or videos (e.g., read, write, edit, copy, manipulate the prescription documents or prepared images or videos). In various instances, the prescription document can be any suitable natural language or plain text sentence or sentence fragment that substantially or semantically conveys clinical findings or observations about a medical patient, including requests, orders, or prescriptions from a medical patient scanned using a medical imaging scanner. It should be noted that the access component can obtain the prescription document from any suitable electronic source (e.g., the prescription document may be typed by or otherwise on behalf of the medical patient's consultant or attending physician; the prescription document may have been uploaded to a Radiological Information System (RIS) associated with the medical imaging scanner; and the access component can retrieve the prescription document from the RIS). In various cases, a preparatory image or video can be one or more pixels or voxel arrays that visually depict or exemplify the medical patient as they prepare for or await the start of an upcoming scan. Therefore, a preparatory image or video can be considered any visual data captured or recorded by the preparatory camera relative to the medical patient. It should be noted that the access component can obtain the preparatory image or video from the preparatory camera.

[0034] In various implementations, the protocol components may electronically store, maintain, control, or otherwise access the first deep learning neural network. In various aspects, the first deep learning neural network may have any suitable deep learning internal architecture. For example, the first deep learning neural network may include any suitable number of layers of any suitable type (e.g., input layers, one or more hidden layers, output layers, any of which may be convolutional layers, dense layers, long short-term memory (LSTM) layers, transformer layers, nonlinear layers, pooling layers, batch normalization layers, or padding layers). As another example, the first deep learning neural network may include any suitable number of neurons in various layers (e.g., different layers may have the same or different numbers of neurons). As yet another example, the first deep learning neural network may include any suitable activation function (e.g., softmax, sigmoid, hyperbolic tangent, corrected linear unit) in various neurons (e.g., different neurons may have the same or different activation functions). As yet another example, the first deep learning neural network may include any suitable inter-neuron or inter-layer connections (e.g., forward connections, skip connections, recursive connections).

[0035] Regardless of its internal architecture, the first deep learning neural network can be configured to receive any suitable text as input and determine, as output, an imaging protocol specified, recorded, invoked, or otherwise defined by such input text. Therefore, the protocol component 116 can electronically execute the first deep learning neural network on the prescription document, thereby enabling the first deep learning neural network to determine which specific imaging protocol the prescription document specifies.

[0036] In some cases, the first deep learning neural network can be constructed as a text classifier. Therefore, the protocol component can electronically execute the first deep learning neural network on a prescription document, and such execution can produce a protocol classification label. More specifically, the protocol component can feed the prescription document into the input layer of the first deep learning neural network, which can perform a forward pass through one or more hidden layers of the first deep learning neural network, and the output layer of the first deep learning neural network can compute a protocol classification label based on the activations provided by one or more hidden layers of the first deep learning neural network. In various aspects, the protocol classification label can indicate the imaging protocol of the medical imaging scanner, which the first deep learning neural network considers or infers to be requested or invoked by the prescription document. In particular, there may be multiple defined imaging protocols that the medical imaging scanner may implement (e.g., different protocols may specify different radiation levels or different gantry speeds for scanning different body parts), and the protocol classification label can indicate which of the multiple defined imaging protocols the prescription document (in the case of the first deep learning neural network) records or specifies for the medical patient.

[0037] In other cases, the first deep learning neural network may alternatively be constructed as a large language model (e.g., ChatGPT). In such cases, the protocol component may execute the first deep learning neural network on the prescription document and the protocol identification prompt, and such execution may produce a protocol indication. More specifically, the protocol identification prompt may be unstructured text or plain text that queries or commands the identification, description, or interpretation of any imaging protocol indicated or specified in the prescription document. In various aspects, the protocol component may cascade the prescription document with the protocol identification prompt. In various instances, the protocol component may feed this cascade into the input layer of the first deep learning neural network, which may perform forward propagation through one or more hidden layers of the first deep learning neural network, and the output layer of the first deep learning neural network may compute the protocol indication based on the activations provided by one or more hidden layers of the first deep learning neural network. In various cases, the protocol indication may be synthetic text based on the prescription document and substantially or semantically respond to the protocol identification prompt. In other words, the protocol indication can be unstructured text or plain text that names, states, describes, or interprets any particular imaging protocol requested or invoked by the prescription document (in relation to the first deep learning neural network).

[0038] In any case, the protocol component may utilize a first deep learning neural network to identify a specific imaging protocol prescribed by the prescription document for the medical patient. This imaging protocol may be referred to as a prescribed imaging protocol. In various aspects, the prescribed imaging protocol may be associated with or otherwise defined by a necessary scan parameter configuration. In other words, the medical imaging scanner may have any suitable scan parameters (e.g., radiation level, gantry speed, field of view, matrix size), and the necessary scan parameter configuration may be any specific combination of values ​​or states that those scan parameters should be set to facilitate the execution or completion of the prescribed imaging protocol. However, in addition to the necessary scan parameter configuration, the prescribed imaging protocol may be associated with or otherwise defined by a necessary body posture or necessary scanner coil position. In practice, in various instances, the prescribed imaging protocol may be intended or designed for use with a patient whose body is physically oriented relative to the medical imaging scanner in a particular manner (e.g., prone, supine, lateral, head forward, feet forward), and the necessary body posture may be or may refer to that particular orientation. Similarly, in various cases, the specified imaging protocol may be intended or designed for use with patients who wear scanner coils at specific anatomical locations (e.g., a coil on the left leg, a coil on the right arm, a coil on the head), and the necessary scanner coil location may be or may refer to that specific anatomical location.

[0039] In various implementations, the preparation component may electronically store, maintain, control, or otherwise access the second deep learning neural network. In various aspects, the second deep learning neural network may have any suitable deep learning internal architecture. For example, the second deep learning neural network may include any suitable number of layers of any suitable type (e.g., input layer, one or more hidden layers, output layer, any of which may be a convolutional layer, a dense layer, an LSTM layer, a transformer layer, a nonlinear layer, a pooling layer, a batch normalization layer, or a padding layer). As another example, the second deep learning neural network may include any suitable number of neurons in various layers (e.g., different layers may have the same or different numbers of neurons). As yet another example, the second deep learning neural network may include any suitable activation function (e.g., softmax, sigmoid, hyperbolic tangent, corrected linear unit) in various neurons (e.g., different neurons may have the same or different activation functions). As yet another example, the second deep learning neural network may include any suitable inter-neuron or inter-layer connections (e.g., forward connections, skip connections, recursive connections).

[0040] Regardless of its internal architecture, the second deep learning neural network can be configured as a computer vision model. That is, the second deep learning neural network can be configured to receive any suitable image or video data as input and locate certain objects of interest within such input image or video data as output. In some aspects, such objects of interest may include various body parts (e.g., head, right eye, left eye, right arm, left arm). In some instances, such objects of interest may include wearable scanner coils. Therefore, the preparation unit 118 can electronically execute the second deep learning neural network on a prepared image or video, causing the second deep learning neural network to generate a set of body part localizations or scanner coil localizations. More specifically, the preparation unit can feed the prepared image or video into the input layer of the second deep learning neural network, which can perform forward propagation through one or more hidden layers of the second deep learning neural network, and the output layer of the second deep learning neural network can compute the set of body part localizations or scanner coil localizations based on the activations provided by one or more hidden layers of the second deep learning neural network.

[0041] In various aspects, this set of body part localizations may include any suitable number of localizations corresponding to any suitable number of different body parts of the medical patient. In various instances, each body part localization may be any suitable electronic data (in the context of the second deep learning neural network) indicating the in-image or in-video location of the corresponding body part of the medical patient (e.g., it may be the landmark coordinates of the corresponding body part; it may be the bounding box surrounding the corresponding body part; it may be the segmentation mask covering the corresponding body part). Similarly, scanner coil localization may be any suitable electronic data (e.g., in the context of the second deep learning neural network) indicating the in-image or in-video location of the scanner coil worn by the medical patient (e.g., it may be the landmark coordinates of the scanner coil; it may be the bounding box surrounding the scanner coil; it may be the segmentation mask covering the scanner coil).

[0042] In various aspects, the preparation unit can electronically determine whether a medical patient is correctly prepared for a prescribed imaging protocol by comparing the following: the set of body part positioning or scanner coil positioning; and the necessary body posture or necessary scanner coil position. In practice, when the preparation camera is viewing or aiming at the gantry, table, or stand of a medical imaging scanner, the set of body part positioning can be considered to collectively indicate or convey the patient's current or present body posture (e.g., head above feet indicates a head-forward posture; head below feet indicates a feet-forward posture). Furthermore, the set of body part positioning and scanner coil positioning together can be considered to collectively indicate the patient's current or present scanner coil position (e.g., scanner coil positioning consistent with right arm positioning indicates that the scanner coil is worn on the patient's right arm). If the current or present body posture does not match the necessary body posture, the preparation unit can determine or ascertain that the medical patient is not prepared for the prescribed imaging protocol. Similarly, if the current or present scanner coil position does not match the required scanner coil position, the preparation component can determine or identify that the medical patient is not ready for the prescribed imaging protocol. However, if the current or present body posture matches the required body posture, and if the current or present scanner coil position matches the required scanner coil position, the preparation component can alternatively determine or identify that the medical patient is ready for the prescribed imaging protocol.

[0043] In various implementations, the guidance component may initiate or perform any suitable electronic action based on the determination or assertion of the preparation component. For example, if the preparation component determines or asserts that the medical patient is ready for a prescribed imaging protocol, the guidance component may initiate or perform any suitable preparation action. Such preparation actions may include: instructing or commanding the medical imaging scanner to begin the prescribed imaging protocol; or prompting the operator of the medical imaging scanner for permission to begin the prescribed imaging protocol via the medical imaging scanner's graphical user interface (GUI). On the other hand, if the preparation component determines or asserts that the medical patient is not yet ready for the prescribed imaging protocol, the guidance component may initiate or perform any suitable guidance action. If the preparation component determines that the medical patient is not ready due to failure to achieve the necessary body posture, such guidance actions may include presenting a notification on the GUI instructing the medical patient not to be ready until the necessary body posture is achieved. Similarly, if the preparation component determines that the medical patient is not ready due to failure to achieve the necessary scanner coil position, such guidance actions may include presenting a notification on the GUI instructing the medical patient not to be ready until the necessary scanner coil position is achieved. In some aspects, there may be actuable or movable light or laser associated with the medical imaging scanner, wherein such light or laser can be directed or aimed at the medical patient's body in a controllable manner. In such cases, if the preparation components determine that the medical patient is not ready due to failure to meet the necessary scanner coil position, the guidance action may include directing the light or laser to illuminate the medical patient's body with a visible beam such that the visible beam falls on or otherwise visually indicates the necessary scanner coil position (e.g., falls on or visually indicates any specific part of the medical patient's body where the scanner coil should be moved). Thus, if the medical patient is not yet ready for a prescribed imaging protocol, the guidance components may be considered (e.g., by indicating how the patient's body posture or scanner coil position should be changed) to help or assist the operator of the medical imaging scanner to quickly or efficiently prepare the medical patient.

[0044] It should be noted that, in order to correctly or accurately identify the prescribed imaging protocol, or to correctly or accurately determine the preparation, the machine learning models described herein (e.g., a first deep learning neural network and a second deep learning neural network) should first undergo training. In various cases, computerized tools can train such machine learning models using any suitable training paradigm (e.g., via supervised training, unsupervised training, or reinforcement learning).

[0045] Various implementation schemes described herein can be employed to solve inherently highly technical problems (e.g., to facilitate camera-based deep learning predictions and guidance for medical imaging protocols) using hardware or software. These problems are not abstract and cannot be performed by humans as a set of mental behaviors. Furthermore, some processes in the execution can be performed by a dedicated computer (e.g., a text classifier, LLM, computer vision model) to implement the defined actions associated with the medical imaging scanner.

[0046] For example, such defined actions may include: inferring a prescribed imaging protocol to be performed on a medical patient by a device operatively coupled to a processor via the execution of a first deep learning neural network (e.g., a text classifier or LLM performed on a prescription document); inferring whether the medical patient is ready for the prescribed imaging protocol by the device via the execution of a second deep learning neural network (e.g., a computer vision model) on a prepared image or video of the medical patient captured by a camera associated with the medical imaging scanner; and initiating an electronic guidance action by the device in response to the inference that the medical patient is not ready for the prescribed imaging protocol, the electronic guidance action explaining or showing how to prepare the medical patient for the prescribed imaging protocol. In various instances, such defined actions may also include: the device presenting a notification on the graphical user interface of the medical imaging scanner in response to the inference that the medical patient is ready for the prescribed imaging protocol, the notification indicating that the prescribed imaging protocol is ready to be performed and requesting the user of the medical imaging scanner to approve the performance of the prescribed imaging protocol; or the device instructing the medical imaging scanner to perform the prescribed imaging protocol. In various scenarios, the second deep learning neural network may receive a preparation image or video as input and produce a location indicating the current body posture or orientation of the medical patient as output. The device may infer that the medical patient is not ready in response to a mismatch between the current body posture or orientation and a necessary body posture or orientation specified in a prescribed imaging protocol. In various instances, the second deep learning neural network may receive a preparation image or video as input and produce a location indicating the current scanner coil position on the medical patient as output. The device may infer that the medical patient is not ready in response to a mismatch between the current scanner coil position and a necessary scanner coil position specified in a prescribed imaging protocol. In various instances, the current scanner coil position of the medical patient inferred by the second deep learning neural network may fail to match the necessary scanner coil position specified in the prescribed imaging protocol, and electronic guidance actions may include irradiating the medical patient's body with light or laser associated with the medical imaging scanner based on the necessary scanner coil position.

[0047] Such defined actions are not performed manually by humans. In fact, neither the human mind nor a person with pen and paper can: electronically perform a text classifier or LLM on a prescription document to identify the imaging protocol to be performed on a given patient by a medical imaging scanner; electronically capture live or real-time images or videos depicting a given patient preparing for or awaiting the start of an imaging protocol; electronically perform computer vision modeling on the images or videos to generate body part localizations or wearable scanner coil localizations for the medical patient; electronically compare those localizations with necessary body poses or necessary scanner coil positions specified by or otherwise associated with the imaging protocol to determine whether the given patient is ready for the imaging protocol; and electronically provide body pose or scanner coil guidance (e.g., via a GUI display or via real-world light or laser beams) in response to determining that the given patient is not yet ready for the imaging protocol. In reality, medical imaging scanners (e.g., MRI scanners, CT scanners, X-ray scanners) are inherently computerized, hardware-based constructions that cannot be meaningfully performed by the human mind in any way without a computer. Additionally, deep learning neural networks (e.g., text classifiers, LLMs, computer vision models) are inherently computerized, software-based constructs that cannot be meaningfully trained or executed by the human mind in any way without a computer. Therefore, computerized tools that utilize the text or image analysis capabilities of deep learning neural networks to automatically identify medical imaging protocols to be applied to patients, automatically check whether the patient's body posture or wearable scanner coil position is consistent with the body posture or scanner coil position required by the medical imaging protocol, and correct any inconsistencies with on-screen instructions or any markings on the body beam are also inherently computerized and cannot be implemented in any compliant, practical, or reasonable manner without a computer.

[0048] Furthermore, the various embodiments described herein can integrate various teachings related to the field of medical imaging scanners into practical applications. As mentioned above, medical imaging scanners can perform countless possible imaging protocols. Also as mentioned above, different imaging protocols can be prescribed for different medical patients. Due to time constraints, long working hours, and widely varying levels of experience, the likelihood of medical imaging technicians performing medical imaging scans incorrectly or erroneously can increase (e.g., they may inadvertently select or load the wrong, incorrect, or unprescribed protocol for a given patient; they may select or load the correct or prescribed protocol for a given patient, but may inadvertently fail to ensure that the patient's body posture or the position of the wearable scanner coil is consistent with the correct or prescribed protocol).

[0049] The various embodiments described herein address one or more of these technical problems. Specifically, given a prescription document (e.g., clinical notes written by a consulting physician and stored in a clinical RIS database) for a particular patient about to undergo a medical scan, the various embodiments described herein can perform a first deep learning neural network (e.g., a text classifier, LLM) on the prescription document to identify a specific imaging protocol prescribed for that particular patient. Furthermore, the various embodiments described herein can capture live images or videos of a particular patient as they prepare for or otherwise await their scan (e.g., in or on a rack, workbench, or stand of a medical imaging scanner; in a donning or dismounting room associated with the medical imaging scanner). In various aspects, the various embodiments described herein can perform a second deep learning neural network (e.g., a computer vision model) on the live images or videos to identify the particular patient's current body posture or current wearable scanner coil position. In various instances, the various embodiments described herein can determine whether the current body posture or current wearable scanner coil position is consistent with those body postures or wearable scanner coil positions listed or specified according to the requirements or needs of a particular imaging protocol. If they are consistent, various implementations can automatically initiate a specific imaging protocol or display a GUI message to the operator requesting permission to begin the specific imaging protocol. On the other hand, if they are inconsistent, the various implementations described herein can provide the operator with real-time guidance (e.g., GUI messages instructing that the body posture of a specific patient or the position of a wearable scanner coil must be corrected; shining visible light onto the body of a specific patient to indicate where the wearable scanner coil should be positioned on the patient). Such implementations can be considered to automatically help or assist the operator to reduce or avoid erroneous or defective scans, thereby saving the operator's time and effort, thus saving the patient's time and effort (e.g., eliminating or reducing the need for repeat scans), and thus improving clinical outcomes (e.g., eliminating or reducing the generation of inaccurate, defective, or artifact-laden medical images). Therefore, the various implementations described herein can be considered as ingenious or creative techniques or pipelines utilizing camera-based deep learning to assist in performing medical imaging scans on medical patients. Therefore, the various implementations described herein certainly constitute tangible and concrete technical improvements or advantages of medical imaging scanners. Thus, such implementations clearly qualify as useful and practical applications of computers.

[0050] Furthermore, the various embodiments described herein can control real-world physical devices based on the disclosed teachings. For example, the various embodiments described herein can direct or cause a real-world medical imaging scanner (e.g., a CT scanner, an MRI scanner) to perform a real-world scan on a real-world patient. Additionally, the various embodiments described herein can cause real-world lighting equipment to illuminate the body of such a real-world patient with visible light beams or visible laser beams.

[0051] It should be understood that the accompanying figures and descriptions provide non-limiting examples of various embodiments and are not necessarily drawn to scale.

[0052] Figure 1 A block diagram of an example non-limiting system 100 according to one or more embodiments described herein is illustrated, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols. As shown, the protocol guidance system 102 may be electronically integrated with a medical imaging scanner 104 or a pre-equipped camera 108.

[0053] In various embodiments, the medical imaging scanner 104 can be any suitable medical image capture modality or apparatus capable of capturing or otherwise generating (e.g., via X-ray emission or electromagnetic fields) medical images. As a non-limiting example, the medical imaging scanner 104 can be an X-ray scanner configured to capture or generate X-ray scan images of any suitable anatomical structure (e.g., organ, tissue, body cavity, body fluid) of the medical patient 106. As another non-limiting example, the medical imaging scanner 104 can be a CT scanner configured to capture or generate CT scan images of any suitable anatomical structure of the medical patient 106. As yet another non-limiting example, the medical imaging scanner 104 can be a positron emission tomography (PET) scanner configured to capture or generate PET scan images of any suitable anatomical structure of the medical patient 106. As yet another non-limiting example, the medical imaging scanner 104 can be a nuclear medicine (NM) scanner configured to capture or generate NM scan images of any suitable anatomical structure of the medical patient 106. As yet another non-limiting example, the medical imaging scanner 104 could be an ultrasound scanner configured to capture or generate ultrasound scan images of any suitable anatomical structures of the medical patient 106. As yet another non-limiting example, the medical imaging scanner 104 could be an MRI scanner configured to capture or generate MRI scan images of any suitable anatomical structures of the medical patient 106.

[0054] Although not explicitly shown in the figures, the medical imaging scanner 104 can be electronically integrated with any suitable human-machine interface device, either remotely or locally. Thus, an operator or user associated with the medical imaging scanner 104 can interact with or otherwise control it. Some non-limiting examples of the human-machine interface device could be a keyboard, a keypad, a touchscreen, or a voice command system for the medical imaging scanner 104.

[0055] Although not explicitly shown in the figures, the medical imaging scanner 104 may have any suitable number and type of scanning parameters or otherwise be associated with any suitable number and type of scanning parameters. In various aspects, the scanning parameters may be any suitable configurable settings of the medical imaging scanner 104 that can be selectively controlled by a user or operator to proportionately control how the medical imaging scanner 104 operates, runs, or otherwise performs scanning. As some non-limiting examples, the scanning parameters may be any of the following: sequence type parameters of the medical imaging scanner 104 (e.g., possible or selectable values ​​or states of such parameters may include T1-weighted sequence type, T2-weighted sequence type, proton density sequence type, diffusion-weighted sequence type, or fluid attenuation inversion recovery sequence type); slice thickness parameters of the medical imaging scanner 104 (e.g., possible or selectable values ​​or states of such parameters may include a slice thickness of 1 mm, 3 mm, or 10 mm); slice orientation parameters of the medical imaging scanner 104 (e.g., possible or selectable values ​​or states of such parameters may include axial slice orientation, coronal slice orientation, or sagittal slice orientation); field of view (FOV) parameters of the medical imaging scanner 104 (e.g., possible or selectable values ​​or states of such parameters may include 100 × 100 mm). 2 FOV, 150×150mm 2 FOV, 200×200mm 2 FOV or 400×400mm 2The FOV of the medical imaging scanner 104; the matrix size parameter of the medical imaging scanner 104 (e.g., possible or selectable values ​​or states of this parameter may include a matrix size of 128 pixels by 128 pixels, a matrix size of 256 pixels by 256 pixels, or a matrix size of 512 pixels by 512 pixels); the repetition time (TR) parameter of the medical imaging scanner 104 (e.g., possible or selectable values ​​or states of this parameter may include a TR of 2 milliseconds (ms), a TR of 500 ms, a TR of 2000 ms, or a TR of 5000 ms); the medical imaging scanner 104 The echo time (TE) parameter of 4 (e.g., possible or selectable values ​​or states of this parameter may include a TE of 2 milliseconds (ms), a TE of 20 ms, a TE of 80 ms, or a TE of 200 ms); or the number of excitations (NEX) parameter of the medical imaging scanner 104 (e.g., possible or selectable values ​​or states of this parameter may include 1 NEX (meaning that the resulting image is formed by a single scan excitation), 2 NEX (meaning that the resulting image is the average of two scan excitations), or 3 NEX (meaning that the resulting image is the average of three scan excitations).

[0056] In various aspects, scanner coil 122 may be associated with medical imaging scanner 104. In various instances, scanner coil 122 may be any suitable wearable device or apparatus that assists or otherwise aids medical imaging scanner 104 in transmitting electromagnetic radiation to or through the body of medical patient 106 in any suitable manner. In some cases, scanner coil 122 may alternatively be referred to as a surface coil. In any case, scanner coil 122 may be physically worn by medical patient 106. In other words, scanner coil 122 may be physically located, physically wound around, or otherwise physically surrounded or encircled, wholly or partially, some external body part of medical patient 106. As some non-limiting examples, the scanner coil 122 may be worn on: the head of the medical patient 106; the right or left shoulder of the medical patient 106; the right or left upper arm of the medical patient 106; the right or left elbow of the medical patient 106; the right or left forearm of the medical patient 106; the right wrist or right hand or left wrist or left hand of the medical patient 106; the torso of the medical patient 106; the right or left hip of the medical patient 106; the right or left thigh of the medical patient 106; the right or left knee of the medical patient 106; the right or left calf of the medical patient 106; or the right ankle or right foot or left ankle or left foot of the medical patient 106.

[0057] In various embodiments, the pre-position camera 108 can be any suitable image capture device that can view the medical patient 106 while the medical patient 106 is preparing, ready, or otherwise waiting for the medical imaging scanner 104 to scan the medical patient. In various aspects, the pre-position camera 108 can be physically integrated or otherwise built into the medical imaging scanner 104. Thus, the pre-position camera 108 can be able to view the medical patient 106 while the medical patient 106 is wearing the scanner coil 122 and is sitting on, in, lying on, or physically occupying the actuated table, rack, or imaging stage of the medical imaging scanner 104. However, this is only a non-limiting example. In other aspects, the pre-position camera 108 can be physically located away from or separate from the medical imaging scanner 104, but can still be in the same room as the medical imaging scanner 104. In such cases, even though the standby camera 108 is physically located away from or separate from the medical imaging scanner 104, the standby camera can still view the medical patient 106 wearing the scanner coil 122 and sitting, lying, or otherwise occupying the workbench, rack, or imaging platform of the medical imaging scanner 104. In other respects, the standby camera 108 may be located in a separate or different room from the medical imaging scanner 104. In practice, a high-volume hospital may have a first room housing the medical imaging scanner 104 and an adjacent or otherwise nearby second room in which patients queuing for scanning wear the scanner coil 122. In such cases, the backup camera 108 may be located in a second room instead of the first room, such that the backup camera 108 cannot view the medical patient 106 sitting on or in the workbench, rack, or stand of the medical imaging scanner 104, lying on or in the workbench, rack, or stand, or otherwise occupying the workbench, rack, or stand, but the backup camera 108 can still view the medical patient 106 wearing the scanner coil 122.

[0058] It should be understood that the pre-camera 108 may have any suitable architecture or construction. For example, the pre-camera 108 may include any suitable type of optical lens, any suitable type of shutter, or any suitable type of photoelectric detection mechanism, or be composed of them in any other way. In some aspects, the pre-camera 108 may capture images or videos of the medical patient 106 in the visible spectrum. In other aspects, the pre-camera 108 may capture images or videos of the medical patient 106 in any suitable form of invisible spectrum (such as infrared images or videos, or such as thermal images or videos).

[0059] In various situations, it may be desirable for the operator or user of the medical imaging scanner 104 to perform a scan on a medical patient 106. As described herein, the protocol guidance system 102 can facilitate such automated assistance.

[0060] In various embodiments, the protocol guidance system 102 may include a processor 110 (e.g., a computer processing unit, microprocessor) and a non-transitory computer-readable storage device 112 operatively or communicatively connected to or coupled to the processor 110. The non-transitory computer-readable storage device 112 may store computer-executable instructions that, when executed by the processor 110, cause the processor 110 or other components of the protocol guidance system 102 (e.g., access component 114, protocol component 116, preparation component 118, guidance component 120) to perform one or more actions. In various embodiments, the non-transitory computer-readable storage device 112 may store computer-executable components (e.g., access component 114, protocol component 116, preparation component 118, guidance component 120), and the processor 110 may execute these computer-executable components.

[0061] In various embodiments, the protocol guidance system 102 may include an access component 114. In various aspects, the access component 114 may electronically access the medical imaging scanner 104 or the pre-reservation camera 108 in any suitable manner or otherwise electronically communicate with the medical imaging scanner or the pre-reservation camera. For example, the access component 114 may electronically send any suitable electronic data to or receive any suitable electronic data from the medical imaging scanner 104 or the pre-reservation camera 108. Therefore, the access component 114 can be considered a proxy or channel through which other components of the protocol guidance system 102 may electronically interact with the medical imaging scanner 104 or the pre-reservation camera 108.

[0062] In various aspects, as described herein, access component 114 can also electronically access prescription documents and prepared images or videos associated with medical patient 106.

[0063] In various implementations, the protocol guidance system 102 may include a protocol component 116. In various aspects, as described herein, the protocol component 116 may electronically identify the prescribed imaging protocol to be performed on the medical patient 106 via performing a first deep learning neural network on a prescription document.

[0064] In various implementations, the protocol guidance system 102 may include a preparation component 118. In various instances, as described herein, the preparation component 118 may electronically determine whether the medical patient 106 is properly prepared to perform the prescribed imaging protocol by performing a second deep learning neural network on a preparation image or video.

[0065] In various implementations, the protocol guidance system 102 may include a guidance component 120. In various cases, as described herein, the guidance component 120 may electronically initiate any appropriate action based on the readiness determination of the readiness component 118 (e.g., initiating a prescribed imaging protocol if the medical patient 106 is ready; or providing electronic calibration guidance to the user or operator of the medical imaging scanner 104 if the medical patient 106 is not yet ready).

[0066] It should be noted that in various instances, access component 114, protocol component 116, preparation component 118, and guidance component 120 can be collectively considered as one or more software components 113 of the protocol guidance system 102. In all aspects, it should be understood that, for ease of explanation and illustration, one or more software components 113 are generally described herein as comprising four components (e.g., access component 114, protocol component 116, preparation component 118, and guidance component 120). However, one or more software components 113 are not limited to being implemented as exactly four components in every embodiment. In fact, in some embodiments, the functionality of these four components described herein can be combined in any suitable manner to be implemented in fewer than four components or by fewer than four components (e.g., in some cases, a single component can perform all the functionality described herein with respect to access component 114, protocol component 116, preparation component 118, and guidance component 120). In other embodiments, the functionality of the four components described herein may alternatively be distributed, separated, split, or segmented in any suitable manner to be implemented in or by more than four components (e.g., two or more components may facilitate functionality that can be performed by access component 114; two or more components may facilitate functionality that can be performed by protocol component 116; two or more components may facilitate functionality that can be performed by preparation component 118; two or more components may facilitate functionality that can be performed by instruction component 120).

[0067] Figure 2A block diagram of an example non-limiting system 200, comprising a prescription document and a prepared image or video, is illustrated according to one or more embodiments described herein. This system facilitates camera-based deep learning prediction and guidance for medical imaging protocols. As shown, in some cases, system 200 may include the same components as system 100 and may also include a prescription document 202 and a prepared image or video 204.

[0068] In various embodiments, access component 114 may electronically access prescription document 202 from any suitable electronic source. As a non-limiting example, access component 114 may electronically receive, retrieve, or otherwise obtain prescription document 202 from a RIS database associated with medical imaging scanner 104 (e.g., the RIS database and medical imaging scanner 104 may belong to the same clinic or hospital). In any case, prescription document 202 may be, or may otherwise include, one or more unstructured or plaintext declarative sentences or sentence fragments that semantically interpret, describe, name, instruct, or otherwise convey any suitable imaging protocol that the consulting or attending physician or other medical professional has requested, invoked, or otherwise specified for medical patient 106 regarding medical imaging scanner 104. In various aspects, prescription document 202 may be electronically typed or otherwise created (e.g., via speech-to-text transcription) by or on behalf of a consulting or attending physical or other medical professional. In various instances, the prescription document 202 may utilize special or other non-standardized language or word choices to interpret, describe, name, instruct, or otherwise convey the imaging protocol of the medical patient 106. As a non-limiting example, there may be various different ways of specifying the scanning of the right knee of the medical patient 106 in textual form, such as “scan right knee,” such as “rt knee,” or such as “knee R.”

[0069] In various embodiments, access component 114 can electronically access the prepared image or video 204 from any suitable electronic source. As a non-limiting example, access component 114 can electronically receive, electronically retrieve, or otherwise electronically acquire the prepared image or video 204 from the preparation camera 108. However, this is merely a non-limiting example. In other aspects, the preparation camera 108 can enable the prepared image or video 204 to be stored or maintained in any suitable electronically accessible database, and access component 114 can electronically receive, electronically retrieve, or otherwise electronically acquire the prepared image or video 204 from that database. In any case, the prepared image or video 204 can be electronically captured or otherwise recorded by the preparation camera 108. Thus, the prepared image or video 204 can be one or more (e.g., time-series) x-by-y pixel arrays, for any suitable positive integers x and y, that visually exemplify or depict the medical patient 106 as the patient prepares or otherwise waits for their turn to be scanned by the medical imaging scanner 104. If the preparation camera 108 is located in the same room as the medical imaging scanner 104, the preparation image or video 204 can visually exemplify a medical patient 106 wearing the scanner coil 122 and sitting on or in the workbench, rack, or stand of the medical imaging scanner 104, lying on or in the workbench, rack, or stand, or otherwise occupying the workbench, rack, or stand. On the other hand, if the preparation camera 108 is instead located in a separate room from the medical imaging scanner 104, the preparation image or video 204 can visually exemplify a medical patient 106 wearing the scanner coil 122 but not sitting on or in the workbench, rack, or stand of the medical imaging scanner 104, lying on or in the workbench, rack, or stand, or otherwise occupying the workbench, rack, or stand.

[0070] Figure 3 A block diagram of an example non-limiting system 300, comprising a first deep learning neural network and a prescribed imaging protocol, is illustrated according to one or more embodiments described herein. This system facilitates camera-based deep learning prediction and guidance for medical imaging protocols. As shown, in some cases, system 300 may include the same components as system 200, and may also include a deep learning neural network 302 and a prescribed imaging protocol 304.

[0071] In various implementations, protocol component 116 may electronically store, electronically maintain, electronically control, or otherwise electronically access the deep learning neural network 302. In various aspects, as described herein, protocol component 116 may electronically utilize the deep learning neural network 302 to identify and specify the imaging protocol 304 based on the prescription document 202. Relative to Figures 4 to 6The non-restrictive aspects are described.

[0072] Figures 4 to 5 Example non-limiting block diagrams 400 and 500 illustrate how a prescribed imaging protocol 304 can be determined from a prescription document 202 according to one or more embodiments described herein.

[0073] First, consider Figure 4 In various implementations, as shown in the figure, the deep learning neural network 302 can be configured as a text classifier with any suitable deep learning internal architecture. For example, in various cases, the deep learning neural network 302 may have an input layer, one or more hidden layers, and an output layer. In various aspects, any of such layers can be coupled together by any suitable inter-neuronal or inter-layer connections (such as forward connections, skip connections, or recursive connections). Furthermore, in various cases, any of such layers can be any suitable type of neural network layer with any suitable learnable or trainable internal parameters. For example, any of such an input layer, one or more hidden layers, or output layer can be a convolutional layer, and the learnable or trainable parameters of the convolutional layer can be a convolutional kernel. As another example, any of such an input layer, one or more hidden layers, or output layer can be a dense layer, and the learnable or trainable parameters of the dense layer can be a weight matrix or a bias value. As yet another example, any of such an input layer, one or more hidden layers, or output layer can be a batch normalization layer, and the learnable or trainable parameters of the batch normalization layer can be a shift factor or a scaling factor. As yet another example, any of such an input layer, one or more hidden layers, or output layer can be an LSTM layer, whose learnable or trainable parameters can be an input state weight matrix or a hidden state weight matrix. As yet another example, any of such an input layer, one or more hidden layers, or output layer can be a transformer layer, whose learnable or trainable parameters can be a single-head or multi-head attention block or other weight matrix. Furthermore, in various cases, any of such layers can be any suitable type of neural network layer with any suitable fixed or untrainable internal parameters. For example, any of such an input layer, one or more hidden layers, or output layer can be a nonlinear layer, a padding layer, a pooling layer, or a cascaded layer.

[0074] Regardless of the specific internal architecture implemented in the deep learning neural network 302 (e.g., the specific number, type, or organization of layers), the deep learning neural network 302 can be configured to receive text data as input and produce classification labels for the input text data as output. Therefore, the protocol component 116 can electronically execute the deep learning neural network 302 on the prescription document 202, and such execution can cause the deep learning neural network 302 to produce protocol classification labels 402. More specifically, the protocol component 116 can feed the prescription document 202 into the input layer of the deep learning neural network 302. In various cases, the prescription document 202 can complete the forward pass through one or more hidden layers of the deep learning neural network 302. In various aspects, the output layer of the deep learning neural network 302 can compute or calculate the protocol classification labels 402 based on the activation maps or feature maps generated by one or more hidden layers of the deep learning neural network 302.

[0075] In various respects, the protocol classification label 402 can be any suitable electronic data (e.g., it can be one or more scalars, one or more vectors, one or more matrices, one or more tensors, one or more strings, or any suitable combination thereof) that represents, conveys, or otherwise indicates (in the case of the deep learning neural network 302) a particular imaging protocol recorded in or otherwise invoked in the prescription document 202. In particular, there can be multiple defined imaging protocols 404. In various instances, the multiple defined imaging protocols 404 may include n protocols, for any suitable positive integer n>1: defined imaging protocol 404(1) to defined imaging protocol 404(n). In various cases, each of the multiple defined imaging protocols may be, or may otherwise represent, a different or unique configuration of scanning parameters that the medical imaging scanner 104 may use to capture medical images of the medical patient 106 (e.g., defined imaging protocol 404(1) may be a first protocol that the medical imaging scanner 104 may use to capture medical images of the medical patient 106; defined imaging protocol 404(n) may be an nth protocol that the medical imaging scanner 104 may use to capture medical images of the medical patient 106).

[0076] In various aspects, the protocol classification label 402 may include a plurality of probability scores 406. In various instances, the plurality of probability scores 406 may each correspond to (e.g., in a one-to-one manner) a plurality of defined imaging protocols 404. Thus, since the plurality of defined imaging protocols 404 may include n protocols, the plurality of probability scores 406 may also include n scores; probability scores 406(1) to probability scores 406(n). In various cases, each of the plurality of probability scores 406 may be a real-valued scalar indicating the probability (as inferred by the deep learning neural network 302) of the prescription document 202 recording, requesting, or otherwise invoking a corresponding one of the defined imaging protocols 404. As a non-limiting example, probability score 406(1) may correspond to defined imaging protocol 404(1). Therefore, probability score 406(1) can be a first scalar estimated by deep learning neural network 302, and its value (e.g., ranging from 0 to 1 or 0% to 100%) indicates the likelihood that prescription document 202 requests the execution of defined imaging protocol 404(1) on medical patient 106. As another non-limiting example, probability score 406(n) can correspond to defined imaging protocol 404(n). Therefore, probability score 406(n) can be an nth scalar estimated by deep learning neural network 302, and its value indicates the likelihood that prescription document 202 requests the execution of defined imaging protocol 404(n) on medical patient 106.

[0077] It should be noted that in some cases, the multiple probability scores 406 may not be independent of each other. As a non-limiting example, the multiple probability scores 406 may be restricted such that their sum can be one (e.g., it can be 1 or 100%). In this case, the deep learning neural network 302 can be considered to determine that the prescription document 202 requests or invokes only one of the multiple defined imaging protocols 404 (e.g., any one of the multiple defined imaging protocols 404 with the highest probability score can be considered to be indicated by the protocol classification label 402). However, in other cases, the multiple probability scores 406 may be independent of each other. As a non-limiting example, each probability score in the multiple probability scores 406 may be in the range of 0 (e.g., 0%) to 1 (e.g., 100%), regardless of the value of any other probability score in the multiple probability scores 406 (e.g., there is no uniform restriction on the sum of the multiple probability scores 406). In this case, the deep learning neural network 302 can be considered to be able to determine several defined imaging protocols among the multiple defined imaging protocols 404 requested or invoked by the prescription document 202 to perform on the medical patient 106 (e.g., any one or more defined imaging protocols among the multiple defined imaging protocols 404 that have a probability score exceeding any suitable threshold can be considered to be indicated by the protocol classification label 402).

[0078] In any case, at least one of the multiple defined imaging protocols 404 may be indicated by the protocol classification label 402 as requested or invoked by the prescription document 202, and such protocol may be referred to as the prescribed imaging protocol 304.

[0079] Now, consider Figure 5 The deep learning neural network 302 can alternatively be configured as a large language model (LLM) instead of a text classifier. In this case, the deep learning neural network 302 may include an encoder section 502 and a synthesizer section 504. In various cases, the encoder section 502 can be considered upstream of the synthesizer section 504. Equivalently, the synthesizer section 504 can be considered downstream of the encoder section 502.

[0080] In various aspects, the encoder section 502 can have any suitable deep learning internal architecture. In fact, in various aspects, the encoder section 502 can have an input layer, one or more hidden layers, and an output layer. In various instances, any of these layers can be coupled together by any suitable inter-neuron or inter-layer connections (such as forward connections, skip connections, or recursive connections). Furthermore, in various cases, any of these layers can be any suitable type of neural network layer with any suitable learnable or trainable internal parameters. For example, any of such an input layer, one or more hidden layers, or output layer can be a convolutional layer, whose learnable or trainable parameters can be convolutional kernels. As another example, any of such an input layer, one or more hidden layers, or output layer can be a dense layer, whose learnable or trainable parameters can be a weight matrix or bias values. As yet another example, any of such an input layer, one or more hidden layers, or output layer can be a batch normalization layer, whose learnable or trainable parameters can be shift factors or scaling factors. As yet another example, any of such an input layer, one or more hidden layers, or output layer can be an LSTM layer, whose learnable or trainable parameters can be an input state weight matrix or a hidden state weight matrix. As yet another example, any of such an input layer, one or more hidden layers, or output layer can be a transformer layer, whose learnable or trainable parameters can be a single-head or multi-head attention block or other weight matrix. Furthermore, in various cases, any of such layers can be any suitable type of neural network layer with any suitable fixed or untrainable internal parameters. For example, any of such an input layer, one or more hidden layers, or output layer can be a nonlinear layer, a padding layer, a pooling layer, or a cascaded layer.

[0081] Similarly, in various instances, the synthesizer section 504 can have any suitable deep learning internal architecture. In fact, in various cases, the synthesizer section 504 can have an input layer, one or more hidden layers, and an output layer. In various instances, any of these layers can be coupled together by any suitable inter-neuron or inter-layer connections (e.g., forward connections, skip connections, recursive connections). Furthermore, in various cases, any of these layers can be any suitable type of neural network layer with any suitable learnable or trainable internal parameters (e.g., any of such an input layer, one or more hidden layers, or output layer can be a convolutional layer, a dense layer, a batch normalization layer, an LSTM layer, or a transformer layer). Further still, in various cases, any of such layers can be any suitable type of neural network layer with any suitable fixed or non-trainable internal parameters (e.g., any of such an input layer, one or more hidden layers, or output layer can be a non-linear layer, a padding layer, a pooling layer, or a cascaded layer).

[0082] Regardless of the specific internal architecture implemented within encoder section 502, encoder section 502 can be configured to receive text data (which may be accompanied by any suitable numerical or graphical data) and generate embeddings based on such input text data. In contrast, regardless of the specific internal architecture implemented within synthesizer section 504, synthesizer section 504 can be configured to receive embeddings generated by encoder section 502 and generate synthesized text content based on such embeddings.

[0083] In various respects, the embedding generated by encoder section 502 in response to a set of input text, numerical, or graphical data can be considered as any suitable mathematical quantity (e.g., scalar, vector, matrix, tensor, or any suitable combination thereof) that represents at least some entity or semantic aspects of the input text, numerical, or graphical data in a low-dimensional manner. In other words, the embedding may be smaller in size or dimension than such input text, numerical, or graphical data (e.g., in some cases, one or more orders of magnitude smaller); however, despite its smaller size, the embedding can still be considered as substantially or semantically representing such input text, numerical, or graphical data. In yet another way, the embedding can be considered as a latent vector representation of such input text, numerical, or graphical data.

[0084] In any case, the deep learning neural network 302 can be constructed as an LLM in some instances. As some non-limiting examples, the deep learning neural network 302 can be any of the following: ChatGPT; GitHub Or Amazon

[0085] In various aspects, there may be a protocol identification prompt 506 that can be fed into the deep learning neural network 302. In various aspects, the protocol identification prompt 506 may be one or more unstructured or plain text sentences or sentence fragments that request or command any imaging protocol conveyed by the prescription document 202 or recorded in the prescription document. As a non-limiting example, the protocol identification prompt 506 may be the following sentence: "What imaging protocol (if any) does the prescription document 202 instruct the medical imaging scanner 104 to use?" As another non-limiting example, the protocol identification prompt 506 may be the following sentence: "Identifies what scanning protocol the prescription document 202 requests."

[0086] Now, in various instances, protocol component 116 may electronically execute deep learning neural network 302 on prescription document 202 and protocol identification prompt 506. In various cases, such execution may cause deep learning neural network 302 to generate protocol indication 508. More specifically, protocol component 116 may cascade prescription document 202 with protocol identification prompt 506. In various aspects, protocol component 116 may feed this cascade to the input layer of encoder section 502. In various aspects, this cascade may complete a forward pass through one or more hidden layers of encoder section 502. In various instances, the output layer of encoder section 502 may compute or otherwise compute one or more embeddings (not shown) based on activation maps or feature maps provided by one or more hidden layers of encoder section 502. In various cases, the one or more embeddings may be routed to the input layer of synthesizer section 504. In each respect, the one or more embeddings can complete the forward pass through one or more hidden layers of synthesizer section 504, and the output layer of synthesizer section 504 can compute or otherwise calculate protocol indication 508 based on the activation map or feature map provided by one or more hidden layers of synthesizer section 504.

[0087] In various respects, protocol instruction 508 may be one or more unstructured or plaintext declarative sentences or sentence fragments that semantically answer or respond to protocol identification prompt 506. That is, protocol instruction 508 may be (in the context of deep learning neural network 302) a synthetic text that names, states, or otherwise identifies a particular imaging protocol described, invoked, or otherwise specified in prescription document 202, and such particular imaging protocol may be referred to as specified imaging protocol 304.

[0088] although Figures 4 to 5The embodiments in which protocol component 116 utilizes a text classifier or LLM to identify the prescribed imaging protocol 304 are described, but these are merely non-limiting examples for ease of explanation and illustration. In various other embodiments, protocol component 116 may alternatively utilize any other suitable natural language processing technique to identify the prescribed imaging protocol 304 from or based on the prescription document 202.

[0089] In any case, protocol component 116 may identify the prescribed imaging protocol 304 based on prescription document 202. It should be noted that in some cases, such identification may be due to the description of a specific body part, tissue type, or pathology in prescription document 202. In practice, different imaging protocols may be considered specialized or customized to capture images of different types of body parts. As a non-limiting example, a first imaging protocol may be optimized to capture images of the head or brain, while a second imaging protocol may be optimized to capture images of the knee joint. Additionally, it should be noted that different imaging protocols may be considered specialized or customized to capture images of different types of tissue, symptoms, pathology, or other anatomical material within a given body part. As a non-limiting example, a third imaging protocol may be optimized to capture images of a healthy brain, while a fourth imaging protocol may be optimized to capture images of a brain with lesions or tumors, and a fifth imaging protocol may be optimized to capture images of a brain with ischemic stroke. Therefore, in some respects, the prescription document 202 may record or specify various body parts or pathological symptoms, and the deep learning neural network 302 (or any other suitable natural language processing technique) may identify which available protocol (e.g., which one in 404) is mapped or associated with those recorded or specified body parts or pathological symptoms, and such protocol may be referred to or considered as the prescribed imaging protocol 304.

[0090] As a non-limiting example, prescription document 202 may or may state "right knee bone (rt.kneebone)," and protocol component 116 may (e.g., via deep learning neural network 302) interpret this to invoke any imaging protocol configured to investigate the bone associated with the right knee. As another non-limiting example, prescription document 202 may or may state "left knee cartilage (knee L cart)," and protocol component 116 may interpret this to invoke any imaging protocol configured to investigate the cartilage associated with the left knee. As yet another non-limiting example, prescription document 202 may or may state "head i-stroke," and protocol component 116 may interpret this to invoke any imaging protocol configured to investigate any ischemic stroke associated with the head.

[0091] Figure 6An example non-limiting block diagram 600 illustrates a specified imaging protocol 304 according to one or more embodiments described herein.

[0092] In various aspects, as shown in the figures, the specified imaging protocol 304 may include, specify, or otherwise associate with the necessary scanner configuration 602. In various instances, the necessary scanner configuration 602 may be any suitable electronic data (e.g., one or more scalars, one or more vectors, one or more matrices, one or more tensors, one or more strings, or any suitable combination thereof) having any suitable format, size, or dimension, indicating or otherwise conveying specific values ​​or states that should be assigned to the medical imaging scanner 104 to enable the medical imaging scanner 104 to perform the specified imaging protocol 304. As a non-limiting example, the necessary scanner configuration 602 may indicate or specify which specific values ​​or states should be assigned to sequence type parameters, slice thickness parameters, slice orientation parameters, FOV parameters, matrix size parameters, TR parameters, TE parameters, or NEX parameters of the medical imaging scanner 104 to enable the medical imaging scanner 104 to perform the specified imaging protocol 304.

[0093] In various cases, as shown in the figure, the specified imaging protocol 304 may include, specify, or otherwise associate with the necessary body posture 604. In various aspects, the necessary body posture 604 may be any suitable electronic data (e.g., one or more scalars, one or more vectors, one or more matrices, one or more tensors, one or more strings, or any suitable combination thereof) having any suitable format, size, or dimension, which indicates or otherwise conveys the specific physical orientation that the specified imaging protocol 304 assumes or otherwise intends to be applied. As a non-limiting example, the specified imaging protocol 304 may assume or otherwise intend to be performed on a patient whose body is oriented in a prone (e.g., face down or abdomen down) position. As another non-limiting example, the specified imaging protocol 304 may assume or otherwise intend to be performed on a patient whose body is oriented in a supine (e.g., face up or abdomen up) position. As yet another non-limiting example, the specified imaging protocol 304 may assume or otherwise intend to be performed on a patient whose body is oriented in a right reclining (e.g., right lateral decubitus) position. As yet another non-limiting example, the specified imaging protocol 304 may be assumed or otherwise intended for a patient whose body is oriented in a left oblique (e.g., left lateral decubitus) position. As yet another non-limiting example, the specified imaging protocol 304 may be assumed or otherwise intended for a patient whose body is oriented in a head-forward (e.g., head entering the scanner before feet) position. As yet another non-limiting example, the specified imaging protocol 304 may be assumed or otherwise intended for a patient whose body is oriented in a feet-forward (e.g., head entering the scanner after feet) position. In any case, if the specified imaging protocol 304 is to be performed on a patient whose body is not oriented in the necessary body posture 604, the specified imaging protocol 304 will not be able to be performed correctly on that patient (e.g., any resulting scan images will be incorrect or filled with artifacts).

[0094] In various aspects, as shown, the defined imaging protocol 304 may include, specify, or otherwise associate with the necessary scanner coil location 606. In practice, as described above, the medical imaging scanner 104 may utilize or otherwise associate with the scanner coil 122. In various aspects, the necessary scanner coil location 606 may be any suitable electronic data (e.g., one or more scalars, one or more vectors, one or more matrices, one or more tensors, one or more strings, or any suitable combination thereof) having any suitable format, size, or dimension, indicating or otherwise conveying a specific on-body location of the scanner coil 122 that the defined imaging protocol 304 assumes or otherwise intends to be applied to. As a non-limiting example, the defined imaging protocol 304 may assume or otherwise intend to be performed on a patient wearing the scanner coil 122 on their head. As another non-limiting example, the defined imaging protocol 304 may assume or otherwise intend to be performed on a patient wearing the scanner coil 122 on their right upper arm. As yet another non-limiting example, the defined imaging protocol 304 may assume or otherwise intend to be performed on a patient wearing the scanner coil 122 on their left calf. In other words, the scanner coil position 606 can be considered as indicating the specific anatomical structure that should be scanned by the medical imaging scanner 104 and the laterality of that specific anatomical structure. In any case, if the prescribed imaging protocol 304 is to be performed on a patient wearing the scanner coil 122 at a location on the body different from the necessary scanner coil position 606, the prescribed imaging protocol 304 will not be able to be performed correctly on that patient (e.g., any resulting scan images will be incorrect or filled with artifacts).

[0095] Figure 7 A block diagram illustrating an example non-limiting system 700 comprising a second deep learning neural network and a preparation determination according to one or more embodiments described herein is shown. This system facilitates camera-based deep learning prediction and guidance for medical imaging protocols. As shown, in some cases, system 700 may include the same components as system 300, and may also include a deep learning neural network 702 and a preparation determination 704.

[0096] In various implementations, the preparation component 118 may electronically store, maintain, control, or otherwise electronically access the deep learning neural network 702. In various cases, the preparation component 118 may utilize the deep learning neural network 702 to generate a preparation determination 704 based on a prescribed imaging protocol 304 and on the prepared image or video 204. Relative to Figures 8 to 12 The non-restrictive aspects are described.

[0097] Figures 8 to 12Example non-limiting block diagrams illustrating how preparation determination 704 can be obtained according to one or more embodiments described herein.

[0098] First, consider Figure 8 A non-limiting block diagram 800 is provided. In various embodiments, as shown, the deep learning neural network 702 can be configured as a computer vision model with any suitable deep learning internal architecture. For example, in various cases, the deep learning neural network 702 may have an input layer, one or more hidden layers, and an output layer. In various aspects, any of such layers can be coupled together by any suitable inter-neuronal or inter-layer connections (such as forward connections, skip connections, or recursive connections). Furthermore, in various cases, any of such layers can be any suitable type of neural network layer with any suitable learnable or trainable internal parameters. For example, any of such an input layer, one or more hidden layers, or output layer can be a convolutional layer, the learnable or trainable parameters of which can be a convolutional kernel. As another example, any of such an input layer, one or more hidden layers, or output layer can be a dense layer, the learnable or trainable parameters of which can be a weight matrix or a bias value. As yet another example, any of such an input layer, one or more hidden layers, or output layer can be a batch normalization layer, the learnable or trainable parameters of which can be a shift factor or a scaling factor. As yet another example, any of such an input layer, one or more hidden layers, or output layer can be an LSTM layer, whose learnable or trainable parameters can be an input state weight matrix or a hidden state weight matrix. As yet another example, any of such an input layer, one or more hidden layers, or output layer can be a transformer layer, whose learnable or trainable parameters can be a single-head or multi-head attention block or other weight matrix. Furthermore, in various cases, any of such layers can be any suitable type of neural network layer with any suitable fixed or untrainable internal parameters. For example, any of such an input layer, one or more hidden layers, or output layer can be a nonlinear layer, a padding layer, a pooling layer, or a cascaded layer.

[0099] Regardless of the specific internal architecture implemented in the deep learning neural network 702 (e.g., the specific number, type, or organization of layers), the deep learning neural network 702 can be configured to receive visual data as input and produce various localizations as output based on the visual data received from that input. Therefore, the preparation unit 118 can electronically execute the deep learning neural network 702 on a preliminary image or video 204, and this execution can cause the deep learning neural network 702 to generate a set of patient body part localizations 802 or scanner coil localizations 804. More specifically, the preparation unit 118 can feed the prepared image or video 204 to the input layer of the deep learning neural network 702. In various cases, the prepared image or video 204 can be forward-propagated through one or more hidden layers of the deep learning neural network 702. In various aspects, the output layer of the deep learning neural network 702 can compute or calculate the set of patient body part localizations 802 or scanner coil localizations 804 based on activation maps or feature maps generated by one or more hidden layers of the deep learning neural network 702.

[0100] In various embodiments, the set of patient body part locations 802 may include any suitable number of patient body part locations. In various aspects, each patient body part location in the set of patient body part locations 802 may be any suitable electronic data (e.g., one or more scalars, one or more vectors, one or more matrices, one or more tensors, or any suitable combination thereof) having any suitable format, size, or dimension, indicating or otherwise representing the location of the corresponding body part of the medical patient 106 within the prepared image or video 204 (as inferred by the deep learning neural network 702). As a non-limiting example, a first patient body part location in the set of patient body part locations 802 may correspond to the right eye of the medical patient 106. Thus, the first patient body part location may be: a landmark coordinate indicating where the right eye of the medical patient 106 is located within the prepared image or video 204 (e.g., according to two-dimensional Cartesian coordinates); a bounding box surrounding the right eye of the medical patient 106 within the prepared image or video 204; or a segmentation mask surrounding or covering the right eye of the medical patient 106 within the prepared image or video 204. As another non-limiting example, the second patient body part location in this set of patient body part locations 802 may correspond to the left ankle of the medical patient 106. Therefore, the second patient body part location may be: landmark coordinates indicating where the left ankle of the medical patient 106 is located within the preparation image or video 204 (e.g., according to two-dimensional Cartesian coordinates); a bounding box surrounding the left ankle of the medical patient 106 within the preparation image or video 204; or a segmentation mask surrounding or covering the left ankle of the medical patient 106 within the preparation image or video 204. As yet another non-limiting example, the third patient body part location in this set of patient body part locations 802 may correspond to the nose of the medical patient 106. Therefore, the third patient body part localization can be: landmark coordinates indicating where the nose of the medical patient 106 is located within the preparation image or video 204 (e.g., according to two-dimensional Cartesian coordinates); a bounding box surrounding the nose of the medical patient 106 within the preparation image or video 204; or a segmentation mask surrounding or covering the nose of the medical patient 106 within the preparation image or video 204. In any case, this set of patient body part localizations 802 can be considered to indicate or show where the various body parts of the medical patient 106 are actually located or positioned when the medical patient 106 is preparing or waiting to be scanned by the medical imaging scanner 104 (in terms of the deep learning neural network 702).

[0101] In various embodiments, scanner coil positioning 804 can be any suitable electronic data (e.g., one or more scalars, one or more vectors, one or more matrices, one or more tensors, or any suitable combination thereof) having any suitable format, size, or dimension, indicating or otherwise representing the position of scanner coil 122 within the prepared image or video 204 (as inferred by deep learning neural network 702). As a non-limiting example, scanner coil positioning 804 can be landmark coordinates indicating where scanner coil 122 is located within the prepared image or video 204 (e.g., according to two-dimensional Cartesian coordinates). As another non-limiting example, scanner coil positioning 804 can be a bounding box surrounding scanner coil 122 within the prepared image or video 204. As yet another non-limiting example, scanner coil positioning 804 can be a segmentation mask surrounding or covering scanner coil 122 within the prepared image or video 204. In any case, scanner coil positioning 804 can be considered as indicating or showing where the medical patient 106 actually wears the scanner coil 122 when the medical patient 106 is ready or waiting to be scanned by the medical imaging scanner 104 (in terms of the deep learning neural network 702).

[0102] Figures 9 to 11 An example non-limiting embodiment depicts a preparation image or video 204 overlapping with an example non-limiting embodiment of the patient body part positioning 802 and scanner coil positioning 804. Specifically, Figure 9 Image 900 depicts a male patient lying supine on the worktable of an MRI scanner. Figure 9 The white circle in the image indicates the location 802 of the patient's body part. That is to say, in... Figure 9 In a non-limiting example, the set of patient body part locations 802 are landmark coordinates that define or represent the image location of the corresponding body part of a male patient. Figure 9 In non-limiting examples, these body parts include: right eye; left eye; nose; right ear; left ear; right shoulder; left shoulder; right elbow; left elbow; right wrist; left wrist; right hip; left hip; right knee; left knee; right ankle; and left ankle. Furthermore, Figure 9 The white rectangle in the image indicates the scanner coil positioning 804. That is, in... Figure 9 In the non-limiting example, scanner coil positioning 804 is a bounding box surrounding the scanner coil worn by the male patient. As shown in the figure, Figure 9 The scanner coil was worn on the left hand of the male patient. Figure 10 Image 1000 depicts the same male patient now lying prone with the scanner coil worn on his right hand. Figure 11Image 1100 depicts the same male patient lying in a supine position with a partially wrapped scanner coil worn on his right ankle.

[0103] Now, consider Figure 12 A non-limiting block diagram 1200. In various embodiments, the readiness determination 704 can be any suitable electronic data (e.g., one or more scalars, one or more vectors, one or more matrices, one or more tensors, one or more strings, or any suitable combination thereof) having any suitable format, size, or dimension, indicating in binary or dichotomous mode whether the medical patient 106 is properly ready to perform the prescribed imaging protocol 304. In various aspects, the readiness component 118 can electronically generate the readiness determination 704 by comparing the following: a set of patient body part positioning 802 and scanner coil positioning 804; with necessary body posture 604 and necessary scanner coil position 606.

[0104] As described above, the necessary body posture 604 can be considered as indicating any particular physical orientation or posture that the body of the medical patient 106 should exhibit, so that the prescribed imaging protocol 304 can be correctly performed on the medical patient 106. In various instances, this set of patient body part positioning 802 can be considered as collectively indicating or conveying any physical orientation or posture that the body of the medical patient 106 actually exhibits at the present, current, or otherwise most recent time (e.g., the pre-positioned camera 108 can capture real-time images or videos of the medical patient 106).

[0105] As a non-limiting example, suppose that the patient body part positioning 802 indicates that the eyes, nose, and ears of the medical patient 106 are located above the ankles of the medical patient 106 (from the perspective of the preparation camera 108). In this case, the preparation component 118 can determine from the patient body part positioning 802 that the medical patient 106 is currently or is oriented in a head-forward position.

[0106] As another non-limiting example, suppose that the patient body part positioning 802 indicates that the eyes, nose, and ears of the medical patient 106 are located below the ankles of the medical patient 106 (from the perspective of the preparation camera 108). In this case, the preparation component 118 can determine from the patient body part positioning 802 that the medical patient 106 is currently or is oriented in a feet-forward position.

[0107] As yet another non-limiting example, suppose that the patient body part positioning 802 indicates that the right shoulder, right hip, right elbow, and right knee of the medical patient 106 are located to the right of the left shoulder, left hip, left elbow, and left knee of the medical patient 106 (from the perspective of the preparation camera 108). In this case, the preparation component 118 can determine from the patient body part positioning 802 that the medical patient 106 is currently or is oriented in a prone position.

[0108] As yet another non-limiting example, suppose that the patient body part positioning 802 indicates that the right shoulder, right hip, right elbow, and right knee of the medical patient 106 are located to the left of the left shoulder, left hip, left elbow, and left knee of the medical patient 106 (from the perspective of the preparation camera 108). In this case, the preparation component 118 can determine from the patient body part positioning 802 that the medical patient 106 is currently or is oriented in a supine position.

[0109] As another non-limiting example, suppose the patient body part positioning 802 indicates that the right and left shoulders of the medical patient 106 are positioned above each other (from the perspective of the preparation camera 108) and that the eyes or nose of the medical patient 106 are positioned to the right of the shoulder or hip (from the perspective of the preparation camera 108). In this case, the preparation component 118 can determine from the patient body part positioning 802 that the medical patient 106 is currently or is oriented in a left-leaning posture.

[0110] As another non-limiting example, suppose that the patient body part positioning 802 indicates that the right and left shoulders of the medical patient 106 are positioned above each other (from the perspective of the preparation camera 108) and that the eyes or nose of the medical patient 106 are positioned to the left of the shoulder or hip (from the perspective of the preparation camera 108). In this case, the preparation component 118 can determine from the patient body part positioning 802 that the medical patient 106 is currently or is oriented in a right-leaning posture.

[0111] Furthermore, as described above, the necessary scanner coil position 606 can be considered as indicating any specific physical location on the body where the medical patient 106 should wear the scanner coil 122, so that the prescribed imaging protocol 304 can be correctly performed on the medical patient 106. In various instances, the scanner coil positioning 804 can be combined with this set of patient body part positioning 802 as an indication or communication of any physical location on the body where the medical patient 106 is currently or is wearing the scanner coil 122.

[0112] As a non-limiting example, assume that the scanner coil positioning 804 coincides with any of the patient body part positioning locations in this set of patient body part positioning locations 802 corresponding to the left wrist of the medical patient 106 (e.g., overlapping with, being on top of, or within any suitable threshold proximity to the patient body part positioning location). Figure 9 (As shown). In this case, the preparation component 118 can determine that the medical patient 106 is currently or at present wearing the scanner coil 122 on his left wrist or left hand.

[0113] As another non-limiting example, suppose that scanner coil positioning 804 coincides with any of the patient body part positioning locations in this set of patient body part positioning locations 802 corresponding to the right wrist of medical patient 106 (e.g., overlaps with, is on top of, or is within any suitable threshold proximity to, that patient body part positioning). Figure 10 (As shown). In this case, the preparation component 118 can determine that the medical patient 106 is currently or at present wearing the scanner coil 122 on his right wrist or right hand.

[0114] As another non-limiting example, suppose that scanner coil positioning 804 coincides with any of the patient body part positioning locations in this set of patient body part positioning locations 802 corresponding to the right ankle of medical patient 106 (e.g., overlaps with, is on top of, or is within any suitable threshold proximity to, that patient body part positioning). Figure 11 (As shown). In this case, the preparation component 118 can determine that the medical patient 106 is currently or at present wearing the scanner coil 122 on his right ankle or right foot.

[0115] As yet another non-limiting example, suppose the scanner coil positioning 804 coincides with any of the patient body part positioning locations in the set of patient body part positioning locations 802 corresponding to the eyes, ears, or nose of the medical patient 106 (e.g., overlapping with, on top of, or within any suitable threshold proximity to that patient body part positioning location). In this case, the preparation component 118 can determine that the medical patient 106 is currently or at present wearing the scanner coil 122 on their head.

[0116] Therefore, in various aspects, the preparation component 118 can generate a preparation determination 704 by comparing the necessary body posture 604 with the actual body posture of the medical patient 106 as conveyed by the set of patient body part positioning 802 and by comparing the necessary scanner coil position 606 with the actual on-body scanner coil position as conveyed by both the set of patient body part positioning 802 and scanner coil positioning 804. In various instances, the preparation determination 704 can instruct the medical patient 106 to be properly prepared for the prescribed imaging protocol 304 if both the necessary body posture 604 and the necessary scanner coil position 606 are satisfied (e.g., if the actual body posture of the medical patient 106 matches the necessary body posture 604, and if the actual on-body position of the scanner coil 122 matches the necessary scanner coil position 606). In contrast, if either the necessary body posture 604 or the necessary scanner coil position 606 is not met (e.g., if the actual body posture of the medical patient 106 does not match the necessary body posture 604, or if the actual body position of the scanner coil 122 does not match the necessary scanner coil position 606), then the preparation determination 704 may indicate that the medical patient 106 is not properly prepared for the prescribed imaging protocol 304.

[0117] Figure 13 A block diagram of an example non-limiting system 1300, comprising one or more preparatory actions and one or more guiding actions, is illustrated according to one or more embodiments described herein. This system facilitates camera-based deep learning prediction and guidance for medical imaging protocols. As shown, in some cases, system 1300 may include the same components as system 700 and may also include one or more preparatory actions 1302, one or more guiding actions 1304, and light or laser 1306.

[0118] In various implementations, the guidance component 120 may electronically initiate or otherwise perform any suitable action based on the readiness determination 704. Specifically, if the readiness determination 704 instructs the medical patient 106 to prepare for the prescribed imaging protocol 304, the guidance component 120 may perform or initiate one or more readiness actions 1302. Conversely, if the readiness determination 704 alternatively instructs the medical patient 106 not to prepare for the prescribed imaging protocol 304, the guidance component 120 may perform or initiate one or more guidance actions 1304.

[0119] In various aspects, one or more preparatory actions 1302 may include electronically sending instructions or commands to the medical imaging scanner 104 to begin executing a prescribed imaging protocol 304. This can be considered as an automatic start-up function of the medical imaging scanner 104 triggered by a preparation determination 704 instructing the medical patient 106 to prepare for the prescribed imaging protocol 304.

[0120] In various other respects, one or more preparatory actions 1302 may alternatively include electronically sending a first notification to the medical imaging scanner 104 (or any other computerized workstation associated with the medical imaging scanner 104), which may be displayed or presented on the GUI of the medical imaging scanner 104 (or on the GUI of other computerized workstations) such that the first notification is viewable by the user or operator of the medical imaging scanner 104. In various instances, the first notification may instruct the medical patient 106 to prepare for the commencement of a prescribed imaging protocol 304, and the first notification may further request or prompt the user or operator for permission to commence the execution of the prescribed imaging protocol 304. In some cases, the user or operator may (via the GUI) press, click, or otherwise select any suitable adjustable GUI button or element to indicate their permission, and the medical imaging scanner 104 may begin executing the prescribed imaging protocol 304 on the medical patient 106 in response to such press, click, or selection.

[0121] In various aspects, one or more guidance actions 1304 may include electronically sending a second notification to the medical imaging scanner 104 (or any other computerized workstation associated with the medical imaging scanner 104), which may be displayed or presented on the GUI of the medical imaging scanner 104 (or on the GUI of other computerized workstations) such that the second notification is viewable by the user or operator of the medical imaging scanner 104. In various instances, the second notification may indicate that the medical patient 106 is not yet ready to begin the prescribed imaging protocol 304. In various cases, the second notification may indicate or otherwise convey a specific reason explaining why the medical patient 106 is not yet ready or prepared. As a non-limiting example, if the preparation component 118 determines that the necessary body posture 604 is not met, the second notification may indicate or otherwise convey that the medical patient 106 will not be ready or prepared until their body is oriented to meet the necessary body posture 604. As another non-limiting example, if the preparation component 118 determines that the necessary scanner coil position 606 is not met, the second notification may indicate or otherwise convey that the medical patient 106 will not be ready or prepared until the scanner coil 122 is worn on their body in a manner conforming to the necessary scanner coil position 606. In some aspects, the second notification may include any suitable illustration or diagram visually indicating the necessary body posture 604 or the necessary scanner coil position 606 to assist the user or operator in understanding (e.g., the second notification may include presenting a preparation image or video 204 on which the necessary scanner coil position 606 may be highlighted or otherwise visually delineated).

[0122] In some instances, one or more guiding actions 1304 may include controlling the adjustment of light or laser 1306. In practice, in various cases, light or laser 1306 may be any suitable physically actuated, oriented, or targetable beam emitting device or laser beam emitting device associated with medical imaging scanner 104. In some aspects, light or laser 1306 may be physically built into or integrated into medical imaging scanner 104. In other aspects, light or laser 1306 may be physically located away from medical imaging scanner 104 or separate from medical imaging scanner but still located in the same room as medical imaging scanner. In still other aspects, light or laser 1306 may be physically located away from medical imaging scanner 104 or separate from medical imaging scanner and located in a different room from medical imaging scanner. In some instances, light or laser 1306 may be physically coupled or integrated into pre-camera 108 such that light or laser 1306 can illuminate anything within the field of view of pre-camera 108. In any case, when a medical patient 106 is preparing for or waiting to be scanned by a medical imaging scanner 104 (e.g., waiting in or on a rack, workbench, or gantry of the medical imaging scanner 104; or waiting in an adjacent donning / removal chamber associated with the medical imaging scanner 104), light or laser 1306 may be able to selectively emit a visible light beam or visible laser beam of any suitable color or intensity onto any suitable external body part of the medical patient. If the preparation component 118 determines that the necessary scanner coil position 606 is not satisfied, one or more guiding actions 1304 may include guiding, commanding, or otherwise aiming the light or laser 1306 (e.g., the light or laser 1306 may be moved by any suitable actuator, such as a servo motor) and significantly illuminating any body part of the medical patient 106 corresponding to the necessary scanner coil position 606. As a non-limiting example, it is assumed that the necessary scanner coil position 606 is the right hand of the medical patient 106. In this scenario, the patient body part positioning 802 can be considered to indicate the current or present real-world position of the right hand of the medical patient 106 relative to the medical imaging scanner 104, relative to the pre-camera 108, or relative to the light or laser 1306, and the guiding component 120 can cause the light or laser 1306 to illuminate the actual real-world right hand of the medical patient 106. As another non-limiting example, suppose the necessary scanner coil position 606 is the left knee of the medical patient 106. In this scenario, the patient body part positioning 802 can be considered to indicate the current or present real-world position of the left knee of the medical patient 106 relative to the medical imaging scanner 104, relative to the pre-camera 108, or relative to the light or laser 1306, and the guiding component 120 can cause the light or laser 1306 to illuminate the actual real-world left knee of the medical patient 106.In any case, the light beam or laser beam emitted by the light or laser 1306 may be considered to significantly or obviously indicate to the user or operator (or even the medical patient 106 itself) the position where the scanner coil 122 should be moved onto the body of the medical patient 106.

[0123] In various aspects, one or more guidance actions 1304 can be considered as showing or instructing the user or operator of the medical imaging scanner 104 how to correct the body posture or scanner coil position of the medical patient 106 to prepare it for the prescribed imaging protocol 304. After the user or operator has adjusted the medical patient 106, the preparation camera 108 may capture or record new preparation images or videos (e.g., a new or later instantiation of 204), the preparation component 118 may generate new preparation determinations (e.g., a new or later instantiation of 704), and the guidance component 120 may accordingly repeat one or more preparation actions 1302 or one or more guidance actions 1304.

[0124] Figure 14 A block diagram illustrating an example non-limiting system 1400 excluding a prescription document according to one or more embodiments described herein is shown. This system facilitates camera-based deep learning prediction and guidance for medical imaging protocols. As shown, in some cases, system 1400 may include the same components as system 1300, but may exclude or omit the prescription document 202.

[0125] In various implementations, the prescription document 202 may not be available. In such cases, the protocol component 116 can still utilize the deep learning neural network 302 to identify the prescribed imaging protocol 304. Relative to Figure 15 The non-restrictive aspects are described.

[0126] Figure 15 Example non-limiting block diagram 1500 illustrates how a prescribed imaging protocol 304 can be determined in the absence of a prescription document, according to one or more embodiments described herein.

[0127] In various implementations, as shown in the figure, the deep learning neural network 302 can be configured as an image classifier rather than as a text classifier or LLM. Even in this case, the deep learning neural network 302 can have any suitable deep learning internal architecture (e.g., any suitable type of layers arranged in any suitable format or layout and having any suitable trainable or non-trainable internal parameters).

[0128] Regardless of its specific internal architecture, Figure 15In non-limiting examples, the deep learning neural network 302 can be configured to receive visual data as input and produce classification labels for the input visual data as output. Therefore, the protocol component 116 can electronically execute the deep learning neural network 302 for a prepared image or video 204 (but not for a prescription document 202), and such execution can cause the deep learning neural network 302 to generate a protocol classification label 402. That is, the protocol component 116 can feed the prepared image or video 204 to the input layer of the deep learning neural network 302, the prepared image or video 204 can perform forward propagation through one or more hidden layers of the deep learning neural network 302, and the output layer of the deep learning neural network 302 can calculate or compute the protocol classification label 402 based on the activation map or feature map generated by one or more hidden layers of the deep learning neural network 302. In various aspects, the protocol classification label 402 can be as described above (e.g., it can include multiple probability scores 406 that may respectively correspond to multiple defined imaging protocols 404).

[0129] In the case where the deep learning neural network 302 generates protocol classification labels 402 based on the prepared image or video 204, the deep learning neural network 302 can be considered to infer the prescribed imaging protocol 304 based on any initial or rough physical position of the medical patient 106 wearing the scanner coil 122. As a non-limiting example, if the user or operator initially or at the beginning places the scanner coil 122 within any suitable threshold proximity to the medical patient 106's right hand, it is reasonably expected (and the deep learning neural network 302 can therefore infer or predict) that the prescribed imaging protocol 304 is any protocol corresponding to or tailored for the right hand. If multiple such protocols exist, any suitable disambiguation technique can be used (e.g., random selection among these protocols). As another non-limiting example, if the user or operator places the scanner coil 122 within any suitable threshold proximity to the medical patient 106's left knee, it is reasonably expected (and the deep learning neural network 302 can therefore infer or predict) that the prescribed imaging protocol 304 is any protocol corresponding to or tailored for the left knee. Again, if multiple such protocols exist, any suitable disambiguation technique can be used. In any case, the absence of a prescription document 202 can be overcome or addressed by inferring the prescribed imaging protocol 304 from the prepared image or video 204 (e.g., from the position of the scanner coil 122 at the beginning or initially placed on the medical patient 106).

[0130] Figure 16A flowchart illustrating an example non-limiting computer-implemented method 1600 according to one or more embodiments described herein is provided, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols. In various cases, a protocol guidance system 102 facilitates the computer-implemented method 1600.

[0131] In various implementations, action 1602 may include access to prescription text (e.g., 202) or video feed (e.g., 204) associated with a patient (e.g., 106) by a device (e.g., via 114) operatively coupled to a processor (e.g., 110).

[0132] In various aspects, action 1604 may include an imaging protocol (e.g., 304) to be inferred by the device (e.g., via 116) through a first neural network (e.g., 302) on a prescription text or video feed, to be implemented for the patient by the scanning modality (e.g., 104).

[0133] In various instances, action 1606 may include inferring the patient’s actual posture (e.g., jointly indicated by 802) or coil position (e.g., jointly indicated by 802 and 804) by the device (e.g., via 118) through a second neural network (e.g., 702) performed on a video feed.

[0134] In various cases, action 1608 may include determining, by the device (e.g., via 118), whether the actual pose or coil position matches the pose or coil position required by the imaging protocol (e.g., 604 or 606). If not, computer-implemented method 1600 may proceed to action 1610. If so, computer-implemented method 1600 may alternatively proceed to action 1612.

[0135] In various aspects, action 1610 may include the device (e.g., via 120) demonstrating, through a display on a screen or light / laser irradiated onto the patient, how the patient's actual posture or coil position can be aligned with or conform to the posture or coil position specified in the imaging protocol. In various cases, computer-implemented method 1600 may return to action 1606 (e.g., continuously updating video feedback over time).

[0136] In various instances, action 1612 may include initiating an imaging protocol on a scanning mode by the device (e.g., via 120) or prompting a user to initiate an imaging protocol on a scanning mode by the device (e.g., via 120).

[0137] To ensure the accuracy, correctness, or reliability of the predictions and guidance provided by the medical imaging protocols described herein, the various machine learning models described herein may first undergo training. Relative toFigure 17 An unrestricted example of this type of training is described.

[0138] Figure 17 Example non-limiting block diagram 1700 illustrates how various artificial intelligence models can be trained according to one or more embodiments described herein.

[0139] In all respects, before training begins, the trainable intrinsic parameters (e.g., convolutional kernels, weight matrices, bias values) of any artificial intelligence model being trained (e.g., deep learning neural network 302, deep learning neural network 702) can be initialized in any suitable manner (e.g., via random initialization).

[0140] In various implementations, a training input 1702 and a benchmark truth annotation 1704 may exist. When it is desired to train the deep learning neural network 302, the training input 1702 may be any suitable training prescription document (or, where appropriate, a cascade of the training prescription document with any suitable training protocol identifier; or, where appropriate, any suitable training image or video), and the benchmark truth annotation 1704 may be any correct or accurate protocol classification label known or considered to correspond to the training input 1702 (or, where appropriate, a correct or accurate protocol indication). When it is desired to train the deep learning neural network 702, the training input 1702 may be any suitable training image or video, and the benchmark truth annotation 1704 may be any correct or accurate patient body part location or correct or accurate scanner coil location known or considered to correspond to the training input 1702.

[0141] In any case, the training AI model can be executed on the training input 1702, thereby causing the AI ​​model to produce an output 1706. More specifically, in some cases, the training input 1702 can be fed or routed to the input layer of the AI ​​model, the training input 1702 can perform forward propagation through one or more hidden layers of the AI ​​model, and the output layer of the AI ​​model can compute the output 1706 based on the activation maps or feature maps provided by one or more hidden layers of the AI ​​model.

[0142] It should be noted that the format, size, or dimension of output 1706 can be determined by the number, arrangement, size, or other characteristics of neurons, convolutional kernels, LSTM weights, or other internal parameters of the output layer (or any other layer) of the AI ​​model. Therefore, output 1706 can be forced to have any desired format, size, or dimension by adding, removing, or otherwise adjusting the characteristics of the output layer (or any other layer) of the AI ​​model.

[0143] In all respects, if output 1706 is generated by deep learning neural network 302, then output 1706 can be considered as a protocol classification label (or, where appropriate, a protocol indication for prediction or inference) that deep learning neural network 302 believes should correspond to the training input 1702. If output 1706 is generated by deep learning neural network 702, then output 1706 can be considered as a patient body part localization or a scanner coil localization that deep learning neural network 702 believes should correspond to the training input 1702. It should be noted that if the trained AI model has not undergone or has undergone little to no training to date, output 1706 may be extremely inaccurate. In other words, output 1706 may differ significantly from the baseline truth annotation 1704.

[0144] In various aspects, the error 1708 between the output 1706 and the baseline truth annotation 1704 can be computed (e.g., mean absolute error (MAE), mean squared error (MSE), cross-entropy error). In various instances, the trainable intrinsic parameters of the artificial intelligence model can be incrementally updated based on the error 1708 via backpropagation (e.g., stochastic gradient descent).

[0145] In various cases, such execution and update procedures can be repeated for any suitable number of input-annotation pairs. This ultimately allows the trainable intrinsic parameters of the artificial intelligence model (e.g., AI models of Deep Learning Neural Network 302, Deep Learning Neural Network 702) to be iteratively optimized to accurately perform its inference task (e.g., protocol classification or determination; body part or scanner coil localization). In all aspects, any suitable training batch size, any suitable error / loss function, or any suitable training termination criterion can be utilized during such training.

[0146] While this document primarily describes various artificial intelligence models as being trained in a supervised manner, this is merely a non-limiting example for ease of explanation and illustration. In various implementations, any other suitable training paradigm, such as unsupervised training or reinforcement learning, can be used to train the deep learning neural network 302 or 702, either jointly or non-jointly.

[0147] Figure 18 A flowchart illustrating an example non-limiting computer-implemented method 1800 according to one or more embodiments described herein is provided, which facilitates camera-based deep learning prediction and guidance for medical imaging protocols. In various cases, a protocol guidance system 102 may facilitate the computer-implemented method 1800.

[0148] In various embodiments, action 1802 may include the execution of a device (e.g., via 116) operatively coupled to a processor (e.g., 110) via a first deep learning neural network (e.g., 302) to infer a prescribed imaging protocol (e.g., 304) to be performed by a medical imaging scanner (e.g., 104) on a medical patient (e.g., 106).

[0149] In various aspects, action 1804 may include the device (e.g., via 118) performing a second deep learning neural network (e.g., 702) on a preliminary image or video of the medical patient captured by a camera (e.g., 108) associated with a medical imaging scanner (e.g., 204) to infer whether the medical patient is ready for the prescribed imaging protocol.

[0150] In various instances, action 1806 may include an electronic guidance action (e.g., 1304) initiated by the device (e.g., via 120) in response to an inference (e.g., 704) that the medical patient is not ready for the prescribed imaging protocol, which explains or shows how to prepare the medical patient for the prescribed imaging protocol.

[0151] Despite Figure 18 Not explicitly shown, but the computer-implemented method 1800 may include: the device (e.g., via 120) responding to an inference that a medical patient is ready for a prescribed imaging protocol (e.g., 704) and presenting a notification (e.g., included in 1302) on the graphical user interface of the medical imaging scanner, the notification indicating that the prescribed imaging protocol is ready to be executed and requesting the user of the medical imaging scanner to approve the execution of the prescribed imaging protocol; or the device (e.g., via 120) instructing the medical imaging scanner to execute the prescribed imaging protocol.

[0152] Despite Figure 18 Not explicitly shown, but the first deep learning neural network may be a large language model that receives the following as input: a text prescription written by a medical professional treating the patient (e.g., 202); and a protocol identification prompt (e.g., 506); and produces synthetic text indicating the prescribed imaging protocol (e.g., 508) as output.

[0153] Despite Figure 18 Not explicitly shown, but the first deep learning neural network could be an image classifier that: receives a preliminary image or video of a medical patient captured by a camera as input; and produces a classification label (e.g., 402) indicating a prescribed imaging protocol as output.

[0154] Despite Figure 18Not explicitly shown, however, the second deep learning neural network may receive a preliminary image as input and produce a localization (e.g., 802) indicating the current body posture or orientation of the medical patient as output, and the device may infer that the medical patient is not ready in response to a mismatch between the current body posture or orientation and the required body posture or orientation specified in the prescribed imaging protocol (e.g., 604).

[0155] Despite Figure 18 Not explicitly shown, however, the second deep learning neural network may receive a preliminary image as input and produce a location (e.g., 804) indicating the current scanner coil position on the medical patient as output, and the device may infer that the medical patient is not ready in response to a mismatch between the current scanner coil position and the necessary scanner coil position specified in the prescribed imaging protocol (e.g., 606).

[0156] Despite Figure 18 Not explicitly shown, but electronic guidance actions may include the following presented on the graphical user interface of the medical imaging scanner: the current body posture or orientation of the medical patient inferred by a second deep learning neural network (e.g., 802) or the current scanner coil position (e.g., indicated jointly by 802 and 804); and the necessary body posture or orientation specified in the imaging protocol (e.g., 604) or the necessary scanner coil position (e.g., 606).

[0157] Despite Figure 18 It is not explicitly shown, but the current scanner coil position of the medical patient inferred by the second deep learning neural network (e.g., indicated jointly by 802 and 804) may fail to match the necessary scanner coil position (e.g., 606) specified in the prescribed imaging protocol, and electronic guidance actions may include irradiating the medical patient's body with light or laser (e.g., 1306) associated with the medical imaging scanner according to the necessary scanner coil position.

[0158] Despite Figure 18 It is not explicitly shown, but the camera may be located in a separate room from the medical imaging scanner.

[0159] The various embodiments described herein may relate to computer program products for facilitating camera-based deep learning prediction and guidance for medical imaging protocols. In various aspects, the computer program product may include a non-transitory computer-readable storage medium (e.g., 112) having program instructions embodied therein. In various instances, the program instructions are executable by a processor (e.g., 110) to cause the processor to execute a first deep learning neural network (e.g., 302) via a physician's prescription (e.g., 202) corresponding to a medical patient (e.g., 106) or a video feed depicting the medical patient (e.g., 204) to infer a prescribed imaging protocol (e.g., 304) to be performed on the medical patient by a magnetic resonance imaging (MRI) scanner (e.g., 104). In various cases, the program instructions can be further executed to cause the processor to infer, via a second deep learning neural network (e.g., 702) performing a video feed, whether the MRI coil position on the medical patient (e.g., jointly indicated by 802 and 804) fails to match the necessary MRI coil position (e.g., 606) specified in the prescribed imaging protocol (e.g., such inference represented by 704). In various aspects, the program instructions can be further executed to cause the processor, in response to the inference that the MRI coil position does not match the necessary MRI coil position, to irradiate the body of the medical patient with an actuable light or laser (e.g., 1306) associated with the MRI scanner, thereby visibly illuminating the necessary MRI coil position on the medical patient (e.g., such action represented by 1304). In various instances, the program instructions can be further executed to cause the processor, in response to the inference that the MRI coil position does match the necessary MRI coil position, to cause the MRI scanner to perform the prescribed imaging protocol (e.g., such action represented by 1302).

[0160] While the various embodiments described herein primarily depict embodiments in which the scanner coil is positioned by a deep learning neural network 702 and compared with the necessary scanner coil positions specified by or otherwise corresponding to the prescribed imaging protocol 304, these are merely non-limiting examples for ease of explanation and illustration. In various cases, the deep learning neural network 702 may be trained or configured to locate any other suitable object of interest that may be worn by a medical patient and is related to or otherwise affects the medical imaging scan (e.g., it may be trained or configured to locate a radiation vest or shield worn by a medical patient; it may be trained or configured to locate a dental mouth opener or lip retractor worn by a medical patient).

[0161] In various instances, machine learning algorithms or models may be implemented in any suitable manner to facilitate any suitable aspect described herein. To facilitate some of the machine learning aspects described above in various implementations, consider the following discussion of artificial intelligence (AI). The various implementations described herein may employ artificial intelligence to facilitate the automation of one or more features or functionalities. These components may employ various AI-based schemes to perform the various implementations / examples disclosed herein. To provide or contribute to the numerous determinations described herein (e.g., determination, detection, inference, accounting, prediction, prognosis, estimation, derivation, forecasting, detection, computation), the components described herein may examine the entirety or a subset of the data to which they have been granted access and may provide inference about or determine the state of a system or environment from a set of observations captured via events or data. For example, determinations may be used to identify specific contexts or actions, or to generate probability distributions of states. These determinations may be probabilistic; that is, the calculation of the probability distribution of states of interest is based on consideration of data and events. Determination may also refer to techniques used to compose higher-level events from a collection of events or data.

[0162] Such determinations can lead to the construction of new events or actions from observed events or a collection of stored event data, regardless of whether the events are closely related in time, and regardless of whether the events and data come from one or more event and data sources. The components disclosed herein can be combined to perform automatic or deterministic actions related to the claimed subject matter using various classification schemes or systems (e.g., support vector machines, neural networks, expert systems, Bayesian confidence networks, fuzzy logic, data fusion engines, etc.) that are explicitly trained (e.g., via training data) and implicitly trained (e.g., via observed behavior, preferences, historical information, received external information, etc.)). Therefore, classification schemes or systems can be used for automatic learning and the performance of multiple functions, actions, or determinations.

[0163] The classifier can take the input attribute vector z = (z1, z2, z3, z4, z5) as input. n This maps the input to a confidence level that it belongs to a certain category, as shown by f(z) = confidence level (category). Such classification can employ probability-based or statistical analysis (e.g., analyzing utility and cost considerations) to determine the action to be automated. Support Vector Machines (SVMs) are an example of a usable classifier. SVMs operate by finding a hypersurface in the space of possible inputs, where the hypersurface attempts to separate triggering criteria from non-triggering events. Intuitively, this makes the classification correct for test data that are close to but different from the training data. Other directed and undirected model classification methods include, for example, Naive Bayes, Bayesian networks, decision trees, neural networks, fuzzy logic models, or any of the probabilistic classification models that provide different independent patterns. Classification, as used in this paper, also includes statistical regression for developing priority models.

[0164] To provide additional context for the various implementation schemes described herein, Figure 19 The following discussion is intended to provide a brief general description of a suitable computing environment 1900 in which various implementations of the embodiments described herein may be implemented. Although the implementations have been described above in the general context of computer-executable instructions that can run on one or more computers, those skilled in the art will recognize that these implementations may also be implemented in combination with other program modules or as a combination of hardware and software.

[0165] Typically, program modules include routines, programs, components, data structures, etc., that perform specific tasks or implement specific abstract data types. Furthermore, those skilled in the art will understand that the methods of this invention can be practiced with other computer system configurations, including single-processor or multi-processor computer systems, minicomputers, mainframe computers, Internet of Things (IoT) devices, distributed computing systems, and personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, each operatively coupled to one or more associated devices.

[0166] The embodiments illustrated in this paper can also be practiced in a distributed computing environment where a specific task is performed by a remote processing device linked via a communication network. In a distributed computing environment, program modules can reside on both local and remote memory storage devices.

[0167] Computing devices typically include a variety of media, which may include computer-readable storage media, machine-readable storage media, or communication media, wherein the two terms are used interchangeably herein, as described below. A computer-readable storage medium or a machine-readable storage medium can be any available storage medium accessible by a computer, and includes volatile and non-volatile media, removable and non-removable media. By way of example and not limitation, a computer-readable storage medium or a machine-readable storage medium can be implemented in conjunction with any method or technology used for storing information, such as computer-readable or machine-readable instructions, program modules, structured data, or unstructured data.

[0168] Computer-readable storage media may include, but is not limited to, random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, optical disc read-only memory (CD-ROM), digital versatile disc (DVD), Blu-ray disc (BD) or other optical disc storage devices, magnetic tape cassettes, magnetic tape, disk storage devices or other magnetic storage devices, solid-state drives or other solid-state storage devices, or other tangible or non-transitory media that can be used to store desired information. In this regard, the terms “tangible” or “non-transitory” used herein to describe storage devices, memories, or computer-readable media should be understood to exclude only the propagation of transient signals themselves as a modifier, and do not waive the rights of all standard storage devices, memories, or computer-readable media that do not only propagate transient signals themselves.

[0169] Computer-readable storage media can be accessed by one or more local or remote computing devices, for example, via access requests, queries or other data retrieval protocols, to perform various operations with respect to the information stored on the media.

[0170] Communication media typically contain computer-readable instructions, data structures, program modules, or other structured or unstructured data in a data signal, such as a modulated data signal, a carrier wave, or other transmission mechanism, and include any information transmission or delivery medium. The terms "modulated data signal" or "signal" refer to a signal whose one or more characteristics are set or altered to encode information in one or more signals. By way of example and not limitation, communication media include wired media (such as wired networks or direct wired connections) and wireless media (such as acoustic, RF, infrared, and other wireless media).

[0171] Refer again Figure 19 An exemplary environment 1900 for implementing various embodiments of the aspects described herein includes a computer 1902, which includes a processing unit 1904, a system memory 1906, and a system bus 1908. The system bus 1908 couples system components, including but not limited to the system memory 1906, to the processing unit 1904. The processing unit 1904 can be any of a variety of commercially available processors. Dual microprocessors and other multi-processor architectures may also be used as the processing unit 1904.

[0172] System bus 1908 can be any of several types of bus structures that can be further interconnected to memory bus (with or without memory controller), peripheral bus, and local bus using any of a variety of commercially available bus architectures. System memory 1906 includes ROM 1910 and RAM 1912. The basic input / output system (BIOS) may be stored in non-volatile memory (such as ROM, erasable programmable read-only memory (EPROM), EEPROM), where the BIOS contains basic routines that facilitate, for example, the transfer of information between components within computer 1902 during startup. RAM 1912 may also include high-speed RAM, such as static RAM for caching data.

[0173] Computer 1902 also includes an internal hard disk drive (HDD) 1914 (e.g., EIDE, SATA), one or more external storage devices 1916 (e.g., floppy disk drive (FDD) 1916, memory stick or flash drive reader, memory card reader, etc.), and drives 1920 (e.g., solid-state drives, optical disc drives) capable of reading from or writing to disks 1922 (such as CD-ROMs, DVDs, BDs, etc.). Alternatively, in cases involving solid-state drives, disks 1922 are not included unless separate. Although the internal HDD 1914 is illustrated as residing within computer 1902, the internal HDD 1914 may also be configured for use outside of suitable infrastructure (not shown). Additionally, although not shown in environment 1900, solid-state drives (SSDs) may be used as a supplement or alternative to HDD 1914. HDD 1914, external storage device 1916, and drive 1920 can be connected to system bus 1908 via HDD interface 1924, external storage interface 1926, and drive interface 1928, respectively. Interface 1924 for the specific implementation of the external drive may include at least one or both of Universal Serial Bus (USB) and Institute of Electrical and Electronics Engineers (IEEE) 1394 interface technologies. Other external drive connection technologies are contemplated in the embodiments described herein.

[0174] Drives and their associated computer-readable storage media provide non-volatile storage of data, data structures, computer-executable instructions, etc. For the computer 1902, drives and storage media are adapted to store any data in a suitable digital format. Although the above description of computer-readable storage media refers to a corresponding type of storage device, those skilled in the art will understand that other types of computer-readable storage media (whether currently existing or developed in the future) can also be used in the example operating environment, and furthermore, any such storage media may contain computer-executable instructions for performing the methods described herein.

[0175] Multiple program modules may be stored in the drive and RAM 1912, including the operating system 1930, one or more application programs 1932, other program modules 1934, and program data 1936. All or part of the operating system, application programs, modules, or data may also be cached in RAM 1912. The systems and methods described herein can be implemented using a variety of commercially available operating systems or combinations of operating systems.

[0176] Computer 1902 may optionally include emulation technology. For example, a hypervisor (not shown) or other intermediary may emulate the hardware environment used for operating system 1930, and the emulated hardware may optionally differ from that of the computer. Figure 19 The hardware illustrated herein. In such an implementation, operating system 1930 may include one of a plurality of virtual machines (VMs) hosted at computer 1902. Furthermore, operating system 1930 may provide a runtime environment, such as the Java Runtime Environment or the .NET Framework, to application 1932. A runtime environment is a consistent execution environment that allows application 1932 to run on any operating system that includes that runtime environment. Similarly, operating system 1930 may support containers, and application 1932 may be in the form of a container of lightweight, standalone, executable software packages, including, for example, application code, runtime, system tools, system libraries, and settings.

[0177] Furthermore, the computer 1902 can be enabled using security modules such as Trusted Processing Modules (TPMs). For example, in the case of a TPM, the boot part is hashed in the next boot part and waits for the result to match a security value before loading the next boot part. This process can occur at any layer of the computer 1902's code execution stack, such as at the application execution level or the operating system (OS) kernel level, thereby achieving security at any code execution level.

[0178] Users can input commands and information into computer 1902 through one or more wired / wireless input devices (e.g., keyboard 1938, touchscreen 1940, and pointing devices such as mouse 1942). Other input devices (not shown) may include microphones, infrared (IR) remote controls, radio frequency (RF) remote controls or other remote controls, joysticks, virtual reality controllers or virtual reality headsets, gamepads, styluses, image input devices (e.g., cameras), gesture sensor input devices, visual motion sensor input devices, emotion or face detection devices, biometric input devices (e.g., fingerprint or iris scanners), etc. These and other input devices are often connected to processing unit 1904 via input device interface 1944, which may be coupled to system bus 1908, but these and other input devices may also be connected via other interfaces (e.g., parallel ports, IEEE 1394 serial ports, game ports, USB ports, IR interfaces, etc.). (Interfaces, etc.) connections.

[0179] Monitor 1946 or other types of display devices may also be connected to system bus 1908 via an interface (such as video adapter 1948). In addition to monitor 1946, the computer typically includes other peripheral output devices (not shown), such as speakers, printers, etc.

[0180] Computer 1902 can operate in a networked environment using a logical connection to one or more remote computers (such as remote computer 1950) via wired or wireless communication. Remote computer 1950 can be a workstation, server computer, router, personal computer, portable computer, microprocessor-based entertainment device, peer-to-peer device, or other public network node, and typically includes many or all of the elements described relative to computer 1902, but for simplicity, only memory / storage device 1952 is illustrated. The depicted logical connection includes wired / wireless connectivity to a local area network (LAN) 1954 or a larger network (e.g., a wide area network (WAN) 1956). Such LAN and WAN networking environments are common in offices and companies and facilitate enterprise-wide computer networks (such as intranets), all of which can be connected to global communications networks (e.g., the Internet).

[0181] When used in a LAN networking environment, computer 1902 can connect to local network 1954 via a wired or wireless communication network interface or adapter 1958. Adapter 1958 facilitates wired or wireless communication with LAN 1954, which may also include wireless access points (APs) configured thereon for communication with adapter 1958 in wireless mode.

[0182] When used in a WAN networking environment, computer 1902 may include modem 1960, or may be connected to a communication server on WAN 1956 via other components (such as via the Internet) for establishing communication over WAN 1956. Modem 1960, which may function as an internal or external device and as a wired or wireless device, may be connected to system bus 1908 via input device interface 1944. In a networking environment, program modules depicted relative to computer 1902 or parts thereof may be stored in remote memory / storage device 1952. It should be understood that the network connections shown are illustrative, and other components for establishing communication links between computers may be used.

[0183] When used in a LAN or WAN networking environment, as a supplement to or alternative to the aforementioned external storage device 1916, computer 1902 can access cloud storage systems or other network-based storage systems, such as, but not limited to, network virtual machines that provide one or more aspects of information storage or processing. Generally, the connection between computer 1902 and the cloud storage system can be established, for example, via adapter 1958 or modem 1960 through LAN 1954 or WAN 1956. When computer 1902 is connected to an associated cloud storage system, external storage interface 1926 can manage the storage provided by the cloud storage system by means of adapter 1958 or modem 1960, just as with other types of external storage devices. For example, external storage interface 1926 can be configured to provide access to cloud storage sources as if those cloud storage sources were physically connected to computer 1902.

[0184] Computer 1902 can be operated to communicate with any wireless device or entity located wirelessly, such as a printer, scanner, desktop or portable computer, portable data assistant, communications satellite, any equipment or location associated with a wirelessly detectable tag (e.g., self-service machine, newsstand, store shelf, etc.), and telephone. This can include Wi-Fi and Wireless technology. Therefore, communication can be a predefined structure like a regular network, or simply self-organizing communication between at least two devices.

[0185] Figure 20This is a schematic block diagram of a sample computing environment 2000 to which the disclosed subject matter can interact. The sample computing environment 2000 includes one or more clients 2010. Clients 2010 can be hardware or software (e.g., threads, processes, computing devices). The sample computing environment 2000 also includes one or more servers 2030. Servers 2030 can also be hardware or software (e.g., threads, processes, computing devices). For example, server 2030 may accommodate threads to perform transformations by employing one or more implementations as described herein. One possible communication between client 2010 and server 2030 may be in the form of data packets suitable for transmission between two or more computer processes. The sample computing environment 2000 includes a communication framework 2050 that can be used to facilitate communication between client 2010 and server 2030. Client 2010 is operatively connected to one or more client data repositories 2020, which can be used to store information local to client 2010. Similarly, server 2030 is operatively connected to one or more server data repositories 2040, which can be used to store information locally on server 2030.

[0186] Various implementations can be systems, methods, apparatus, or computer program products at any possible level of integrated technical detail. A computer program product may include a computer-readable storage medium (or media) having computer-readable program instructions thereon for causing a processor to implement aspects of various implementations. The computer-readable storage medium may be a tangible device that holds and stores instructions for use by an instruction execution device. The computer-readable storage medium may be, for example, but not limited to, electronic storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. A less complete list of more specific examples of computer-readable storage media may also include: portable computer floppy disks, hard disks, solid-state drives such as M.2 (including Fast Non-Volatile Memory (NVMe) or Serial Advanced Technology Attachment (SATA)), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable optical disc read-only memory (CD-ROM), digital versatile disk (DVD), memory sticks, floppy disks, mechanical encoding devices (such as punch cards or raised structures in grooves on which instructions are recorded), and any suitable combination of the foregoing. As used herein, computer-readable storage media should not be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses passing through fiber optic cables), or electrical signals transmitted through wires.

[0187] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to a suitable computing / processing device, or downloaded to an external computer or external storage device via a network (e.g., the Internet, a local area network, a wide area network, or a wireless network). The network may include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them for storage in a computer-readable storage medium within the suitable computing / processing device. The computer-readable program instructions used to implement the various embodiments can be assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, integrated circuit configuration data, or source code or object code written in any combination of one or more programming languages ​​(including object-oriented programming languages ​​such as Smalltalk, C++, etc.) and procedural programming languages ​​(such as the "C" programming language or similar programming languages). Computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet through an Internet service provider). In some implementations, electronic circuitry, including, for example, programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), may execute computer-readable program instructions to personalize the electronic circuitry for various purposes by utilizing state information from the computer-readable program instructions.

[0188] The various aspects described herein are illustrated by flowchart illustrations or block diagrams of methods, apparatus (systems), and computer program products according to various embodiments. It should be understood that each block of a flowchart illustration or block diagram, and combinations of blocks in a flowchart illustration or block diagram, can be implemented by computer-readable program instructions. These computer-readable program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create components for implementing the functions / actions specified in one or more blocks of the flowchart or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that instructs a computer, programmable data processing apparatus, or other device to function in a particular manner, such that the computer-readable storage medium having the instructions stored therein includes an article of writing comprising instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart or block diagram. Computer-readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operations to be performed on the computer, other programmable apparatus, or other device to produce a computer-implemented process, such that the instructions, which execute on the computer, other programmable apparatus, or other device, implement the functions / actions specified in one or more blocks of the flowchart or block diagram.

[0189] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible specific implementations of systems, methods, and computer program products according to various embodiments. In this regard, each box in a flowchart or block diagram may represent a module, segment, or portion of instructions, comprising one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions indicated in the boxes may not occur in the order shown in the figures. For example, two boxes shown consecutively may actually be executed substantially simultaneously, or sometimes they may be executed in reverse order, depending on the functionality involved. It will also be noted that each box illustrated in a block diagram or flowchart, and combinations of boxes illustrated in a block diagram or flowchart, may be implemented by a dedicated hardware-based system that performs the specified function or action or implements a combination of dedicated hardware and computer instructions.

[0190] Although the subject matter has been described above in the general context of computer-executable instructions of a computer program product running on one or more computers, those skilled in the art will recognize that the present disclosure may also be implemented in combination with other program modules. Typically, program modules include routines, programs, components, data structures, etc., that perform a specific task or implement a specific abstract data type. Furthermore, those skilled in the art will understand that various aspects can be practiced using other computer system configurations, including single-processor or multi-processor computer systems, small computing devices, mainframe computers, and computers, handheld computing devices (e.g., PDAs, telephones), microprocessor-based or programmable consumer or industrial electronics, etc. The illustrated aspects can also be practiced in a distributed computing environment in which tasks are performed by remote processing devices linked via a communication network. However, some (if not all) aspects of the present disclosure can be practiced on a standalone computer. In a distributed computing environment, program modules may reside on both local and remote memory storage devices.

[0191] As used herein, the terms “component,” “system,” “platform,” “interface,” etc., may refer to or include computer-related entities or entities associated with an operator having one or more specific functionalities. Entities disclosed herein may be hardware, a combination of hardware and software, software, or software in execution. For example, a component may be, but is not limited to, a process running on a processor, a processor, an object, an executable file, a thread of execution, a program, or a computer. By way of example, both an application running on a server and the server itself can be components. One or more components may reside within a process or a thread of execution, and components may be located on a single computer or distributed across two or more computers. In another example, a corresponding component may execute on various computer-readable media on which various data structures are stored. Components may communicate via local or remote processes, such as based on signals having one or more data packets (e.g., data from one component that interacts with another component in a local system, a distributed system, or a network (such as the Internet with other systems) via signals). As another example, a component may be a device having specific functionalities provided by mechanical parts operated by electrical or electronic circuitry, which is operated by a software or firmware application executed by a processor. In such cases, the processor may be internal or external to the device and may execute at least a portion of a software or firmware application. As yet another example, a component may be a means of providing specific functionality through electronic parts rather than mechanical components, wherein the electronic parts may include a processor or other components for executing software or firmware that at least partially imparts functionality to the electronic parts. In one aspect, a component may be emulated, for example, within a cloud computing system via a virtual machine.

[0192] Furthermore, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or.” That is, unless otherwise specified or clear from the context, “X adopts A or B” is intended to mean any natural inclusive substitution. That is, if X adopts A; X adopts B; or X adopts both A and B, then “X adopts A or B” is satisfied in any of the foregoing cases. As used herein, the term “and / or” is intended to have the same meaning as “or.” Furthermore, unless otherwise specified or clear from the context to be directed to the singular form, the articles “a” and “an” used in this specification and figures should generally be understood to mean “one or more.” As used herein, the terms “example” or “exemplary” are used to mean used as an example, instance, or illustration. For the avoidance of doubt, the subject matter disclosed herein is not limited to such examples. Additionally, any aspect or design described herein as an “example” or “exemplary” is not necessarily to be construed as preferred or advantageous over other aspects or designs, nor does it exclude equivalent exemplary structures and techniques known to those skilled in the art.

[0193] The disclosure herein describes non-limiting examples. For ease of description or explanation, various parts of the disclosure herein use the terms "each," "every," or "all" when discussing various examples. Such use of the terms "each," "every," or "all" is non-limiting. In other words, when the disclosure herein provides a description applicable to "each," "every," or "all" of a particular object or component, it should be understood that this is a non-limiting example, and it should also be understood that in various examples, such a description may apply to fewer than "each," "every," or "all" of that particular object or component.

[0194] As used herein, the term "processor" can refer substantially to any computing processing unit or device, including but not limited to a single-core processor; a single processor with software multithreading capabilities; a multi-core processor; a multi-core processor with software multithreading capabilities; a multi-core processor with hardware multithreading technology; a parallel platform; and a parallel platform with distributed shared memory. Additionally, a processor can refer to an integrated circuit, an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a field-programmable gate array (FPGA), a programmable logic controller (PLC), a complex programmable logic device (CPLD), discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. Furthermore, processors can utilize nanoscale architectures (such as, but not limited to, molecular and quantum dot-based transistors, switches, and gates) to optimize space usage or enhance the performance of user equipment. Processors can also be implemented as a combination of computing processing units. In this disclosure, terms such as "repository," "storage device," "data repository," "data storage device," "database," and substantially any other information storage component related to the operation and functionality of a component are used to refer to a "memory component," an entity specifically embodied in "memory," or a component that includes memory. It should be understood that the memory or memory component described herein may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. By way of illustration and not limitation, non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM), flash memory, or non-volatile random access memory (RAM) (e.g., ferroelectric RAM (FeRAM)). For example, volatile memory may include RAM that can act as external cache memory. By way of illustration and not limitation, RAM can be provided in a variety of forms, such as synchronous RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), direct Rambus RAM (DRRAM), direct Rambus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM). Furthermore, the memory components disclosed in the systems or computer-implemented methods herein are intended to include, but are not limited to, these and any other suitable types of memory.

[0195] The foregoing description includes only examples of systems and computer-implemented methods. Of course, it is impossible to describe every conceivable combination of components or computer-implemented methods for the purposes of describing this disclosure, but many other combinations and substitutions of this disclosure are possible. Furthermore, regarding the extent to which the terms “comprising,” “having,” “possessing,” etc., are used in the detailed description, claims, appendices, and drawings, such terms are intended to be inclusive in a manner similar to the term “comprising,” as interpreted when “comprising” is used as a transitional word in the claims.

[0196] Various embodiments have been described for illustrative purposes, but these descriptions are not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and variations will be apparent without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best illustrate the principles of the embodiments, the practical application of or improvement of technology found in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A system comprising: A processor (e.g., 110) that executes computer-executable components stored in a non-transitory computer-readable memory (e.g., 112), wherein the computer-executable components include: Protocol component (e.g., 116) infers a prescribed imaging protocol (e.g., 304) to be performed by a medical imaging scanner (e.g., 104) on a medical patient (e.g., 106) via execution of a first deep learning neural network (e.g., 302); A preparation component (e.g., 118) that infers whether the medical patient is ready for the prescribed imaging protocol by executing a second deep learning neural network (e.g., 702) on preparation images or videos (e.g., 204) of the medical patient captured by a camera (e.g., 108) associated with the medical imaging scanner; and A guidance component (e.g., 120) initiates an electronic guidance action (e.g., 1304) in response to an inference (e.g., 704) that the medical patient is not ready for the prescribed imaging protocol, the electronic guidance action explaining or showing how to prepare the medical patient for the prescribed imaging protocol.

2. The system of claim 1, wherein the guidance component responds to an inference that the medical patient is ready for the prescribed imaging protocol (e.g., an alternative embodiment of 704): A notification is presented on the graphical user interface of the medical imaging scanner, indicating that the prescribed imaging protocol is ready to be executed and requesting the user of the medical imaging scanner to approve the execution of the prescribed imaging protocol; or The medical imaging scanner is instructed to execute the prescribed imaging protocol.

3. The system according to claim 1, wherein the first deep learning neural network is a large language model, the large language model being: Receives a text prescription (e.g., 202) prepared by a medical professional who has treated the patient as input; and Generate synthetic text (e.g., 508) as output, which indicates the prescribed imaging protocol based on anatomical structure or laterality.

4. The system according to claim 1, wherein the first deep learning neural network is an image classifier, the image classifier: Receive the prepared image or video of the medical patient captured by the camera as input; and Generate a classification label (e.g., 402) indicating the specified imaging protocol as output.

5. The system of claim 1, wherein the second deep learning neural network receives the preparation image or video as input and generates a localization (e.g., 802) indicating the current body posture or orientation of the medical patient as output, and wherein the preparation component infers that the medical patient is not ready in response to a mismatch between the current body posture or orientation and a necessary body posture or orientation specified in the prescribed imaging protocol (e.g., 604).

6. The system of claim 1, wherein the second deep learning neural network receives the preparation image or video as input and generates a positioning (e.g., 804) indicating the current scanner coil position on the medical patient as output, and wherein the preparation component infers that the medical patient is not ready in response to a mismatch between the current scanner coil position and a necessary scanner coil position (e.g., 606) specified in the prescribed imaging protocol.

7. The system of claim 1, wherein the electronically guided action is presented on the graphical user interface of the medical imaging scanner: The current body posture or orientation of the medical patient inferred by the second deep learning neural network (e.g., 802) or the current scanner coil position (e.g., 804); and The necessary body pose or orientation (e.g., 604) or necessary scanner coil position (e.g., 606) specified in the prescribed imaging protocol.

8. The system of claim 1, wherein the current scanner coil position (e.g., 804) of the medical patient inferred by the second deep learning neural network does not match the necessary scanner coil position (e.g., 606) specified in the prescribed imaging protocol, and wherein the electronic guidance action includes irradiating the body of the medical patient with light or laser (e.g., 1306) associated with the medical imaging scanner according to the necessary scanner coil position.

9. The system of claim 1, wherein the camera is located in a separate room from the medical imaging scanner.

10. A computer-implemented method, the computer-implemented method comprising: A device (e.g., via 116) operatively coupled to a processor (e.g., 110) infers a prescribed imaging protocol (e.g., 304) to be performed by a medical imaging scanner (e.g., 104) on a medical patient (e.g., 106) via the execution of a first deep learning neural network (e.g., 302). The device (e.g., via 118) infers whether the medical patient is ready for the prescribed imaging protocol by performing a second deep learning neural network (e.g., 702) on a prepared image or video of the medical patient captured by a camera (e.g., 108) associated with the medical imaging scanner (e.g., 204). as well as The device (e.g., via 120) initiates an electronic guidance action (e.g., 1304) in response to an inference (e.g., 704) that the medical patient is not ready for the prescribed imaging protocol, the electronic guidance action explaining or showing how to prepare the medical patient for the prescribed imaging protocol.

11. The computer-implemented method according to claim 10, further comprising: The device (e.g., via 120) responds to an inference that the medical patient is ready for the prescribed imaging protocol (e.g., an alternative embodiment of 704) and presents a notification on the graphical user interface of the medical imaging scanner, the notification indicating that the prescribed imaging protocol is ready to be executed and requesting the user of the medical imaging scanner to approve the execution of the prescribed imaging protocol; or The medical imaging scanner is directed by the device (e.g., via 120) to perform the prescribed imaging protocol.

12. The computer-implemented method of claim 10, wherein the first deep learning neural network is a large language model, the large language model being: Receives a text prescription (e.g., 202) prepared by a medical professional who has treated the patient as input; and Generate synthetic text (e.g., 508) as output, which indicates the prescribed imaging protocol based on anatomical structure or laterality.

13. The computer-implemented method of claim 10, wherein the first deep learning neural network is an image classifier, the image classifier: Receive the prepared image or video of the medical patient captured by the camera as input; and Generate a classification label (e.g., 402) indicating the specified imaging protocol as output.

14. The computer-implemented method of claim 10, wherein the second deep learning neural network receives the preparation image or video as input and generates a localization (e.g., 802) indicating the current body posture or orientation of the medical patient as output, and wherein the device infers that the medical patient is not ready in response to a mismatch between the current body posture or orientation and a necessary body posture or orientation specified in the prescribed imaging protocol (e.g., 604).

15. The computer-implemented method of claim 10, wherein the second deep learning neural network receives the preparation image or video as input and generates a positioning (e.g., 804) indicating the current scanner coil position on the medical patient as output, and wherein the device infers that the medical patient is not ready in response to a mismatch between the current scanner coil position and a necessary scanner coil position (e.g., 606) specified in the prescribed imaging protocol.

Images