Lung ultrasound imaging

Abstract

An ultrasound system (100) comprises a processor (1010) configured to: obtain (S315), in an ultrasound procedure, an ultrasound image using a transducer array (113) of an ultrasound probe (110); generating (S350) a score for reverberation in the ultrasound image; estimating (S355) the presence of one or more A-line structures in the ultrasound image based on the score for reverberation; and estimating (S360) an intensity of one or more A-line structures in the ultrasound image based on the score for reverberation.

Description

Technical Field

[0001] This invention relates to medical imaging, particularly lung ultrasound imaging. Background Technology

[0002] Lung ultrasound imaging is a point-of-care technique used to aid in the assessment of lung conditions such as pulmonary edema and respiratory infections. Healthy, air-filled lungs typically appear as numerous A-lines on ultrasound. A-lines are recurring, hyperechoic, horizontal artifacts that appear at regular intervals from the pleural line, the intervals equal to the distance between the skin and the pleural line. Specifically, A-lines appear when the ultrasound beam hits the highly reflective surface of the pleura and reflects back to the transducer. The first return results in a true image of the pleural surface on the monitor. However, the beam reflects back from the transducer surface again, and this cycle repeats. Reverberation can be defined as the prolongation or persistence of sound after it has stopped propagating. Multiple reverberations result in multiple A-lines at multiples of the pleural depth. In contrast, fluid-filled lungs (as is often the case in patients with heart failure or pneumonia) are typically characterized by B-lines. B-lines appear as dynamic structures extending vertically from the pleural line to the bottom of the screen. B-lines are a pathological feature resulting from fluid accumulation in the interstitial spaces, appearing as a "comet tail" artifact.

[0003] Accurately identifying A-lines and B-lines, distinguishing them from each other, and differentiating them from other visually similar lung ultrasound structures are all challenging aspects of existing quantitative B-line detection algorithms. The difficulty in accurately identifying A-lines and B-lines and differentiating them from other lung ultrasound structures stems from their visual similarity. The effectiveness of lung ultrasound imaging may depend on operator experience, image quality, and the choice of imaging settings. For ultrasound users, the assessment of B-lines can play a crucial role in screening, diagnosis, or the management of disease progression. However, even for experienced ultrasound users, a standardized measurement method is not provided. Summary of the Invention

[0004] The purpose of this invention is particularly to assist in the identification of A-lines in ultrasound imaging.

[0005] This invention is defined by the independent claims. The dependent claims define advantageous embodiments.

[0006] According to an embodiment, a computer-implemented method for ultrasound imaging is provided, the method comprising:

[0007] In the ultrasound process, an array of transducers from an ultrasound probe is used to receive ultrasound images.

[0008] Select candidate regions of interest from the ultrasound images.

[0009] Each candidate region of interest includes a basic rectangular region.

[0010] In this context, the major axis of each basic rectangular region is substantially perpendicular to the potential line A and / or parallel to the potential line B, and the minor axis of each rectangular region is substantially parallel to the potential line A and / or perpendicular to the potential line B.

[0011] The basic rectangular regions do not overlap with each other, and

[0012] The basic rectangular regions are substantially parallel to each other; and

[0013] Generate a reverberation feature score for each candidate region of interest.

[0014] An aggregate score for reverberation in ultrasound images is generated based on the reverberation feature score for each candidate region.

[0015] Based on the score for reverberation and / or the aggregate score for reverberation, the presence of one or more A-line structures in the ultrasound image is estimated; and

[0016] Based on the score for reverberation and / or the aggregate score for reverberation, the intensity of one or more A-line structures in the ultrasound image is estimated.

[0017] In this way, since the major axis of the rectangular search or evaluation window is vertical, candidate structures in the ultrasound image can be searched column-wise (i.e., along each column), and then the scores in each search window are aggregated row-wise. This provides a simple and low-cost two-step algorithm that can be implemented in near real-time due to its simplicity and low computational cost.

[0018] In the example, the method further includes:

[0019] One or more B-line structures in an ultrasound image are classified based on the score for reverberation and / or the aggregate score for reverberation.

[0020] In the example, the method further includes:

[0021] Generate pixel data for the ultrasound image; and

[0022] Determine the pixel intensity of the pixel data; and

[0023] The reverberation feature score for each candidate region of interest is based on the pixel intensity within the candidate region of interest.

[0024] In the example, selecting candidate regions of interest in an ultrasound image includes:

[0025] Based on the pixel intensity in the ultrasound image, one or more B-line candidates are identified in the ultrasound image; and

[0026] Define a candidate region of interest containing each identified B-line candidate.

[0027] In the example, generating a reverberation feature score for each candidate region of interest based on pixel intensity within the region of interest includes, for each region of interest:

[0028] The pixel intensity data is averaged.

[0029] Fit the averaged pixel intensity data;

[0030] The averaged pixel intensity data is compared with the fitted averaged pixel intensity data; and

[0031] Identify one or more maximum absolute differences between the averaged pixel intensity data and the fitted averaged pixel intensity data.

[0032] In the example, the method further includes:

[0033] Identify the pleural line in the ultrasound image; and

[0034] The candidate region of interest in the ultrasound image extends from the pleural line.

[0035] In the example, the method further includes:

[0036] Generate display data for ultrasound images, and a label for at least one of the A-line or B-line structures.

[0037] In the example, the ultrasound images include a film sequence, wherein the film sequence includes multiple ultrasound images, and wherein estimating the score for reverberation in the ultrasound images includes estimating an aggregate score for the reverberation based on at least a portion of the multiple ultrasound images in the film sequence.

[0038] According to an embodiment, an ultrasound system is provided, comprising:

[0039] A processor configured to execute any of the methods disclosed herein.

[0040] In the example, the system also includes:

[0041] A display that communicates with the processor, wherein the display is configured to display the ultrasound image and markings for at least one of the A-line structure or the B-line structure.

[0042] In the example, the system also includes:

[0043] An ultrasonic transducer array that communicates with the processor and is configured to acquire the ultrasonic images.

[0044] According to an embodiment, a computer program product including instructions, which, when executed by a processor, cause the processor to perform any of the methods disclosed herein. In this disclosure, the concepts of "computer program," "software program," and "program" have the same meaning.

[0045] These and other aspects of the invention will become apparent and will be set forth with reference to the embodiments described herein. Attached Figure Description

[0046] The exemplary embodiments can be best understood by reading in conjunction with the accompanying drawings, based on the following detailed description. It should be emphasized that the various features are not necessarily drawn to scale. In fact, dimensions may be increased or decreased arbitrarily for clarity of discussion. Where applicable and feasible, the same reference numerals refer to the same elements.

[0047] Figure 1 An exemplary ultrasound system according to a representative embodiment is illustrated.

[0048] Figure 2 Another exemplary ultrasound system according to a representative embodiment is illustrated.

[0049] Figure 3A An exemplary method for A-line detection in lung ultrasound, according to a representative embodiment, is illustrated.

[0050] Figure 3B Another exemplary method for A-line detection in lung ultrasound, according to a representative embodiment, is illustrated.

[0051] Figure 4 An example of A-line and B-line detection in lung ultrasound according to a representative embodiment is illustrated.

[0052] Figure 5 The illustration shows a candidate region of interest for A-line detection in lung ultrasound according to a representative embodiment.

[0053] Figure 6 The illustration shows a candidate region of interest for A-line detection in lung ultrasound according to a representative embodiment.

[0054] Figure 7 The illustration shows the difference between candidate contours and two types of fitted data in lung ultrasound A-line detection according to a representative embodiment.

[0055] Figure 8 The illustration shows the reverberation feature score evaluated within a reduced window during A-line detection of lung ultrasound according to a representative embodiment.

[0056] Figure 9 The illustration shows a cinematic sequence hierarchy assessment of A-line detection in lung ultrasound according to a representative embodiment.

[0057] Figure 10 The illustration shows a computer system according to a representative embodiment, on which a method for detecting A-lines in lung ultrasound is implemented. Detailed Implementation

[0058] In the detailed description below, exemplary embodiments with specific details disclosed are set forth for purposes of explanation and not limitation, in order to provide a thorough understanding of embodiments according to the present teachings. However, other embodiments consistent with the present disclosure but departing from the specific details disclosed herein remain within the scope of the appended claims. Descriptions of known systems, apparatuses, materials, methods of operation, and methods of manufacture may be omitted to avoid obscuring the description of representative embodiments. Nevertheless, systems, apparatuses, materials, and methods within the scope of the present teachings are within the capabilities of those skilled in the art and may be used according to representative embodiments. It should be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. The definitions and interpretations of terminology herein supplement the technical and scientific meanings of terms commonly understood and accepted in the art of the present teachings.

[0059] It should be understood that although the terms first, second, third, etc., may be used herein to describe various elements or components, these elements or components should not be limited by these terms. These terms are only used to distinguish one element or component from another. Therefore, without departing from the teachings of the inventive concept, the first element or component discussed below may also be referred to as the second element or component.

[0060] As used in the specification and claims, the singular forms of the terms “a,” “an,” and “the” are intended to include both the singular and plural forms, unless the context clearly specifies otherwise. Additionally, when used herein, the terms “comprising” and / or “including” and / or similar terms specify the presence of the recited features, elements, and / or components, but do not exclude the presence or addition of one or more other features, elements, components, and / or groups thereof. As used herein, the term “and / or” includes any and all combinations of one or more of the associated listed items.

[0061] Unless otherwise stated, when an element or component is said to be “connected to,” “coupled to,” or “proximity to” another element or component, it will be understood that the element or component can be directly connected or coupled to the other element or component, or that there may be intermediate elements or components. That is, these and similar terms include cases where one or more intermediate elements or components may be used to connect two elements or components. However, when an element or component is referred to as being “directly connected” to another element or component, this only includes cases where two elements or components are connected to each other without any intermediate or intermediary elements or components.

[0062] In the context of this disclosure, "substantially" means that the described attribute or characteristic substantially conforms to the stated quality, form, or orientation, without requiring absolute precision. For example, "substantially rectangular" means that the shape is substantially similar to a rectangle, but some minor deviations may exist. Similarly, "substantially parallel" means that elements or parts are aligned primarily in a parallel manner, allowing for slight variations. "Substantially perpendicular" indicates that the orientation is primarily perpendicular, while allowing for slight angular deviations. Furthermore, "substantially non-overlapping" means that objects or elements primarily avoid overlapping configurations, acknowledging possible edge intersections that do not significantly obscure the uniqueness of the included entities.

[0063] This disclosure, through its various aspects, embodiments, and / or specific features or sub-components, is intended to provide one or more of the advantages specifically stated below.

[0064] As described in this paper, A-lines in lung ultrasound video sequences can be identified based on the reverberant appearance along the ultrasound beam profile. Image metrics describing the appearance of A-lines can be generated to provide an assessment of overall lung health at the cinematic sequence level. The presence and intensity of reverberation can also be used to identify B-lines. Detected A-lines and B-lines can be displayed on the ultrasound images, and results at the cinematic sequence level across all lung regions can be summarized via an overall summary screen. Such challenges could potentially be addressed through automated algorithms, especially when resources and trained professionals are limited. By identifying and characterizing the reverberant appearance on the ultrasound beam profile, A-lines and B-lines can be distinguished, and this can be used to enhance the counting of B-lines. This distinction could provide an improved ability to identify A-lines as indicators of healthy, air-filled lungs.

[0065] Figure 1 The illustration shows a system 100 for lung ultrasound A-line detection according to a representative embodiment.

[0066] Figure 1 System 100 is a system for detecting A-lines in lung ultrasound and includes components that may be provided together or may be distributed. System 100 includes an ultrasound probe 110, an ultrasound base 120, and a display 180.

[0067] The ultrasound probe 110 includes processing circuitry 115 and a transducer array 113. Processing circuitry 115 may include a memory for storing data and instructions, and an application-specific integrated circuit (ASIC) and / or processor for processing data and instructions. Transducer array 113 includes an array of transducer elements, at least a first transducer element 1131, a second transducer element 1132, and an Xth transducer element 113X. That is, the ultrasound probe 110 includes an ultrasound transducer and associated electronic components for generating an image, which is then rendered onto a display 180 for viewing by a user. The ultrasound probe 110 includes transducer array 113. Transducer array 113 converts electrical energy into sound waves, which are reflected back from human tissue, and the transducer array receives the echoes of the sound waves and converts the echoes back into electrical energy. Transducer array 113 may include dozens, hundreds, or thousands of individual transducer elements. The ultrasound probe 110 can emit a sound beam to generate an image and can detect echoes. The processing circuit 115 can process the ultrasound images captured by the transducer array 113 of the ultrasound probe 110.

[0068] The ultrasonic base 120 may include an ultrasonic trolley or a portion thereof. The ultrasonic base 120 includes a first interface 121, a second interface 122, a third interface 123, and a controller 150. Figure 6 The image depicts a computer that can be used to implement the ultrasonic base 120, but the ultrasonic base 120 may contain more than... Figure 1 or Figure 6 The number of components may vary. One or more interfaces may include ports, disk drives, wireless antennas, or other types of receiver circuitry that connect controller 150 to other electronic components. A first interface 121 connects ultrasound base 120 to ultrasound probe 110 and may include ports, antennas, and / or other types of physical components for wired or wireless communication. A second interface 1220 connects ultrasound base 120 to display 180 and may also include ports, antennas, and / or other types of physical components for wired or wireless communication. A third interface 123 is a user interface and may include buttons, keys, a mouse, a microphone, a speaker, a switch, a touchscreen, or other types of displays separate from display 180, and / or other types of physical components that allow the user to interact with ultrasound base 120, such as inputting commands and receiving outputs.

[0069] The controller 150 includes at least a memory 151 storing instructions and a processor 152 executing the instructions. The instructions stored in the memory 151 may include software programs for generating scores to estimate the presence and intensity of one or more A-line structures in an ultrasound image by generating scores for reverberation in the ultrasound image. The reverberation characteristic scores, along with markers or other types of labels for A-line and B-line structures, may be generated as display data and displayed on the display 180, superimposed on the ultrasound image, or provided adjacent to the ultrasound image. The software program in the memory 151 helps ensure that appropriate ultrasound images are obtained during the ultrasound examination before the patient leaves.

[0070] Display 180 may be located locally on ultrasound base 120 or may be remotely (e.g., wirelessly) connected to ultrasound base 120. Display 180 may be connected to ultrasound base 120 via a local wired interface such as an Ethernet cable or a local wireless interface such as a Wi-Fi connection. Display 180 may be connected to other user input devices (including a mouse, keyboard, scroll wheel, etc.) through which a user can input commands. Display 180 may be a monitor, such as a computer monitor, a display on a mobile device, an augmented reality display, a television, an electronic whiteboard, or another screen configured to display electronic images. Display 180 may also include: one or more input interfaces (e.g., those mentioned above) that can be connected to other elements or components; and an interactive touchscreen configured to display prompts to the user and collect the user's touch input. Display 180 is configured to display markings for at least one of the A-line and B-line structures in the ultrasound image.

[0071] Controller 150 may directly perform some of the operations described herein, and may indirectly perform other operations described herein. For example, controller 150 may indirectly control operations such as generating and transmitting content to be displayed on display 180. Controller 150 may directly control other operations, such as processor 152 executing instructions from memory 151 based on input received from electronic components and / or from the user via an interface, thereby performing logical operations. Therefore, when processor 152 executes instructions from memory 151, the process implemented by controller 150 may include steps that controller 150 does not directly execute.

[0072] Figure 2 The illustration shows another system for lung ultrasound A-line detection according to a representative embodiment.

[0073] exist Figure 2 In the process, the ultrasound probe 210 is connected to smartphones A and B via network 201.

[0074] Network 201 may include a local wireless network, such as a WiFi network, but network 201 may also or alternatively include wired components, such as wires that connect smartphone A and smartphone B via a USB cable.

[0075] The ultrasound probe 210 includes a transducer array 213, a lens 214, a user interface 223, a controller 250, and a wireless communication circuit 290. The transducer array 213 includes at least a first transducer element 2131, a second transducer element 2132, and an Xth transducer element 213X. The transducer array 213 may correspond to a transducer array 113 and may include an array of individually controllable transducer elements. The transducer array 213 converts electrical energy into sound waves, which are reflected back from human tissue, and the transducer array receives the echoes of the sound waves and converts the echoes back into electrical energy. The transducer array 213 may include dozens, hundreds, or thousands of individual transducer elements. The ultrasound probe 210 can emit a sound beam to generate an image and can detect echoes. The lens 214 can be used to emit the ultrasonic beam and receive the echoes of the ultrasonic beam. A user can interact with the ultrasound probe 210 using the user interface 223. Wireless communication circuitry 290 can be used to communicate with smartphones A and B via network 201. That is, ultrasound probe 210 includes an ultrasound transducer and associated electronic circuitry for generating images, which are then rendered onto the displays of smartphones A and / or B.

[0076] Smartphone A stores and runs ultrasound application 299A. Smartphone B stores and runs ultrasound application 299B. Ultrasound applications 299A and 299B can be configured to enable smartphones A and B to interact with ultrasound probe 210 via network 201. For example, ultrasound applications 299A and 299B can be configured to enable the display of ultrasound images from ultrasound probe 210. Ultrasound applications 299A and 299B can also be configured to overlay scores and labels and / or other markings on the ultrasound image display from ultrasound probe 210. Ultrasound probe 210 is configured to connect to smartphones A and / or B, which are external devices in the ultrasound procedure, via ultrasound applications 299A and / or 299B. Ultrasound images can be displayed on smartphones A and / or B, which are external devices, and the external devices are configured to display markings of at least one of the A-line and B-line structures in the ultrasound images.

[0077] The controller 250 includes at least a memory 251 for storing instructions and a processor 252 for executing those instructions. The instructions stored in the memory 251 may include a software program with a model (e.g., an artificial intelligence model) and the same or different software programs for generating a user interface on the ultrasound device. Figure 2 In this system, the user interface can be generated by instructions stored in memory 251 and can be displayed on the screen of smartphone A or smartphone B. The software program in memory 251 generates scores, labels, and / or other markings for A-lines and B-lines present (if present) in the ultrasound image.

[0078] Controller 250 may directly perform some of the operations described herein, and may indirectly perform other operations described herein. For example, controller 250 may indirectly control operations such as generating and transmitting content to be displayed on the screen of smartphone A or smartphone B. Controller 250 may directly control other operations, such as logical operations performed by processor 252 based on input received from electronic components and / or users via an interface, executing instructions from memory 251. Therefore, when processor 252 executes instructions from memory 251, the process implemented by controller 250 may include steps that controller 250 does not directly execute.

[0079] Figure 3A The illustration depicts a method for A-line detection in lung ultrasound according to a representative embodiment.

[0080] Figure 3A The method can be composed of, including Figure 1 The system 100, which is the controller of the ultrasonic base 120, performs the operation, or is controlled by... Figure 2 The ultrasonic probe 210 in the middle is used to perform the operation.

[0081] At S310, the ultrasound procedure is initiated. The ultrasound procedure may include a cinematic sequence of ultrasound images obtained using an ultrasound probe (e.g., ultrasound probe 110 or ultrasound probe 210).

[0082] At S315, ultrasound images are acquired during the ultrasound procedure. The ultrasound images at S315 may include a single ultrasound image captured at a single time point, but in the context of a movie sequence, a sequence of two or more ultrasound images captured. Each ultrasound image is acquired using a transducer array of an ultrasound probe (e.g., ultrasound probe 110 or ultrasound probe 210).

[0083] At S320, pixel data is generated based on the ultrasound images obtained from the transducer array. Ultrasound probe 110 or ultrasound probe 210 can use the transducer array to generate multi-row and multi-column pixel data for each ultrasound image.

[0084] At S330, one or more candidate structures are identified. For example, a frame may contain multiple candidate structures along the A-line. These candidate structures can be identified using an image analysis program that examines patterns in the ultrasound image.

[0085] The B-line candidate structure is detected as a local peak in the vertical direction, which can be detected based on the image intensity profile. Intensity profiles are calculated along each column of the image to generate a pattern such as... Figure 5 The graph shown in the lower left corner. Each peak in the graph corresponds to a candidate B line.

[0086] In the second step, candidate A-line structures are determined by analyzing the smaller rectangular regions of interest generated from each candidate B-line structure (local peak), as can be seen in 681B and will be discussed later. Figure 6 The A-line is characterized by a peak along the horizontal direction and can be detected as a local peak in the intensity curve for each smaller region of interest. Here, unlike above, the intensity curve is calculated along each row of the region of interest. It is worth noting that because the A-line extends horizontally in the image, the same A-line tends to appear as a local peak in multiple adjacent regions of interest.

[0087] Alternatively, candidate structures for line A can be detected directly from the image based on pixel intensity (e.g., along each row of pixels in the image) and local peaks can be found. This eliminates the need to detect candidate structures for line B before identifying candidate structures for line A.

[0088] At S340, intensity is determined. The ultrasound probe 110, ultrasound base 120, and / or ultrasound probe 210 can be configured to determine the intensity in rows and columns of pixel data. Intensity is determined for the pixel data of each candidate structure identified in the frame at S330, therefore, at S340, intensity can be determined separately and independently for the pixel data of multiple candidate structures in the current frame. Intensity can be determined in some or all rows and columns of pixel data. In some embodiments, intensity can be determined for fewer than all pixels represented by the pixel data in the ultrasound image. For example, for a smaller window in the ultrasound image. In some embodiments, candidate structures in the ultrasound image can be identified, and the intensity determined at S340 can be limited to the intensity of rows and columns of pixel data for a candidate structure or a set of such candidate structures.

[0089] At S345, the averaged pixel data is compared with the fitted data. The ultrasonic base 120 or ultrasonic probe 210 can average the pixel data for the candidate structure to generate, for example... Figure 7 Any jagged line in the image. Then, the average pixel data for all candidate structures can be fitted to generate, for example... Figure 7 A smooth, continuous line is formed. Then, the ultrasonic base 120 or ultrasonic probe can identify one or more maximum absolute differences between the average pixel data and the fitted data for each candidate structure. In other words, it is possible to perform tests on structures such as... Figure 7The large deviations between noisy lines and smooth, continuous lines are identified and classified. Comparisons can be made between the original intensity curve and the linear or nonlinear best-fit curve. The comparison at S345 is performed on the pixel data of a candidate structure, and comparisons can be performed separately and independently on the average pixel data of each candidate structure in the frame at S345.

[0090] At S350, a reverberation score is generated. The reverberation characteristic score can be, or can reflect, an image metric describing the appearance of an A-line in order to assess overall lung health. The reverberation characteristic score is a reverberation characteristic score that measures the likelihood of the presence of an A-line in an ultrasound image based on the intensity determined at S340.

[0091] In some embodiments, the reverberation feature score generated at S350 can measure the likelihood of the presence of an A-line, but it can also be generated based on an initial set of one or more potential B-line candidates detected in each ultrasound frame. B-line candidates can be identified as local peaks in the intensity profile along the lateral width direction of the ultrasound frame. For each detected candidate, a small region of interest can be defined, and the reverberation feature score can be calculated based on these regions of interest. The concept of identifying B-line candidates to assess the likelihood of the presence of an A-line is... Figure 5 As shown in and below regarding Figure 5 It was described. In contrast, Figure 6 An example of a candidate region of interest calculated from an ultrasound image frame containing many A-lines but no B-lines is shown. Therefore, based on the initial set of candidate regions for detecting A-lines, a reverberation feature score at S350 is generated.

[0092] At S355, the existence of the A-line structure is estimated. The existence of the A-line structure can be estimated based on the comparison at S345 and the reverberation feature score generated at S350.

[0093] At S360, the intensity of the A-line structure is estimated. The intensity of the A-line structure can be estimated based on a reverberant feature score, and can be similar to a confidence score, indicating the confidence level that a candidate structure is an A-line structure. For example, the intensity of the A-line structure can be assigned a score from 0 to 100 to provide greater granularity.

[0094] In S370, B-line structures are classified. The B-line structures in each frame of the movie sequence can be classified using reverberation feature scoring. Candidate structures can be classified as B-line structures based on the intensity determined by the rows and columns of pixel data for each candidate structure. In some embodiments, reverberation feature scoring can be used in conjunction with other image features to classify B-lines in each frame. Each A-line and / or B-line identified in each frame can be displayed on display 170, ultrasound probe 210, or one of smartphone A or smartphone B, for example, in each frame of the movie sequence. For example, detected A-lines and B-lines can be displayed as bounding boxes.

[0095] Figure 3A The method identifies A-lines in ultrasound images from lung ultrasound video playback and distinguishes A-lines from B-lines based on the reverberant appearance on the ultrasound beam profile.

[0096] Figure 3B The illustration shows another method for A-line detection in lung ultrasound according to a representative embodiment.

[0097] Figure 3B The method can be composed of, including Figure 1 The system 100 executes the control 150 of the ultrasonic base 120, or by... Figure 2 The ultrasonic probe 210 in the middle is used to perform the operation.

[0098] Figure 3B Methods and Figure 3A The methods are largely overlapping, the difference being that this method is performed on multiple ultrasound images within a movie sequence. At S310, the ultrasound procedure is initiated. At S315, ultrasound images are acquired during the ultrasound procedure. At S320, pixel data is generated based on the ultrasound images. At S330, one or more candidate structures are identified. At S340, intensity is determined. At S345, the averaged pixel data is compared with fitted data. At S350, a score is generated for reverberation. The reverberation characteristic score can be or reflects an image metric describing the appearance of the A-line for assessing overall lung health and can be used in… Figure 3B In this context, it provides an evaluation of the film sequence hierarchy. Figure 3B The method identifies A-lines in ultrasound images from lung ultrasound video playback and distinguishes A-lines from B-lines based on the reverberant appearance on the ultrasound beam profile.

[0099] At S355, the presence of the A-line structure is estimated. At S360, the intensity of the A-line structure is estimated. At S370, the B-line structure is classified. Each A-line and / or B-line identified in each frame can be displayed on display 170, ultrasound probe 210, or one of smartphone A or smartphone B, for example, in each frame of a movie sequence. For example, the detected A-lines and B-lines can be displayed as bounding boxes.

[0100] At S380, the aggregate score is determined. After scoring the reverberation for each ultrasonic frame at S350, the aggregate score can be determined at S380, and for... Figure 3B When a movie sequence is re-evaluated or subsequently determined, it can be considered an update to the previous aggregated score. A movie sequence-level assessment of the presence of A-lines can also be provided to the user as an overall measure of lung health. The aggregated score from S380 can be an overall score for the entire movie sequence, or a score for each frame or subset of frames within the movie sequence.

[0101] At S390, display data is generated. The display data generated at S390 can be the display data of the aggregated score determined at S380. The display data can be generated by... Figure 1 The controller 150 generates the data, which is then provided to the display 180. The display data can be generated by... Figure 2 The controller 250 generates the data and then provides it to the display of the ultrasound probe 210, or to smartphone A or smartphone B for display.

[0102] At S395, it is determined whether the ultrasound image obtained at S315 is the final ultrasound image. If the ultrasound image obtained at S315 is not the final ultrasound image (S395 = No), the method returns to S315 and obtains another ultrasound image.

[0103] If the ultrasound image obtained in S315 is the final ultrasound image (S395 = Yes), then display data is generated again in S399. The display data generated in S395 can be an aggregate score for A-line structures identified in the cinnabar sequence. Figure 3A and Figure 3B The quantification results embodied in the method can serve as input for a prediction of whether the imaged lung is normal or abnormal, and this prediction can also be output as display data at S399. In some embodiments, such predictions may be limited, for example, for portable ultrasound systems usable by end users, and can alternatively be made by a system controlled by a healthcare provider.

[0104] Figure 4 An example of A-line and B-line detection in lung ultrasound according to a representative embodiment is illustrated.

[0105] exist Figure 4 In the image, the left side shows an example of line A on the first user interface 481A, and the right side shows an example of line B on the second user interface 481B. Figure 4 The A-line and B-line in the image are from a lung ultrasound. The A-line is a sign of a healthy, air-filled lung, while the B-line is a pathological sign of fluid accumulation in the lungs.

[0106] Figure 5 The illustration shows a candidate region of interest for A-line detection in lung ultrasound according to a representative embodiment.

[0107] As described above regarding S350, the score can measure the likelihood of an A-line presence, but it can also be generated based on an initial set of one or more potential B-line candidates detected in each ultrasound frame. Figure 5 In this section, pixel data in the ultrasound frame displayed in the left-hand user interface 581A is analyzed to identify B-line candidates as local peaks in an intensity profile calculated along the lateral width of the ultrasound frame. For each detected candidate, a small region of interest can be defined as shown in user interface 581B, and reverberation feature scores are calculated based on these regions of interest. The example candidate regions of interest displayed on user interface 581B are calculated based on an ultrasound image frame containing many B-lines displayed in user interface 581A. Each B-line candidate has a fairly uniform intensity, extending vertically from the pleural line to the bottom of the image.

[0108] Figure 6 The illustration shows a candidate region of interest for A-line detection in lung ultrasound according to a representative embodiment.

[0109] exist Figure 6 In this context, the candidate region of interest is calculated based on ultrasound image frames containing many A-lines but no B-lines; therefore, the reverberation feature score at S350 is generated based on an initial set of groups that detect A-line candidates. Figure 6 In the process, candidate objects are displayed in the user interface 681A on the left, and then the region of interest is obtained for each candidate object in the user interface 681B on the right. Figure 6 The candidate regions of interest are calculated based on ultrasound image frames that contain many A-lines but no B-lines. Figure 6 The local peaks along the intensity profile for each candidate region of interest show obvious spikes.

[0110] For example, Figure 6The bounding box in the image can contain a rectangle with a width of 10 pixels and a height of 200 pixels. The intensity can be provided as a row average, for example, taking 10 pixels for each point of the height. Using 10x200 pixels as the bounding box size is just an example, and candidate bounding boxes may be smaller or larger in one or two dimensions. In other examples, the height-to-width ratio can be 2, 4, 6, 8, 10, 12, 14, 16, or 18. In other words, for example, 2 pixels wide and 4 pixels high, 20 pixels wide and 40 pixels high, 10 pixels wide and 100 pixels high, etc. Figure 6 As shown, rectangle 681B may include a major axis in the vertical direction and a minor axis in the horizontal direction.

[0111] Figure 7 The illustration shows the difference between candidate contours and two types of fitted data in lung ultrasound A-line detection according to a representative embodiment.

[0112] As described above, the reverberation appearance of each candidate region can be calculated within the corresponding candidate region of interest. Specifically, for each candidate region, the reverberation feature score is calculated as the average absolute difference between the original intensity profile at the center of the region of interest and the linear best fit of that profile. If y is considered as the intensity profile on the center line of the candidate region of interest, and y' is the linear best fit of the intensity profile, then the reverberation feature score can be calculated as the mean of the absolute differences between the intensity profile on the center line and the linear best fit of the intensity profile on the center line. Figure 7 The user interface 781A on the left shows a visual representation of the reverberation features calculated based on the linear best fit. Figure 7 The user interface 781B on the right shows a visual representation of the reverberation features calculated based on a smoothed nonlinear best fit. This can be described as a(y) instead of y' for a linear best fit, where "a" represents a smoothing operator, such as Gaussian smoothing, median smoothing, frequency / Fourier-based smoothing, or any other denoising filter. The smoothing operator "a" can be applied directly to the original intensity profile y, or it can be applied to the intensity profile y after the intensity profile signal has been preprocessed by a separate operator. Figure 7 In the two user interfaces, the jagged lines with peaks are referred to as the "A line," representing the intensity point readings from top to bottom. In both user interfaces, the center line of the three parallel lines represents the best fit, thus being linear in user interface 781A and smooth nonlinear in user interface 781B. The upper and lower lines of the three parallel lines in both user interfaces respectively indicate confidence intervals, although confidence intervals are not particularly necessary to demonstrate linear and smooth nonlinear best fits.

[0113] Instead of the length of the entire contour, a smaller evaluation window can be used. Ultrasound probe 110, ultrasound base 120, and / or ultrasound probe 210 can be configured to select regions (e.g., rows and columns of pixels) in the ultrasound image as windows and generate a score for reverberation within the window. The reverberation feature score can then be calculated based on only a portion of the contour. For example, the middle 50% of the contour can be used as the evaluation window. Furthermore, the reverberation feature score can include a weighted sum of sub-scores from different evaluation sub-windows. Evaluating within a smaller window can be used to avoid regions on the candidate contour that may contain noise peaks caused by features other than the A-line, such as portions of the pleural line that might appear above the region of interest, or organ structures that might appear at the bottom of the region of interest. Other types of fitting, metrics, window sizes, and sub-window sizes can be used to calculate feature scores reflecting the likelihood of the presence of the A-line, making these features not limited to the specific examples listed herein.

[0114] Figure 8 The illustration shows the reverberation feature score evaluated within a reduced window during A-line detection of lung ultrasound according to a representative embodiment.

[0115] Figure 8 User interface 881 shows a visual example of a smaller evaluation window. The reduced window on user interface 881 is positioned between vertical dashed lines. For example, instead of using 200 pixels as the window width, a smaller evaluation window might contain 40 pixels with the greatest intensity variation.

[0116] In other embodiments, the reverberation feature score may be calculated based on alternative metrics, such as maximum absolute difference, median absolute difference, sum of absolute differences, or a combination of these alternative measures.

[0117] Similarly, the squared difference of intensity profiles can be used instead of absolute differences and can highlight larger local peaks corresponding to larger differences. The maximum / median / sum of squared differences, or combinations of these alternative measures, can also be used.

[0118] Statistical measures can also be used instead of best-fit profiles. These measures include, for example, the standard deviation or variance of the intensity profile. These types of statistical measures can be used to avoid calculating the best-fit line or curve, thus simplifying the process.

[0119] Fast Fourier Transform (FFT) contouring can also be used, as long as the shape patterns in the reverberation contour can be evaluated in the frequency domain. For example, it is possible to... Figure 7 The contour shown is subjected to a Fast Fourier Transform (FFT) to reveal its spectrum. The peaks of line A, repeating with uniform increments, are clearly visible in the FFT spectrum and can be extracted as a metric.

[0120] Alternatively, a data-driven approach can be used, leveraging machine learning techniques to learn the exact form of the nonlinear best fit a(y) and / or to compute a metric for the reverberation score. Here, the parameters of the fitting function "a" and / or the metric are not predefined but determined empirically using the training dataset.

[0121] In some embodiments, the presence of an A-line can be detected in an image frame because the ultrasound frame may contain multiple candidates, and the A-line may span multiple candidate regions. Reverberation feature scores for each candidate can be combined to generate a single A-line score for that frame, thereby calculating an average reverberation feature score across all candidates. In other embodiments, the reverberation feature score can be calculated as the sum of the reverberation feature scores across all candidates. The reverberation feature score can indicate the likelihood of an A-line being present in the frame.

[0122] Figure 9 The illustration shows a cinematic sequence hierarchy assessment of A-line detection in lung ultrasound according to a representative embodiment.

[0123] Figure 9 The cinematic sequence-level evaluation in the A-line is based on the aggregation of individual candidates across frames, and ultimately the aggregation of the entire cinematic sequence. The results of the cinematic sequence-level evaluation will be provided to the user as an overall measure of lung health.

[0124] At position 930, candidates are obtained. At position 950, a reverberation feature score is calculated for each candidate in the frame. At position 980, an aggregate score is obtained for the frame. At position 981, an aggregate score is obtained for the video. The result of the aggregate score for the video can be output as an indicator of overall lung health. The A-line scores for each frame can be summed to produce an overall score for the movie sequence, and then the video-level score can be used to assess the likelihood and number of A-lines in the video. The presence and number of A-lines indicate the health status (i.e., lack of fluid) in the lung region. Therefore, the reverberation feature score may be helpful in assessing overall lung health.

[0125] In some embodiments, reverberation feature scores can be used in combination with other image features to identify B-lines. That is, in addition to using reverberation feature scores to detect A-lines, reverberation feature scores can also be used to detect B-lines. Unlike A-lines, which are markers of healthy, air-filled lungs, B-lines are pathological markers of pulmonary effusion. Since B-lines tend to appear in the absence of A-lines, a high reverberation feature score (indicating a higher probability of A-lines) will indicate a lower probability of B-lines, and therefore a lower probability of pulmonary effusion. A low reverberation score (indicating a higher probability of B-lines) indicates a lower probability of A-lines.

[0126] Figure 10The illustration shows a computer system according to another representative embodiment, which implements a method for detecting A-lines in lung ultrasound.

[0127] refer to Figure 4 The computer system 1000 includes a set of software instructions that can be executed to cause the computer system 1000 to perform any of the methods or computer-based functions disclosed herein. The computer system 1000 can operate as a standalone device or can be connected to other computer systems or peripheral devices, for example, using a network 1001. In an embodiment, the computer system 1000 performs logic processing based on digital signals received via an analog-to-digital converter.

[0128] In a networked deployment, computer system 1000 can operate as a server or client user computer in a server-client user network environment, or as a peer-to-peer (or distributed) computer system in a peer-to-peer (or distributed) network environment. Computer system 1000 can also be implemented as or incorporated into various devices, such as workstations including controllers, fixed computers, mobile computers, personal computers (PCs), laptops, tablets, or any other machine capable of executing a set of software instructions (sequential or otherwise) specifying the actions to be taken by the machine. Computer system 1000 can be incorporated as a device or incorporated into a device, which is then included in an integrated system containing additional devices. In embodiments, computer system 1000 can be implemented using electronic devices that provide voice, video, or data communication. Furthermore, while computer system 1000 is shown as a monolith, the term "system" should also be considered as a collection of any systems or subsystems that individually or jointly execute one or more sets of software instructions to perform one or more computer functions.

[0129] like Figure 10As shown, computer system 1000 includes processor 1010. Processor 1010 can be considered a representative example of a processor of a controller and executes instructions to implement some or all aspects of the methods and processes described herein. Processor 1010 is an article of manufacture and / or machine part. Processor 1010 is configured to execute software instructions to perform the functions described in the various embodiments herein. Processor 1010 may be a general-purpose processor or may be part of an application-specific integrated circuit (ASIC). Processor 1010 may also be a microprocessor, microcomputer, processor chip, controller, microcontroller, digital signal processor (DSP), state machine, or programmable logic device. Processor 1010 may also be logic circuitry, including a programmable gate array (PGA) such as a field-programmable gate array (FPGA), or another type of circuitry including discrete gate and / or transistor logic. Processor 1010 may be a central processing unit (CPU), a graphics processing unit (GPU), or both. Furthermore, any processor described herein may include multiple processors, parallel processors, or both. Multiple processors may be included in or coupled to a single device or multiple devices.

[0130] As used herein, the term "processor" encompasses any electronic component capable of running programs or machine-executable instructions. References to computing devices including "processor" should be interpreted as including more than one processor or processing core, such as a multi-core processor. A processor can also refer to a collection of processors within a single computer system or distributed across multiple computer systems. The term computing device should also be interpreted as including a collection or network of computing devices, each including one or more processors. A program has software instructions that are executed by one or more processors, which may be within the same computing device or distributed across multiple computing devices.

[0131] Computer system 1000 also includes main memory 1020 and static memory 1030, wherein the memories in computer system 1000 communicate with each other and with processor 1010 via bus 1008. Either or both of main memory 1020 and static memory 1030 can be considered representative examples of the memory of a controller and store instructions for implementing some or all aspects of the methods and processes described herein. The memory described herein is a tangible storage medium for storing data and executable software instructions, and is non-transient during the time the software instructions are stored. As used herein, the term "non-transient" should not be interpreted as a permanent characteristic of a state, but rather as a characteristic of a state that will persist for a period of time. The term "non-transient" specifically negates fleeting characteristics, such as the characteristics of a carrier wave or signal, or other forms of characteristics that exist only transiently at any time and place. Main memory 1020 and static memory 1030 are articles of manufacture and / or machine parts. Main memory 1020 and static memory 1030 are computer-readable media from which a computer (e.g., processor 1010) can read data and execute software instructions. Each of main memory 1020 and static memory 1030 can be implemented as one or more of random access memory (RAM), read-only memory (ROM), flash memory, electrically programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, removable disks, magnetic tapes, optical disc read-only memory (CD-ROM), digital versatile disks (DVDs), floppy disks, Blu-ray discs, or any other form of storage medium known in the art. The memory can be volatile or non-volatile, secure and / or encrypted, insecure and / or unencrypted.

[0132] “Memory” is an example of a computer-readable storage medium. Computer memory is any memory that a processor can directly access. Examples of computer memory include, but are not limited to, RAM, registers, and register files. The reference to “computer memory” or “memory” should be interpreted as potentially referring to multiple memories. Memory can be, for example, multiple memories within the same computer system. Memory can also be multiple memories distributed across multiple computer systems or computing devices.

[0133] As shown in the figure, the computer system 1000 also includes a video display unit 1050, such as a liquid crystal display (LCD), an organic light-emitting diode (OLED), a flat panel display, a solid-state display, or a cathode ray tube (CRT). Additionally, the computer system 1000 includes input devices 1060, such as a keyboard / virtual keyboard or a touch input screen or voice input with voice recognition, and cursor control devices 1070, such as a mouse or a touch input screen or touchpad. The computer system 1000 may also optionally include a disk drive unit 1080, a signal generation device 1090 (e.g., a speaker or remote control), and / or a network interface device 1040.

[0134] In one embodiment, such as Figure 10 As shown, the disk drive unit 1080 includes a computer-readable medium 1082 in which one or more sets 1084 of software instructions (software) are embedded. The set 1084 of software instructions is read from the computer-readable medium 1082 and executed by the processor 1010. Furthermore, when the processor 1010 executes the software instructions 1084, one or more steps of the methods and processes described herein are performed. In one embodiment, the software instructions 1084 reside wholly or partially within main memory 1020, static memory 1030, and / or processor 1010 during execution by the computer system 1000. Additionally, the computer-readable medium 1082 may include the software instructions 1084 or receive and execute the software instructions 1084 in response to a propagation signal, causing a device connected to the network 1001 to transmit voice, video, or data through the network 1001. The software instructions 1084 may be transmitted or received on the network 1001 via a network interface device 1040.

[0135] In one embodiment, a dedicated hardware implementation, such as an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a programmable logic array, and other hardware components, is constructed to implement one or more methods described herein. One or more embodiments described herein may use two or more specific interconnected hardware modules or devices to implement functionality, said hardware modules or devices having associated control and data signals that can communicate between and through modules. Therefore, this disclosure includes software, firmware, and hardware implementations. Nothing in this application should be construed as being implemented or achievable solely in software and not in hardware such as tangible non-transient processors and / or memory.

[0136] In this embodiment, these instructions may form part of a computer program product. The computer program product may be software containing the instructions, which can be downloaded from a server (e.g., via the Internet). Alternatively, the instructions may be stored on a suitable (non-transient) computer-readable medium, such as an optical storage medium or a solid-state medium, which may or may not be provided with other hardware, or may be part of other hardware.

[0137] According to various embodiments of this disclosure, the methods described herein can be implemented using a hardware computer system that executes software programs. Furthermore, in exemplary non-limiting embodiments, implementations may include distributed processing, component / object distributed processing, and parallel processing. Virtual computer system processing can implement one or more methods or functions as described herein, and the processors described herein can be used to support virtual processing environments.

[0138] Therefore, A-line detection in lung ultrasound can distinguish between A-lines and B-lines. A-line detection in lung ultrasound is computationally efficient and can be performed in real-time in various imaging environments, including organ structures with different A-line and B-line appearances.

[0139] Although A-line detection in lung ultrasound has been described with reference to several exemplary embodiments, it should be understood that the terms used are descriptive and illustrative, not restrictive. Modifications may be made within the scope of the appended claims, as described herein and as amended, without departing from the scope of the various aspects of A-line detection in lung ultrasound. While A-line detection in lung ultrasound has been described with reference to specific methods, materials, and embodiments, it is not intended to be limited to the disclosed details; A-line detection in lung ultrasound extends to all functionally equivalent structures, methods, and uses, such as those within the scope of the appended claims.

[0140] The illustrations of the embodiments described herein are intended to provide a general understanding of the structure of various embodiments. These illustrations are not intended to be a complete description of all elements and features of the disclosure described herein. Many other embodiments will likely be apparent to those skilled in the art after reviewing this disclosure. Other embodiments can be utilized and derived from this disclosure, allowing structural and logical substitutions and changes to be made without departing from the scope of this disclosure. Furthermore, these illustrations are representative only and may not be drawn to scale. Some scales in the illustrations may be enlarged, while others may be reduced. Therefore, this disclosure and the accompanying drawings should be considered illustrative rather than restrictive.

[0141] The foregoing description of the disclosed embodiments is provided to enable any person skilled in the art to practice the concepts described in this disclosure. Therefore, the subject matter disclosed above should be considered illustrative rather than restrictive, and the appended claims are intended to cover all such modifications, enhancements, and other embodiments falling within the scope of this disclosure. Accordingly, to the fullest extent permitted by law, the scope of this disclosure will be determined by the broadest permissible interpretation of the following claims and their equivalents, and should not be limited to or restricted by the foregoing detailed description.

[0142] Reference numerals in the claims should not be construed as limiting the claims. Measures recited in dependent claims may be advantageously combined.

Claims

1. A computer-implemented method for ultrasound imaging, the method comprising: receiving (S315), in an ultrasound procedure, an ultrasound image using a transducer array (113) of an ultrasound probe (110); selecting (S330) candidate regions of interest (581B, 681B) in the ultrasound image, wherein each candidate region of interest comprises a substantially rectangular region, wherein a long axis of each substantially rectangular region is substantially perpendicular to a potential A-line and / or parallel to a potential B-line, and wherein a short axis of each rectangular region is substantially parallel to a potential A-line and / or perpendicular to a potential B-line, wherein the substantially rectangular regions are substantially non-overlapping with each other, and wherein the substantially rectangular regions are substantially parallel to each other; and generating (S350) a reverberation feature score for each candidate region of interest, generating (S350) an aggregated score for reverberation in the ultrasound image based on the reverberation feature score for each candidate region; estimating (S355) a presence of one or more A-line structures in the ultrasound image based on the score for reverberation and / or the aggregated score for reverberation; and estimating (S360) an intensity of one or more A-line structures in the ultrasound image based on the score for reverberation and / or the aggregated score for reverberation.

2. The method of claim 1, further comprising: classifying (S370) one or more B-line structures in the ultrasound image based on the score for reverberation and / or the aggregated score for reverberation.

3. The method of claim 1 or 2, further comprising: generating (S320) pixel data for the ultrasound image; and determining (S340) pixel intensities for the pixel data; and wherein the reverberation feature score for each candidate region of interest is based on the pixel intensities within the candidate region of interest.

4. The method of claim 3, wherein, Selecting candidate regions of interest in the ultrasound image comprises: identifying one or more B-line candidates in the ultrasound image based on the pixel intensities in the ultrasound image; and defining a candidate region of interest that encompasses each identified B-line candidate.

5. The method of claim 3 or 4, wherein, Generating a reverberation feature score for each candidate region of interest based on the pixel intensities within the candidate region of interest comprises, for each region of interest: averaging the pixel intensity data; fitting the averaged pixel intensity data; comparing (S345) the averaged pixel intensity data to the fitted averaged pixel intensity data; and identifying one or more maximum absolute differences between the averaged pixel intensity data and the fitted averaged pixel intensity data.

6. The method of any one of the preceding claims, further comprising: identifying a pleural line in the ultrasound image; and wherein the candidate regions of interest in the ultrasound image extend from the pleural line.

7. The method of any one of the preceding claims, further comprising: generating (S390, S399) display data for the ultrasound image, and a marker for at least one of an A-line structure or a B-line structure.

8. The method according to any of the preceding claims, wherein, The ultrasound images comprise a cine sequence, wherein the cine sequence comprises a plurality of ultrasound images, and wherein estimating the score for reverberation in the ultrasound images comprises estimating an aggregated score for the reverberation based on at least part of the plurality of ultrasound images in the cine sequence.

9. An ultrasound system (100) comprising: a processor (152, 252) configured to perform the method according to any one of claims 1 to 8.

10. The ultrasound system according to claim 9, further comprising: a display (180) in communication with the processor, wherein the display is configured to display the ultrasound images and a marker for at least one of an A-line structure or a B-line structure.

11. The ultrasound system according to claim 9 or 10, further comprising: an ultrasound transducer array (113, 213) in communication with the processor and configured to acquire the ultrasound images.

12. A computer program product comprising instructions which, when executed by a processor, cause the processor to perform the method according to any one of claims 1 to 8.

Images