System and method for automatic estimation of maximum vertical amniotic fluid pool

Abstract

A maximum vertical amniotic fluid pool (MVP) estimation system is provided. The MVP estimation system includes a controller. The controller is configured to receive imaging data of a patient's uterus. The imaging data includes multiple frames. The controller is also configured to detect multiple amniotic fluid pool candidates from the multiple frames via an amniotic fluid pool detector. Each of the multiple amniotic fluid pool candidates has an amniotic fluid pool depth. The controller is further configured to determine multiple interpretable amniotic fluid pools from the multiple amniotic fluid pool candidates via an amniotic fluid pool quality analyzer. The controller is also configured to determine the estimated MVP depth based on the multiple interpretable amniotic fluid pools via an amniotic fluid pool depth estimator.

Description

Technical Field

[0001] This disclosure relates generally to fetal ultrasound diagnosis, and more specifically to systems and methods for automatically estimating the depth of the maximum vertical pocket of amniotic fluid in the uterus. Background Technology

[0002] Maximum vertical amniotic fluid pool (MVP) assessment is considered a reliable method for evaluating the amount of amniotic fluid surrounding the fetus as seen on ultrasound scans. Amniotic fluid volume can be an important indicator of fetal health. MVP assessment is typically performed manually by an experienced clinician, involving evaluating and measuring the maximum depth of the amniotic fluid pool excluding the umbilical cord and fetal portion. This manual assessment usually requires high-quality ultrasound scans to accurately determine the MVP depth. Summary of the Invention

[0003] This disclosure generally relates to systems and methods for automatically estimating the maximum vertical amniotic fluid pool (MVP) within the uterus. Specifically, this disclosure provides systems and methods for automatically estimating the MVP depth based on ultrasound imaging data (such as cine scan data comprising multiple frames) in a manner similar to a manual evaluation performed by an experienced clinician. The imaging data is processed by an amniotic fluid pool detector (such as a “You Only Look Once” (YOLO) neural network) to identify amniotic fluid pool candidates and evaluate their associated amniotic fluid pool depths. These amniotic fluid pool candidates are then processed by an amniotic fluid pool quality analyzer to generate an interpretable subset of amniotic fluid pools by filtering out those that would be overlooked by the clinician during manual evaluation. An amniotic fluid pool depth estimator then processes the remaining subset of interpretable amniotic fluid pools to determine the estimated MVP depth. Thus, the disclosed system enables the estimation of MVP depth based on low- to moderate-quality ultrasound scans captured by inexperienced or non-specialist ultrasound technicians (such as midwives with only basic training).

[0004] The amniotic fluid pool quality detector includes a series of filters to remove amniotic fluid pools from a subset of interpretable amniotic fluid pools. These filters include a confidence filter, an anatomical structure filter, an echogenicity filter, a continuity filter, and a pair of fan-beam intersection filters. The confidence filter processes amniotic fluid pool candidates using a deep learning model trained on historical amniotic fluid pool data, allowing the candidate most similar to the historical amniotic fluid pool to pass through, thus removing irregular or unusual amniotic fluid pools. The anatomical structure filter removes amniotic fluid pool candidates corresponding to non-uterine anatomical structures of the patient (such as their heart or bladder), as these anatomical structures may resemble amniotic fluid pools in the imaging data. The fan-beam intersection filter then removes any amniotic fluid pool candidates that appear to intersect or overlap with the edges of the fan-beam shapes of the frames. The echogenicity filter evaluates the echogenicity of the amniotic fluid pool candidates and filters out those with an echogenicity above an echogenicity threshold. The echogenicity filter also evaluates the thickness of the amniotic fluid pool candidates and removes those with a thickness below a thickness threshold. A continuity filter evaluates consecutive frames to remove amniotic fluid pool candidates present in irregular or erroneous frames. Once a subset of interpretable amniotic fluid pools is generated, the amniotic fluid pool depth estimator analyzes the depth of the interpretable amniotic fluid pools to determine the estimated MVP depth. In some examples, the estimated MVP depth corresponds to the 99th percentile of the depth of the interpretable amniotic fluid pools.

[0005] Typically, in one aspect, an MVP estimation system is provided. The MVP estimation system includes a controller. The controller is configured to receive imaging data of a patient's uterus. The imaging data includes multiple frames.

[0006] The controller is also configured to detect multiple amniotic fluid pool candidates from multiple frames via an amniotic fluid pool detector. Each of the multiple amniotic fluid pool candidates has an amniotic fluid pool depth.

[0007] The controller is also configured to identify multiple interpretable amniotic fluid pools from multiple amniotic fluid pool candidates via an amniotic fluid pool quality analyzer.

[0008] The controller is also configured to determine the estimated MVP depth based on multiple interpretable amniotic fluid pools via an amniotic fluid pool depth estimator.

[0009] As shown in the example, the amniotic fluid pool quality analyzer includes a confidence filter. The confidence filter is configured to remove one or more amniotic fluid pool candidates from a pool of candidates whose confidence scores are below a confidence threshold. Confidence scores are configured via a confidence scoring model. The confidence model can be a deep learning model. Historical amniotic fluid pool data can be used to train the confidence model.

[0010] As an example, the amniotic fluid pool quality analyzer includes an anatomical structure filter. This filter is configured to remove one or more amniotic fluid pool candidates from a pool of candidates that correspond to non-uterine anatomical structures of the patient.

[0011] According to the example, the amniotic fluid pool quality analyzer includes a fan-beam intersection filter. The fan-beam intersection filter is configured to remove one or more amniotic fluid pool candidates from a plurality of amniotic fluid pool candidates that intersect with the fan beams of corresponding frames in a plurality of frames.

[0012] As shown in the example, the amniotic fluid pool quality analyzer includes a continuity filter. The continuity filter is configured to remove one or more amniotic fluid pool candidates from a plurality of amniotic fluid pool candidates based on consecutive frames across multiple frames.

[0013] According to the example, the amniotic fluid pool quality analyzer includes an echogenicity filter. The echogenicity filter is configured to remove one or more amniotic fluid pool candidates from a plurality of amniotic fluid pool candidates whose echogenicity is greater than an echogenicity threshold. The echogenicity of one of the plurality of amniotic fluid pool candidates can be determined based on the intensity distribution corresponding to the region of interest surrounding one of the amniotic fluid pool candidates. Echogenicity can be determined by dividing the peak-to-valley values ​​of the intensity distribution by the minimum edge intensity value of the intensity distribution. The echogenicity filter can also be configured to remove one or more amniotic fluid pool candidates from a plurality of amniotic fluid pool candidates whose amniotic fluid pool thickness is less than a thickness threshold.

[0014] According to the example, the amniotic fluid pool detector is a YOLO neural network.

[0015] Based on the example, the estimated MVP corresponds to the 99th percentile amniotic fluid pool depth of multiple interpretable amniotic fluid pools.

[0016] Based on the example, the imaging data is movie scan data.

[0017] Typically, in another example, a method for MVP estimation is provided. This method includes receiving imaging data of a patient's uterus via a controller. The imaging data comprises multiple frames.

[0018] The method further includes detecting a plurality of amniotic fluid pool candidates from the plurality of frames via an amniotic fluid pool detector of the controller, wherein each of the plurality of amniotic fluid pool candidates has an amniotic fluid pool depth;

[0019] The method also includes identifying multiple interpretable amniotic fluid pools from multiple amniotic fluid pool candidates via an amniotic fluid pool quality analyzer controlled by a controller.

[0020] The method also includes an amniotic fluid pool depth estimator via a controller to determine the estimated MVP depth based on multiple interpretable amniotic fluid pools.

[0021] In various implementations, the processor or controller may be associated with one or more storage media (collectively referred to herein as "memory," such as volatile and non-volatile computer memories, including RAM, PROM, EPROM, EEPROM, floppy disk, optical disk, magnetic tape, SSD, etc.). In some implementations, the storage medium may be encoded with one or more programs that, when executed on one or more processors and / or controllers, perform at least some of the functions discussed herein. Various storage media may be fixed within the processor or controller, or may be transportable, such that one or more programs stored thereon can be loaded into the processor or controller to implement the various aspects discussed herein. The terms "program" or "computer program" are used herein in a general sense to refer to any type of computer code (e.g., software or microcode) that can be used to program one or more processors or controllers.

[0022] It should be understood that all combinations of the foregoing concepts and the additional concepts discussed in more detail below (assuming that such concepts do not contradict each other) are considered part of the inventive subject matter disclosed herein. In particular, all combinations of the claimed subject matter appearing at the end of this disclosure are considered part of the inventive subject matter disclosed herein. It should also be understood that terms expressly adopted herein that may also appear in any disclosure incorporated by reference should be given the meaning most consistent with the specific concepts disclosed herein.

[0023] These and other aspects of the various embodiments will be apparent from and illustrated with reference to the embodiments described below. Attached Figure Description

[0024] In the accompanying drawings, similar reference numerals often refer to the same parts throughout the different views. Furthermore, the drawings are not necessarily drawn to scale, but rather typically focus on illustrating the principles of various embodiments.

[0025] Figure 1 This is a diagram of the maximum vertical amniotic fluid pool estimation system based on the example.

[0026] Figure 2 This is a block diagram of the controller of the system estimated based on the maximum vertical amniotic fluid pool in the example.

[0027] Figure 3 It is based on the example. Figure 2 Block diagram of the amniotic fluid pool quality analyzer for the maximum vertical amniotic fluid pool estimation system.

[0028] Figure 4 It is based on the example. Figure 3 Block diagram of the echo filter of the amniotic fluid pool quality analyzer.

[0029] Figure 5 This is a frame of an ultrasound scan showing an amniotic fluid pool candidate detected by the amniotic fluid pool detector of the maximum vertical amniotic fluid pool estimation system, according to an example.

[0030] Figure 6 The frame is an ultrasound scan of an extended region of interest, as shown in the example, illustrating a candidate amniotic fluid pool.

[0031] Figure 7 It is based on the intensity distribution of the candidate amniotic fluid pool in the example.

[0032] Figure 8 This is a schematic diagram of the controller of the maximum vertical amniotic fluid pool estimation system.

[0033] Figure 9 yes Figure 8 A schematic diagram of the processor of the controller.

[0034] Figure 10 This is a flowchart of a method for estimating the maximum vertical amniotic fluid pool, based on an example. Detailed Implementation

[0035] This disclosure generally relates to systems and methods for automatically estimating the maximum vertical amniotic fluid pool (MVP) within the uterus. Specifically, in its various embodiments and implementations, this disclosure provides systems and methods for automatically estimating the depth of the MVP based on ultrasound imaging data (such as cine scan data comprising multiple frames) in a manner similar to a manual evaluation performed by an experienced clinician. The imaging data is processed by an amniotic fluid pool detector (such as a “You Only Look Once” (YOLO) neural network) to identify amniotic fluid pool candidates and evaluate their associated amniotic fluid pool depths. These amniotic fluid pool candidates are then processed by an amniotic fluid pool quality analyzer to generate an interpretable subset of amniotic fluid pools by filtering out those that would be overlooked by the clinician during manual evaluation. An amniotic fluid pool depth estimator then processes the remaining subset of interpretable amniotic fluid pools to determine the estimated MVP depth.

[0036] Now turn to the attached diagram. Figure 1 The illustration shows non-limiting examples of an MVP estimation system 10 according to various embodiments of the present disclosure. Figure 1 As shown, system 10 typically includes a controller 100, a display screen 200, and an ultrasound scanner 300. Clinicians or other healthcare professionals use the ultrasound scanner 300 to scan the uterus U of patient P. Figure 1In a non-limiting example, the controller 100, display screen 200, and ultrasound scanner 300 are integrated into a single device, such as an ultrasound cart. However, in other examples, one or more components of system 10 may be separate from the others. For example, display screen 200 may be embodied as a tablet computer screen or television screen configured to receive a visual representation of the imaging data 102 captured by the ultrasound scan. System 10 described in this disclosure enables inexperienced healthcare professionals or amateur ultrasound practitioners (such as minimally trained midwives) to generate an estimated MVP depth 118 by capturing low- to medium-quality ultrasound scans. The estimated MVP depth 118 generated by system 10 is interpretable, allowing clinicians to logically interpret the estimated MVP depth 118 based on the provided frame 104. Therefore, the estimated MVP depth 118 should approximately match such an estimate observed and measured by an experienced clinician.

[0037] Figure 2 A block diagram of the controller 100 of the MVP estimation system 10 is shown. (As follows) Figure 7 and 8 As shown, controller 100 typically includes a processor 125 for processing data, a memory 175 for storing data, and a transceiver 185 for sending and / or receiving data. Broadly, controller 100 is configured to receive imaging data 102 from ultrasound scanner 300 and generate an estimated MVP depth 118. Once generated, the estimated MVP depth 118 can then be displayed on display screen 200. Imaging data 102 is generated from an ultrasound scanning protocol such as cine scan. In some examples, the scanning protocol may be a 3x3 or 5x5 blind scan designed to cover most of the uterus U. Each scan includes multiple frames 104 showing different locations of the uterus.

[0038] like Figure 2 As shown, the controller 100 includes three main aspects: an amniotic fluid pool detector 106, an amniotic fluid pool quality analyzer 112, and an amniotic fluid pool depth estimator 116. The amniotic fluid pool detector 106 is configured to generate one or more amniotic fluid pool candidates 108 in each frame 104 of the imaging data 102. Figure 5 An example of amniotic fluid pool candidate 108 in frame 104 is shown. Amniotic fluid pool candidate 108 is a region in frame 104 representing the amniotic fluid within the uterus U of patient P. Figure 5 As shown, the area of ​​amniotic fluid is relatively dark, which means the scanned area is hypoechoic. Figure 5 As further illustrated, each of the amniotic fluid pool candidates 108 may be defined by an amniotic fluid pool depth 110. In some examples, the amniotic fluid pool depth 110 may be provided in centimeters.

[0039] The amniotic fluid pool detector 106 can be any model capable of identifying amniotic fluid pool candidates 108. In some examples, the amniotic fluid pool detector 106 can be a deep learning neural network model, such as the "You Only Look Once" (YOLO) neural network. The amniotic fluid pool detector 106 can be trained using historical amniotic fluid pool data 128. The historical amniotic fluid pool data 128 can include previously generated ultrasound frames paired with identified amniotic fluid pools. The amniotic fluid pool detector 106 is typically trained offline before processing new imaging data 102.

[0040] Once the amniotic fluid pool detector 106 generates amniotic fluid pool candidates 108, it provides the amniotic fluid pool candidate 108 to the amniotic fluid pool quality analyzer 112. (The remaining text appears to be unrelated and possibly machine-generated gibberish.) Figure 3 and 4 To explain in more detail, the amniotic fluid pool quality analyzer 112 includes multiple filters configured to remove amniotic fluid pool candidates 108 that would not be selected by an experienced clinician performing manual analysis. For example, filters could remove amniotic fluid pool candidates 108 that overlap with non-uterine anatomical structures of the patient (such as their bladder), amniotic fluid pool candidates 108 that are too thin, amniotic fluid pool candidates 108 that intersect with the fan-shaped bundle 136 of frame 104, etc. Thus, the amniotic fluid pool quality analyzer 112 generates multiple interpretable amniotic fluid pools 114 that would have been manually selected by an experienced clinician. Each of these interpretable amniotic fluid pools 114 has an associated amniotic fluid pool depth 110.

[0041] The amniotic fluid pool depth 110 of the interpretable amniotic fluid pool 114 is then provided to the amniotic fluid pool depth estimator 116 to determine the estimated MVP depth 118. In some examples, the estimated MVP depth 118 may simply be the maximum amniotic fluid pool depth 110 of all interpretable amniotic fluid pools 114 across all frames 104 of imaging data 102. In other examples, the amniotic fluid pool depth estimator 116 may perform statistical analysis to generate the estimated MVP depth 118. In some examples, the estimated MVP depth 118 is the 99th percentile amniotic fluid pool depth 158 of all amniotic fluid pool depths 110 across all frames 104. By using the 99th percentile value instead of the absolute maximum, the amniotic fluid pool depth analyzer 116 can filter out extreme outliers that may potentially indicate anomalies or system errors. In other examples, other types of statistical analysis may be performed to determine the estimated MVP depth 118.

[0042] Figure 3 The diagram shows Figure 2 Various aspects of the amniotic fluid pool quality analyzer 112 shown. Figure 3A non-limiting example depicts the amniotic fluid pool quality analyzer 112 as including a confidence filter 120, an anatomical structure filter 130, a first sector beam intersection filter 134a, an echogenicity filter 140, a second sector beam intersection filter 134b, and a continuity filter 138. The amniotic fluid pool quality analyzer 112 is generally configured to receive amniotic fluid pool candidates 108 generated by the amniotic fluid pool detector 106 and to determine a subset of the amniotic fluid pools 114 that can be interpreted from the candidates 108. As described above, the various filters of the amniotic fluid pool quality analyzer 112 are configured to remove amniotic fluid pool candidates 108 that would typically be overlooked by experienced clinicians manually evaluating the MVP in frame 104.

[0043] Confidence filter 120 processes amniotic fluid pool candidates 108 using confidence scoring model 126 to generate a confidence score 122 for each candidate. Confidence scoring model 126 may be a deep learning model trained on a large amount of historical amniotic fluid pool data 128 from patients. Therefore, the confidence score 122 represents the similarity of amniotic fluid pool candidate 108 to previously evaluated amniotic fluid. Confidence filter 120 may then use a confidence threshold 124 to remove amniotic fluid pool candidates 108 whose confidence score 122 is lower than the threshold 124. In some examples, instead of evaluating the confidence score 122 against a constant threshold, confidence filter 120 may remove all amniotic fluid pool candidates 108 except those with the highest (e.g., top 10 highest) confidence score 122. The remaining amniotic fluid pool candidates 108 are then output by the confidence filter 120 as a confidence filter subset 108a of the amniotic fluid pool candidates 108. The confidence filter subset 108a includes the amniotic fluid pool candidates 108 that are most similar to historical amniotic fluid pools from which irregular or unusual amniotic fluid pools have been removed.

[0044] The confidence filter subset 108a is then provided to the anatomical structure filter 130 for further filtering. The anatomical structure filter 130 is configured to remove amniotic fluid pool candidates 108 corresponding to non-uterine anatomy structures of patient P. For example, certain anatomical aspects of patient P or their fetus (such as the bladder or heart) may appear as hypoechoic areas in frame 104 of imaging data 102, resembling amniotic fluid pools. When visualized, these hypoechoic areas are shown as dark areas in frame 104. Therefore, the amniotic fluid pool detector 108 may mistake parts of the bladder or heart of patient P or the fetus for amniotic fluid pools. The anatomical structure filter 130 may receive non-uterine anatomy data 132 describing these non-uterine regions to remove amniotic fluid pool candidates 108 corresponding to these regions. The remaining amniotic fluid pool candidates 108 are then output by the anatomical structure filter 130 as an anatomical structure filter subset 108b of amniotic fluid pool candidates 108.

[0045] The anatomical structure filter subset 108b is then provided to the first sector beam intersecting filter 134a. Figure 5 The example fan beam 134 of frame 104 is shown. The first fan beam intersection filter 134b removes any amniotic fluid pool candidates 108 that appear to intersect or overlap with one or more edges of the fan beam 136. Figure 5 As shown, the upper right corner of an amniotic fluid pool candidate 108i intersects with the edge of the fan-beam 136. Such overlapping fan-beam amniotic fluid pool candidates 108i are typically overlooked by experienced clinicians because they will not be able to accurately measure the depth of the overlapping amniotic fluid pools. Therefore, the first fan-beam intersection filter 134a removes the amniotic fluid pool candidate 108i. The remaining amniotic fluid pool candidates 108 are then output by the first fan-beam intersection filter 134a as the first fan-beam filter subset 108c of the amniotic fluid pool candidates 108.

[0046] Then the first sector beam filter subset 108c is provided to the echo filter 140. (Refer to...) Figure 4 The echo filter 140 is described in more detail. In a general sense, the echo filter 140 determines the echo 142 of each of the remaining amniotic fluid pool candidates 108 and filters out amniotic fluid pool candidates 108 whose echo 142 is higher than an echo 144. The echo filter 140 may also evaluate the amniotic fluid pool thickness 154 of each of the remaining amniotic fluid pool candidates 108 and remove amniotic fluid pool candidates 108 whose amniotic fluid pool thickness 154 is lower than a thickness threshold 156. The remaining amniotic fluid pool candidates 108 are then output by the echo filter 140 as an echo filter subset 108d of the amniotic fluid pool candidates 108.

[0047] The echo filter subset 108d is then provided to the second sector beam intersecting filter 134b. As part of the echo filter 140, an expanded region of interest 148 is generated for each amniotic fluid pool candidate 108. These regions of interest 148 extend each amniotic fluid pool candidate 108 by an offset in both the vertical and horizontal directions. In some examples, this offset can be 1.5 cm. Figure 6 It shows the corresponding Figure 5 The example of candidate amniotic fluid pool 108a extends the region of interest 148a. (See reference...) Figure 4 and 6 The description describes generating an intensity distribution 146 for each extended region of interest 148 to evaluate the echogenicity 142 of the corresponding amniotic fluid pool candidate 108. A second sector beam 134b removes amniotic fluid pool candidates 108 that have extended regions of interest 148 that intersect or overlap with sector beam 136. The remaining amniotic fluid pool candidates 108 are then output by the second sector beam intersection filter 134b as a second sector beam filter subset 108e of the amniotic fluid pool candidates 108.

[0048] The second sector filter subset 108e is then provided to the continuity filter 138. The continuity filter 138 evaluates a consecutive series of frames 104 to remove amniotic fluid pool candidates 108 shown in irregular or erroneous frames 104. The continuity filter 138 can perform this analysis by evaluating consecutive frames 104 for amniotic fluid pool candidates 108 with similar amniotic fluid pool depths 110. Abnormal amniotic fluid pool candidates 108 are then removed by the continuity filter 138, and the remaining amniotic fluid pool candidates 108 are then output by the continuity filter 138 as a plurality of interpretable amniotic fluid pools 114. The interpretable amniotic fluid pools 114 are then provided to the amniotic fluid pool depth estimator 116 to generate an estimated MVP depth 118.

[0049] Figure 4 The diagram shows Figure 3 Various aspects of the echo filter 140 are shown. Figure 3 A non-limiting example depicts the amniotic fluid pool quality analyzer 112 as including an image intensifier 160, a main edge extractor 162, a region of interest expander 164, an intensity distribution extractor 166, an intensity valley filter 168, and an amniotic fluid pool thickness filter 170. The echo filter 140 is typically configured to receive a first sector filter subset 108c of amniotic fluid pool candidates 108 from the first sector filter 134a, and to determine an echo filter subset 108d of amniotic fluid pool candidates 108 to be provided to the second sector intersecting filter 134b. Like the other filters described above, the echo filter 140 is configured to remove amniotic fluid pool candidates 108 that would typically be ignored by experienced clinicians manually evaluating the MVP depth in frame 104.

[0050] Image intensifier 160 processes frame 104, which contains amniotic fluid pool candidates 108, by increasing the visual contrast between a portion of frame 104 within amniotic fluid pool candidates 108 (and thus representing amniotic fluid) and a portion of frame 104 outside of amniotic fluid pool candidates 108. Contrast is increased by darkening the amniotic fluid portion of frame 104 and brightening the other non-amniotic fluid portions. Image intensifier 160 then outputs a series of enhanced amniotic fluid pool candidates 108c1 within each frame 104.

[0051] The enhanced amniotic fluid pool candidates 108c1 are then provided to the master edge extractor 162. The master edge extractor 162 is configured to output the master edge 172 of each amniotic fluid pool candidate 108 after image enhancement.

[0052] The main edge 108c3 of each amniotic fluid pool candidate 108 is then provided to the region of interest expander 164. The region of interest expander 164 generates an expanded region of interest 148 for each amniotic fluid pool candidate 108. In some examples, the expanded region of interest 148 of each amniotic fluid pool candidate 108 surrounds a 1.5 cm area around the amniotic fluid pool candidate 108 in both the vertical and horizontal directions. Figure 6 The diagram illustrates the corresponding Figure 5 The example of the amniotic fluid pool candidate 108a is an extended region of interest 148a in frame 104.

[0053] The expanded region of interest 148 for each amniotic fluid pool candidate 108 is then provided to the intensity distribution extractor 166. The intensity distribution extractor 166 generates an intensity distribution 146 for each expanded region of interest 148. Figure 7 Example intensity distribution 146 is shown. Intensity distribution 146 depicts the intensity (or brightness) within a pixel range of the region of interest 148 in frame 104. Intensity distribution 146 representing the amniotic fluid pool without non-amniotic material or imaging processing artifacts would appear as a good or valley shape, as shown. Figure 7 As shown. The intensity distribution 146 can be defined by peak-valley value 150 and two edge intensity values ​​152a, 152b. Edge intensity values ​​152a, 152b correspond to the main edges 172a, 172b of the amniotic fluid pool candidate 108 as determined by the main edge extractor 162.

[0054] The intensity distribution 146 is then provided to the intensity valley filter 168. The intensity valley filter 168 is configured to remove amniotic fluid pool candidates 108 whose echogenicity 142 is greater than the echogenicity threshold 144. This filtering corresponds to a manual review performed by an experienced clinician who will ignore amniotic fluid pool candidates 108 with non-amniotic material or imaging artifacts. Typically, the echogenicity 142 is determined by comparing the intensity of the valleys of the intensity distribution 146 to the intensity of the edges. In one example, the echogenicity 142 is calculated by dividing the minimum edge intensity value 152 by the peak-valley value 150. Figure 7 In the example, the minimum edge intensity value 152 corresponds to the second edge 172b with a second edge intensity value 152b. Therefore, the minimum edge intensity value 152 is approximately 120 joules. Furthermore, the peak-valley value 150 is approximately 30 joules. Therefore, the echogenicity 142 of the intensity distribution 146 is approximately 0.25. In other examples, the echogenicity 142 can be determined by dividing the minimum edge intensity value 152 by a certain percentile value of the valley (such as the 70th percentile), instead of using the peak-valley value 150. Amniotic fluid pool candidates 108 with an echogenicity 142 below the echogenicity threshold 144 are then output as an echo subset 108c2.

[0055] The echo subset 108c2 is then provided to the amniotic fluid pool thickness filter 170. The amniotic fluid pool thickness filter 170 is configured to remove amniotic fluid pool candidates 108 whose amniotic fluid pool thickness 154 is less than a thickness threshold 156. This filtering corresponds to manual review performed by an experienced clinician who will ignore very thin amniotic fluid pool candidates 108. The amniotic fluid pool thickness 154 is... Figure 7 The intensity distribution 146 is illustrated as the distance between the two main edges 152a and 152b. The remaining amniotic fluid pool candidates are then output as a subset 108d of the echo filter to be provided to the second sector beam intersecting filter 134b, as shown. Figure 3 As shown.

[0056] Figure 8 This schematically illustrates the previous... Figure 1 and 2 The controller 100 is depicted in the image. The controller 100 includes a processor 125, a memory 175, and a transceiver 185. The processor 125... Figure 9 The image is shown in more detail below. Memory 175 is configured to store imaging data 102, multiple amniotic fluid pool candidates 108, estimated MVP depth 118, confidence threshold 124, historical amniotic fluid pool data 128, non-uterine anatomical data 134, fan-bundle data 136, echogenicity threshold 144, and amniotic fluid pool thickness threshold 156. Imaging data 102 includes multiple frames 104. For each amniotic fluid pool candidate 108, the candidate includes an amniotic fluid pool depth 110, a confidence score 122, echogenicity 142, an extended region of interest 148, and an intensity distribution 146 (including peak-to-trough values ​​150 and minimum edge intensity values ​​152). Amniotic fluid pool candidates 108 may also include a subset of interpretable amniotic fluid pools 114 and the 99th percentile value 158 of the amniotic fluid pool depth that interprets amniotic fluid pools 158.

[0057] Figure 9 yes Figure 8 A schematic diagram of the processor 125 of the controller 100 is shown. The processor 125 is configured to execute the amniotic fluid pool detector 106, the amniotic fluid pool quality analyzer 112, and the amniotic fluid pool depth estimator 116. The amniotic fluid pool quality analyzer 112 includes a confidence filter 120 (configured to execute a confidence scoring model 126), an anatomical structure filter 130, a first sector beam intersection filter 134a and a second sector beam intersection filter 134b, a continuity filter 138, and an echo filter 140. The echo filter 140 is configured to execute an image intensifier 160, a main edge extractor 162, a region of interest expander 164, an intensity distribution extractor 166, an intensity valley filter 168, and an amniotic fluid pool thickness filter 170.

[0058] Figure 10This is a flowchart of method 900 for estimating the maximum vertical amniotic fluid pool. (Reference) Figure 1-10 Method 900 includes receiving imaging data 102 of the uterus U of patient P via a controller in step 902. The imaging data 102 includes a plurality of frames 104.

[0059] Method 900 further includes, in step 904, detecting a plurality of amniotic fluid pool candidates 108 from a plurality of frames 104 via an amniotic fluid pool detector 106 of controller 100, wherein each of the plurality of amniotic fluid pool candidates 108 has an amniotic fluid pool depth 110.

[0060] Method 900 further includes, in step 906, determining a plurality of interpretable amniotic fluid pools 114 from a plurality of amniotic fluid pool candidates 108 via an amniotic fluid pool quality analyzer 112 of the controller 100.

[0061] Method 900 further includes, in step 908, determining an estimated MVP depth 118 based on a plurality of interpretable amniotic fluid pools 114 via an amniotic fluid pool depth estimator 116 of controller 100.

[0062] All definitions defined and used herein should be understood as control dictionary definitions, definitions in incorporated documents by reference, and / or the general meaning of the defined terms.

[0063] Unless explicitly indicated to the contrary, the words “a” and “an” as used herein in the specification and claims shall be understood to mean “at least one”.

[0064] The phrase “and / or” as used herein in the specification and claims should be understood to mean “any one or both” of the elements so combined, that is, elements that are combined in some cases and separate in others. Multiple elements listed using “and / or” should be interpreted in the same way, that is, “one or more” of the elements so combined. In addition to the elements specifically identified by the “and / or” clause, other elements may optionally be present, whether related to or unrelated to those specifically identified.

[0065] As used herein in the specification and claims, “or” should be understood to have the same meaning as “and / or” as defined above. For example, when separating items in a list, “or” or “and / or” should be interpreted as inclusive, i.e., including multiple elements or at least one element in the list of elements, but also including more than one, and optionally additional unlisted items. Only terms that explicitly indicate the opposite (such as “only one of…” or “exact one of…”, or “consisting of…” as used in the claims) will refer to multiple elements or exactly one element in the list of elements. Generally, when preceded by an exclusive term (such as “any,” “one of…,” “only one of…,” or “exact one of…”), the term “or” as used herein should be interpreted only as indicating an exclusive alternative (i.e., “one or the other but not both”).

[0066] As used herein in the specification and claims, the phrase "at least one" in relation to a list of one or more elements should be understood to mean at least one element selected from any one or more elements in the list of elements, but does not necessarily include at least one of each and every element specifically listed in the list of elements, and does not exclude any combination of elements in the list of elements. This definition also allows for the optional presence of elements other than those specifically identified in the list of elements referred to by the phrase "at least one," whether related to or unrelated to those specifically identified elements.

[0067] It should also be understood that, unless expressly indicated to the contrary, in any method claimed herein that includes more than one step or action, the order of the steps or actions of the method is not necessarily limited to the order in which the steps or actions of the method are described.

[0068] In the claims and the foregoing description, all transitional phrases (such as "comprising," "including," "carrying," "having," "containing," "involving," "holding," "consisting of," etc.) shall be understood as open-ended, meaning including but not limited to. Only the transitional phrases "consisting of" and "substantially consisting of" shall be closed or semi-closed transitional phrases, respectively.

[0069] The examples of the described topics described above can be implemented in any of a variety of ways. For example, some aspects can be implemented using hardware, software, or a combination thereof. When any aspect is implemented at least partially in software, the software code can execute on any suitable processor or collection of processors, whether provided in a single device or computer or distributed across multiple devices / computers.

[0070] This disclosure can be implemented as a system, method, and / or computer program product at any possible level of technical detail integration. The computer program product may include a computer-readable storage medium (or media) having computer-readable program instructions thereon for causing a processor to execute aspects of this disclosure.

[0071] Computer-readable storage media can be tangible devices capable of retaining and storing instructions for use by an instruction execution device. Computer-readable storage media can 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 non-exhaustive list of more specific examples of computer-readable storage media includes the following: portable computer disks, hard disks, 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 disc (DVD), memory sticks, floppy disks, mechanical encoding devices (such as punched 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 through fiber optic cables), or electrical signals transmitted through wires.

[0072] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to a suitable computing / processing device, or downloaded via a network (e.g., the Internet, a local area network, a wide area network, and / or a wireless network) to an external computer or external storage device. The network may include copper cables, optical fibers, wireless transmission, routers, firewalls, switches, gateway computers, and / 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 to a computer-readable storage medium within the suitable computing / processing device.

[0073] Computer-readable program instructions used to perform the operations of this disclosure may be assembly 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). The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a stand-alone 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 wide area network (WAN)) or may make a connection to an external computer (e.g., via the Internet using an Internet service provider). In some examples, electronic circuits including, for example, programmable logic circuits, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs) may execute computer-readable program instructions to personalize the electronic circuits in order to perform aspects of this disclosure by utilizing the status information of the computer-readable program instructions.

[0074] This document describes aspects of the disclosure with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to examples of the disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0075] Computer-readable program instructions may be provided to a processor of a 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 modules for implementing one or more functions / actions specified in flowcharts and / or block diagrams. These computer-readable program instructions may also be stored in a computer-readable storage medium that can instruct a computer, programmable data processing apparatus, and / or other device to operate in a particular manner, such that the computer-readable storage medium in which the instructions are stored includes an article of writing comprising instructions for implementing aspects of the functions / actions specified in the flowcharts and / or block diagrams or blocks.

[0076] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer-implemented process, and to cause the instructions to be executed on the computer, other programmable apparatus or other device to perform one or more functions / actions specified in a flowchart and / or block diagram box.

[0077] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various examples of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of instructions, including one or more executable instructions for implementing one or more specified logical functions. In some alternative implementations, the functions marked in the blocks may occur in a non-consecutive order. For example, two blocks shown consecutively may actually be executed substantially simultaneously, or these blocks may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified function or action or executes a combination of dedicated hardware and computer instructions.

[0078] Other embodiments are within the scope of the following claims and other claims that the applicant may enjoy.

[0079] While various examples have been described and illustrated herein, those skilled in the art will readily conceive of various other means and / or structures for performing functions and / or obtaining results and / or one or more advantages described herein, and each of such variations and / or modifications is considered to be within the scope of the examples described herein. More generally, those skilled in the art will readily understand that all parameters, dimensions, materials, and configurations described herein are exemplary, and actual parameters, dimensions, materials, and / or configurations will depend on one or more specific applications using the teachings. Those skilled in the art will recognize or be able to determine many equivalents of the specific examples described herein using only conventional experimentation. Therefore, it should be understood that the foregoing examples are presented by way of example only, and that the examples may be practiced in ways different from the specific descriptions and claims within the scope of the appended claims and their equivalents. The examples of this disclosure relate to each individual feature, system, article, material, kit, and / or method described herein. Furthermore, any combination of two or more such features, systems, articles, materials, kits, and / or methods is included within the scope of this disclosure if such features, systems, articles, materials, kits, and / or methods do not contradict each other.

Claims

1. A maximum vertical amniotic fluid pool (MVP) estimation system (10), comprising a controller (100) configured to: Receive imaging data (102) of the uterus (U) of the patient (P), wherein, The imaging data (102) includes multiple frames (104); Multiple amniotic fluid pool candidates (108) are detected from the multiple frames (104) via an amniotic fluid pool detector (106), wherein each of the multiple amniotic fluid pool candidates (108) has an amniotic fluid pool depth (110); Multiple explanatory amniotic fluid pools (114) are identified from the plurality of amniotic fluid pool candidates (108) via an amniotic fluid pool quality analyzer (112); and The estimated MVP depth (118) is determined based on the plurality of interpretable amniotic fluid pools (114) via an amniotic fluid pool depth estimator (116).

2. The MVP estimation system (10) according to claim 1, wherein, The amniotic fluid pool quality analyzer (112) includes a confidence filter (120) configured to remove one or more amniotic fluid pool candidates (108) whose confidence score (122) is less than a confidence threshold (124), and wherein the confidence score (122) is assigned via a confidence score model (126).

3. The MVP estimation system (10) according to claim 2, wherein, The confidence score model (126) is a deep learning model.

4. The MVP estimation system (10) according to claim 2, wherein, The confidence scoring model (126) is trained using historical amniotic fluid pool data (128).

5. The MVP estimation system (10) according to claim 1, wherein, The amniotic fluid pool quality analyzer (112) includes an anatomical structure filter (130) configured to remove one or more amniotic fluid pool candidates (108) from the plurality of amniotic fluid pool candidates (108) that correspond to a non-uterine anatomical structure (132) of the patient (P).

6. The MVP estimation system (10) according to claim 1, wherein, The amniotic fluid pool quality analyzer (112) includes a fan-beam intersecting filter (134) configured to remove one or more amniotic fluid pool candidates (108) from the plurality of amniotic fluid pool candidates (108) whose fan-beams (136) intersect with the corresponding frames (104) in the plurality of frames (104).

7. The MVP estimation system (10) according to claim 1, wherein, The amniotic fluid pool quality analyzer (112) includes a continuity filter (138) configured to remove one or more amniotic fluid pool candidates (108) based on consecutive frames (104) in the plurality of frames (104).

8. The MVP estimation system (10) according to claim 1, wherein, The amniotic fluid pool quality analyzer (112) includes an echo filter (140) configured to remove one or more amniotic fluid pool candidates (108) whose echo (142) is greater than an echo threshold (144).

9. The MVP estimation system (10) according to claim 8, wherein, The echogenicity (142) of one of the plurality of amniotic fluid pool candidates (108) is determined based on the intensity distribution (146) corresponding to the region of interest (148) around the one of the plurality of amniotic fluid pool candidates (108).

10. The MVP estimation system (10) according to claim 9, wherein, The echo (142) is determined by dividing the peak-valley value (150) of the intensity distribution (146) by the minimum edge intensity value (152) of the intensity distribution (146).

11. The MVP estimation system (10) according to claim 8, wherein, The echo filter (140) is also configured to remove one or more amniotic fluid pool candidates (108) whose amniotic fluid pool thickness (154) is less than a thickness threshold (156).

12. The MVP estimation system (10) according to claim 1, wherein, The amniotic fluid pool detector (106) is a "YouOnly Look Once" (YOLO) neural network.

13. The MVP estimation system (10) according to claim 1, wherein, The estimated MVP depth (118) corresponds to the 99th percentile amniotic fluid pool depth (158) of the plurality of interpretable amniotic fluid pools (114).

14. The MVP estimation system (10) according to claim 1, wherein, The imaging data (102) is cinematic scan data.

15. A method (900) for estimating the maximum vertical amniotic fluid pool (MVP), comprising: Imaging data of the patient's uterus is received (902) via a controller, wherein the imaging data includes multiple frames; The amniotic fluid pool detector of the controller detects (904) a plurality of amniotic fluid pool candidates from the plurality of frames, wherein each of the plurality of amniotic fluid pool candidates has an amniotic fluid pool depth; The amniotic fluid pool quality analyzer of the controller identifies (906) multiple interpretable amniotic fluid pools from the plurality of amniotic fluid pool candidates; and The amniotic fluid pool depth estimator via the controller determines the estimated MVP depth (908) based on the plurality of interpretable amniotic fluid pools.

Images