An ultrasound imaging method of a fetal face, an ultrasound imaging apparatus, and a medium

Abstract

This disclosure relates to an ultrasound imaging method, ultrasound imaging device, and medium for fetal facial imaging. The method includes: acquiring three-dimensional volumetric data of the fetal face; determining a three-dimensional eye region and a three-dimensional facial region of the fetus based on the acquired three-dimensional volumetric data; generating a VR image of the face and eyes based on the three-dimensional eye region and the three-dimensional facial region; determining interocular distance parameters, including intraocular and lateral interocular distances, based on the three-dimensional eye region; and presenting the determined interocular distance parameters in association with the VR image of the face and eyes. Thus, the three-dimensional eye region (and three-dimensional facial region) of the fetus can be automatically and accurately determined directly by considering the spatial correlation of voxels, and accurate interocular distance parameters can be derived accordingly for presentation in association with the three-dimensional eye region (e.g., its VR image). The three-dimensional VR image of the face can also be presented comparatively, facilitating doctors to obtain comprehensive information on facial deformities for efficient screening.

Description

Technical Field

[0001] This disclosure relates to a medical imaging method, a medical imaging device, and a medium, and more specifically, to an ultrasound imaging method, an ultrasound imaging device, and a medium. Background Technology

[0002] Ultrasound examinations are widely used in clinical practice due to their safety, convenience, lack of radiation, and low cost, becoming one of the main auxiliary tools for doctors to diagnose diseases. Prenatal ultrasound, as a primary imaging examination in prenatal care, provides important imaging evidence for measuring fetal growth and development and screening for structural abnormalities. Prenatal ultrasound examinations are now a necessary examination in early, mid, and late pregnancy.

[0003] Abnormalities in the distance and shape of the fetal eyes are ultrasound manifestations of many congenital defects, such as chromosomal abnormalities. Three-dimensional ultrasound imaging can help medical personnel quickly and intuitively screen for fetal eye malformations.

[0004] Eye malformations are common fetal defects. Fetal interocular distance is one of the ultrasound indicators for diagnosing chromosomal abnormalities and congenital diseases such as holoprosencephaly. Closely spaced eyes are often associated with holoprosencephaly, and this condition can be detected by prenatal ultrasound in 86% of cases. Furthermore, some other chromosomal abnormalities may also present with this characteristic. Widely spaced eyes are a major feature of chromosomal abnormalities, midfacial cleft syndrome, frontal encephalocele, or meningocele.

[0005] Existing methods for screening fetal eye deformities mostly involve doctors manually measuring while viewing ultrasound images. The screening results are greatly affected by the doctor's experience and technique, and the doctor's workload is also quite heavy. Summary of the Invention

[0006] Therefore, there is a need for an ultrasound imaging method, device, and medium for fetal facial features that can automatically and accurately determine the three-dimensional eye region (and the three-dimensional facial region) of the fetus directly from the three-dimensional volumetric data of the fetal face, taking into account the spatial correlation of voxels. Based on this, accurate interocular distance parameters can be derived and presented in conjunction with the three-dimensional eye (e.g., its VR image). Furthermore, according to user needs, the three-dimensional eye can be presented simultaneously with a three-dimensional VR image of the fetal face, facilitating doctors to obtain comprehensive information on facial deformities, including interocular distance parameters and three-dimensional anatomical details of the face. This enables efficient screening of various fetal facial deformities over a wider range.

[0007] According to a first aspect of this disclosure, an ultrasound imaging method for a fetal face is provided. The method may include acquiring three-dimensional volumetric data of the fetal face; determining, by a processor, a three-dimensional eye region and a three-dimensional facial region of the fetus based on the acquired three-dimensional volumetric data of the fetal face; generating, by the processor, a VR image of the face and eyes together based on the three-dimensional eye region and the three-dimensional facial region of the fetus; determining, by the processor, binocular distance-related parameters including intraocular distance and lateral distance based on the three-dimensional eye region of the fetus; and presenting, by the processor, the determined binocular distance-related parameters in association with the VR image of the face and eyes together.

[0008] According to a second aspect of this disclosure, an ultrasound imaging method for a fetal face is provided. The method may include: acquiring three-dimensional volumetric data of the fetal face; determining, by a processor, a three-dimensional eye region of the fetus based on the acquired three-dimensional volumetric data of the fetal face; determining, by the processor, binocular distance-related parameters, including intraocular distance and lateral distance, based on the three-dimensional eye region of the fetus; and presenting, by the processor, the determined binocular distance-related parameters in association with the three-dimensional eye region of the fetus.

[0009] According to a third aspect of this disclosure, an ultrasound imaging method for a fetal face is provided, characterized by comprising: acquiring three-dimensional volumetric data of the fetal face; determining, by a processor, three-dimensional eye regions of the fetus based on the acquired three-dimensional volumetric data of the fetal face; determining, by the processor, binocular distance-related parameters including intraocular distance and lateral distance based on the three-dimensional eye regions of the fetus; extracting, by the processor, a cross-section passing through the center of each eyeball based on the three-dimensional eye regions of the fetus; and presenting, by the processor, the determined binocular distance-related parameters in association with the extracted cross-section passing through the center point of each eyeball.

[0010] According to a fourth aspect of this disclosure, an ultrasound imaging apparatus for a fetal face is provided. The ultrasound imaging apparatus includes a processor configured to perform an ultrasound imaging method for a fetal face according to various embodiments of this disclosure.

[0011] According to a fifth aspect of this disclosure, a computer-readable storage medium is provided that stores computer-executable instructions thereon, which, when executed by a processor, implement an ultrasound imaging method for a fetal face according to various embodiments of this disclosure.

[0012] By utilizing the ultrasound imaging methods, apparatus, and media for fetal facial structures according to various embodiments of this disclosure, the three-dimensional eye region (and the three-dimensional facial region) of the fetus can be automatically and accurately determined directly from the three-dimensional volumetric data of the fetal face, taking into account the spatial correlation of voxels. Based on this, accurate interocular distance parameters can be derived and presented in association with the three-dimensional eye (e.g., its VR image). Furthermore, according to user needs, the three-dimensional eye can be presented in conjunction with a three-dimensional VR image of the fetal face, facilitating doctors to obtain comprehensive information on facial deformities, including interocular distance parameters and three-dimensional anatomical details of the face. This enables efficient screening of various fetal facial deformities over a wider range. Attached Figure Description

[0013] The features, advantages, and technical and industrial significance of exemplary embodiments of the present invention will now be described with reference to the accompanying drawings, wherein like reference numerals denote like elements, and wherein:

[0014] Figure 1(a) shows a structural diagram of an ultrasound imaging system for the fetal face according to an embodiment of the present disclosure;

[0015] Figure 1(b) shows a structural diagram of an example of an ultrasound imaging apparatus for the fetal face according to an embodiment of the present disclosure;

[0016] Figure 2 A flowchart illustrating a first example of an ultrasound imaging method for a fetal face according to an embodiment of the present disclosure;

[0017] Figure 3 An illustration shows the interface presented on a display using the ultrasound imaging method of the first example;

[0018] Figure 4 A flowchart illustrating a second example of an ultrasound imaging method for a fetal face according to an embodiment of the present disclosure;

[0019] Figure 5 An illustration shows the interface presented on the display using the ultrasound imaging method of the second example;

[0020] Figure 6 A flowchart illustrating a third example of an ultrasound imaging method for a fetal face according to embodiments of the present disclosure; and

[0021] Figure 7 An illustration shows the interface presented on the display using the ultrasound imaging method of the third example. Detailed Implementation

[0022] Embodiments of the invention will be described below; however, the invention is not intended to be limited to these embodiments. Not all components of these embodiments are always essential.

[0023] Figure 1(a) shows a structural diagram of a three-dimensional ultrasound imaging system for a fetal face according to an embodiment of the present disclosure. As shown in Figure 1(a), the ultrasound imaging system 100 may include a probe 101, a transmitting circuit 102 for exciting the probe 101 to emit ultrasound waves to the pregnant woman being examined, a receiving circuit 103 for controlling the probe 101 to receive ultrasound echo signals returned from the pregnant woman being examined, and a processor 104.

[0024] Various types of probes 101 can be used, such as, but not limited to, at least one of ultrasound volume probes, area array probes, and conventional ultrasound array probes (such as linear array probes, convex array probes, etc.). The doctor can move the probe 101 to select a suitable position and angle. The transmitting circuit 102 can send a set of delayed-focused pulses to the probe 101, which emits ultrasound waves along the 2D scanning plane towards the pregnant woman being examined (i.e., towards the fetus's face). The receiving circuit 103 receives the reflected ultrasound waves and converts them into electrical signals for processing by the processor 104.

[0025] In some embodiments, processor 104 may be a processing device including one or more general-purpose processing devices, such as a microprocessor, a central processing unit (CPU), a graphics processing unit (GPU), etc. More specifically, the processor may be a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a processor running other instruction sets, or a processor running a combination of instruction sets. The processor may also be one or more special-purpose processing devices or circuits, such as application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), system-on-a-chip (SoCs), etc.

[0026] In the ultrasound imaging system 100, the processor 104 can be configured to perform beamforming, 3D reconstruction, post-processing, and other processing to obtain 3D data of the fetal face. Various processing methods can be implemented using dedicated circuit or software modules. Specifically, the probe 101 emits ultrasound waves along the 2D scanning plane towards the pregnant woman being examined (i.e., towards the fetal face). After the receiving circuit 103 receives the reflected ultrasound waves, it converts them into electrical signals. The processor 104 can then perform corresponding delay and weighted summation processing on the signals obtained from multiple transmissions / receptions to achieve beamforming. Furthermore, the probe 101 can emit / receive ultrasound waves in a series of scanning planes, converting them into electrical signals. The information from these signals is integrated according to the 3D spatial relationships to achieve scanning of the fetal face in 3D space and reconstruction of the 3D image. Next, the reconstructed 3D image information of the fetal face is post-processed, such as through denoising, smoothing, and enhancement, to obtain the 3D data of the fetal face. The three-dimensional data of the fetus's face, processed by processor 104, can be displayed on display 106. Display 106 can be an LCD, CRT, or LED display.

[0027] In addition to the above processing, the processor 104 can also perform automatic detection and analysis of the fetal facial and eye regions, for example, but not limited to, performing ultrasound imaging methods of the fetal face according to various embodiments of this disclosure. Thus, the three-dimensional ultrasound imaging system of the fetal face can itself be used as an ultrasound imaging device for automatically detecting and analyzing the fetal facial and eye regions, but this is merely an example.

[0028] In some embodiments, the ultrasound imaging apparatus for automatically detecting and analyzing the facial and eye regions of the fetus can also be implemented as a separate but communicable device 100' from a three-dimensional ultrasound imaging system of the fetal face. Note that the technical term "ultrasound imaging apparatus" in this disclosure is not limited to a device that includes an ultrasound probe and transmits / receives ultrasound to form an image, but may also include a device for detecting and analyzing images derived from ultrasound, such as an image station, a remote image analysis platform, etc. For example, the device 100' may be a computer customized for image data acquisition and image data processing tasks, or a server located in the cloud.

[0029] As shown in FIG1(b), the fetal face imaging device 100' according to an embodiment of the present disclosure may include a processor 104', which may be configured to perform an ultrasound imaging method of the fetal face according to various embodiments of the present disclosure. The device 100' may include a communication interface 102' to acquire three-dimensional volumetric data of the fetal face, for example, from the three-dimensional ultrasound imaging system shown in FIG1(a), from an image database, from a PACS system, etc., which will not be elaborated here.

[0030] In some embodiments, the communication interface 102' may include a network adapter, cable connector, serial connector, USB connector, parallel connector, high-speed data transmission adapter such as fiber optic, USB 9.0, Lightning, wireless network adapter such as WiFi adapter, telecommunications (4G, LTE, 5G, etc.) adapter. The device 100' can be connected to a network via the communication interface 102'. The network can provide the functionality of a local area network (LAN), wireless network, cloud computing environment (e.g., software as a service, platform as a service, infrastructure as a service, etc.), client-server, wide area network (WAN), etc.

[0031] The hardware configuration of processor 104' can be found in Figure 1(a), and will not be described in detail here.

[0032] In some embodiments, device 100' may additionally include at least one of input / output 105' and display 106'. Input / output 105' may be configured to allow device 100' to receive and / or transmit data. Input / output 105' may include one or more digital and / or analog communication devices that allow device 100' to communicate with a user or other machines and devices. For example, input / output 105' may include a keyboard and mouse that allow the user to provide input.

[0033] As shown in FIG1(a), the device 100' may include a read-only memory (ROM), flash memory, random access memory (RAM), dynamic random access memory (DRAM) such as synchronous DRAM (SDRAM) or Rambus DRAM, static memory (e.g., flash memory, static random access memory), etc., on which computer-executable instructions are stored in any format. In some embodiments, the memory 100' may store computer-executable instructions for one or more image processing programs (such as a detection and analysis program for the fetal face and eyes), which, when executed by the processor 104', implement an ultrasound imaging method for the fetal face according to various embodiments of the present disclosure. Specifically, the computer program instructions may be accessed by the processor 104', read from the ROM or any other suitable storage location, and loaded into the RAM for execution by the processor 104'.

[0034] Figure 2 A flowchart illustrating a first example of an ultrasound imaging method for a fetal face according to an embodiment of the present disclosure is provided. The ultrasound imaging method may begin at step 201, acquiring three-dimensional volumetric data of the fetal face. At step 202, a processor may determine the three-dimensional eye region and the three-dimensional facial region of the fetus based on the acquired three-dimensional volumetric data of the fetal face; that is, automatically detect the three-dimensional eye region and the three-dimensional facial region of the fetus. Compared to detecting the eye region and the three-dimensional facial region based on a two-dimensional cross-section of the fetal face, using three-dimensional volumetric data of the fetal face can comprehensively consider the relationships between voxels in three-dimensional space, especially the relationships between voxels missing in the two-dimensional cross-section, thus more accurately determining the three-dimensional eye region and the three-dimensional facial region of the fetus. The automatically detected three-dimensional eye region of the fetus can more comprehensively reflect the spatial dimensions of the eye region. The three-dimensional eye region can take into account various deformations of the three-dimensional shape of the eye region compared to a standard sphere, thereby facilitating the acquisition of more accurate eye parameters, such as various interocular distance parameters mentioned below.

[0035] In step 203, the processor can generate a VR image of the face and eyes based on the three-dimensional eye region and the three-dimensional facial region of the fetus. This VR image of the face and eyes allows users to immerse themselves in viewing the fetus's face and eyes from various stereoscopic angles, thereby obtaining richer information for diagnosis. By limiting the presentation of the VR image to the face and eyes, while meeting the user's observation needs, it significantly reduces the computational load compared to a non-selective full-body VR image of the fetus, making it smoother for users to switch viewing angles.

[0036] In step 204, the processor can determine interocular distance parameters, including intraocular distance and interocular lateral distance, based on the three-dimensional eye region of the fetus. Automatically analyzing and deriving interocular distance parameters directly based on the three-dimensional eye region of the fetus allows for a more comprehensive consideration of the spatial dimensions of the eye region. It also eliminates the need for manual or semi-manual selection of eye profiles, reducing the user's workload and avoiding significant deviations in interocular distance parameters caused by inappropriate profile selection.

[0037] Next, in step 205, the processor can present the determined interocular distance parameters in association with the VR image of the face and eyes. This helps the user to compare and analyze the VR image of the face and eyes (immersive image information) with the determined interocular distance parameters (anatomical values), improving the stability of the user's fetal anomaly screening. For example, a wide face, low nose combined with wide interocular distance and small palpebral fissures can clearly indicate a diagnosis of Down syndrome. Conversely, excessively narrow interocular distance may indicate a diagnosis of holoprosencephaly. Furthermore, some malformations are not significantly reflected in the interocular distance parameters, but can be accurately determined by referring to the facial features, such as micrognathia.

[0038] Using the aforementioned ultrasound imaging method for fetal face, the three-dimensional eye region and the three-dimensional facial region of the fetus can be automatically determined. Based on this, a VR image of the fetal face and eyes can be presented, and the interocular distance parameters can be determined more accurately. Users can enjoy a smooth and immersive viewing experience of the VR image while comparing and analyzing the interocular distance parameters. This significantly optimizes the workflow for screening fetal facial abnormalities, thereby improving work efficiency and enhancing the stability of the acquired interocular distance parameters. Consequently, it improves the stability of the screening results for various fetal abnormalities (manifested in interocular distance parameters, manifested in the face, or manifested in a combination of interocular distance parameters and the face), promoting the widespread adoption and application of interocular distance and morphological abnormality screening.

[0039] Figure 3 An illustration shows the interface displayed on a monitor using the ultrasound imaging method of the first example. Figure 3 As shown, the intraocular distance 302b and lateral interocular distance 302a can be presented in conjunction with the VR image 301 of the face and eyeballs, providing a useful reference for users to screen for facial abnormalities. In some embodiments, the ratio of the intraocular distance 302b to the lateral interocular distance 302a, i.e., the intraocular-to-interocular distance ratio 302e, can also be presented, along with the intraocular distance 302b and the lateral interocular distance 302a, as a parameter related to interocular distance for users to use as a reference to screen for facial abnormalities. In this way, users can compare the absolute value of the intraocular and lateral interocular distances and the relative ratio of the intraocular and lateral interocular distances to more accurately and efficiently determine whether the fetus's face has abnormalities.

[0040] Based on the three-dimensional eye region of the fetus, the intraocular distance 302b and the lateral eye distance 302a can be determined in various ways. In some embodiments, paired medial boundary points and paired lateral boundary points of both eyeballs can be determined based on the three-dimensional eye region of the fetus. The distance between the paired medial boundary points of both eyeballs can be determined as the intraocular distance. The distance between the paired lateral boundary points of both eyeballs can be determined as the lateral eye distance.

[0041] In other words, once the paired inner boundary points are determined, the intraocular distance can be calculated, and once the paired outer boundary points are determined, the intraocular distance can be calculated.

[0042] Various methods can be used to obtain the inner boundary points and the outer boundary points.

[0043] For example, the distances between boundary points on the boundaries of each eye's three-dimensional eye region can be obtained by traversing the region. The two closest boundary points can be considered the inner boundary points, and the two farthest boundary points can be considered the outer boundary points. This method does not perform any idealization of the three-dimensional eye region (e.g., imagine it as a sphere), resulting in more accurate calculations of the intraocular and lateral interocular distances. However, this method of sampling, combining, and comparing a large number of points results in a high computational load.

[0044] For example, the center of each eyeball can be determined based on the three-dimensional eye region of the fetus, and the boundary points where the lines connecting the three-dimensional eye region to the centers of each eyeball are intersected can be determined as paired inner and outer boundary points. This method utilizes the near-spherical geometric property of the three-dimensional eye region, which can significantly reduce the computational load and obtain calculation results for the intraocular and lateral interocular distances with acceptable accuracy.

[0045] Various methods can be used to visualize the determined interpupillary distance parameters in conjunction with a VR image of the face and eyeballs. Specifically, such as... Figure 3 As shown, a first line 302c connecting the paired medial boundary points of both eyes and a second line 302d connecting the paired lateral boundary points of both eyes can be presented on the VR image of the face and eyeballs. The intraocular distance 302b can be presented in association with the first line 302c, and the lateral distance 302a can be presented in association with the second line 302d. Through the first line 302c and the second line 302d, the intraocular and lateral distances of both eyes are intuitively associated with the corresponding parts in the VR image, facilitating doctors to view the corresponding presentation in the VR image while examining the intraocular and lateral distances, thus enabling more efficient and accurate assessment of morphological abnormalities. Specifically, the association between each distance and its corresponding line can be achieved in various ways, for example, as shown in... Figure 3 As shown, this relationship is represented by the connection of guide lines. Alternatively, this relationship can also be represented by using the same color. Figure 3In this system, various interocular distance (IAD) parameters are displayed in blank areas of the interface. This avoids obscuring anatomical details in the VR image of the face and eyes. This is just an example; IAD parameters can also be presented as floating windows. For instance, they can initially be displayed adjacent to the corresponding anatomical structures, allowing doctors to intuitively associate the anatomical structures with the IAD parameters. After this intuitive association, the floating window can be freely moved elsewhere (e.g., to a blank area) to view anatomical details without obstruction, thus more accurately and efficiently determining whether the fetus's face is abnormal. In some embodiments, IAD parameters are presented in a proximal manner so that doctors can simultaneously compare and view various IAD parameters without shifting their gaze, reducing attentional fatigue and thus avoiding cognitive errors caused by fatigue.

[0046] Figure 4 A flowchart illustrating a second example of an ultrasound imaging method for a fetal face according to an embodiment of the present disclosure is provided. The ultrasound imaging method may begin at step 401, acquiring three-dimensional volumetric data of the fetal face. At step 402, the three-dimensional eye region of the fetus may be determined based on the acquired three-dimensional volumetric data of the fetal face. Similar to step 202, the three-dimensional eye region of the fetus is extracted and determined directly based on the three-dimensional volumetric data of the fetal face, rather than via a two-dimensional cross-section of the fetal face (regardless of whether it is derived from the three-dimensional volumetric data of the fetal face). This allows for a more comprehensive consideration of the spatial relationships between voxels, resulting in a more accurate determination of the three-dimensional eye region of the fetus. By narrowing the target region from the three-dimensional eye region and the facial region of the fetus to the three-dimensional eye region, the analysis and computational load of the three-dimensional volumetric data can be significantly reduced. Consequently, the subsequent rendering requires less computational resources, enabling the extraction and rendering of the three-dimensional eye region of the fetus to be performed smoothly even in ultrasound imaging devices with lower computational performance.

[0047] In step 403, binocular distance-related parameters, including intraocular distance and lateral distance, can be determined based on the three-dimensional eye region of the fetus. The definitions and calculation methods of binocular distance-related parameters according to various embodiments of this disclosure can be incorporated herein, and will not be elaborated upon here.

[0048] In step 404, the determined interocular distance parameters can be presented in association with the three-dimensional eye region of the fetus. In some embodiments, similar to step 203, a VR image of the eyeball can be generated based on the three-dimensional eye region of the fetus for presentation using a VR processing unit; however, the three-dimensional eye region can also be presented using other 3D rendering methods. For example, the determined interocular distance parameters can be presented only in association with the three-dimensional eye region of the fetus, significantly reducing the computational load required for rendering and, to some extent, meeting the visualization requirements of doctors for screening fetal facial abnormalities. Furthermore, when screening for fetal facial abnormalities, especially eye abnormalities, doctors sometimes desire to be able to focus on observing the three-dimensional eye region without interference from other anatomical structures. This requirement is met by presenting the determined interocular distance parameters in association with the three-dimensional eye region of the fetus. Note that presenting the determined interocular distance parameters in association with the three-dimensional eye region of the fetus can be limited to presenting only the three-dimensional eye region of the fetus and presenting the determined interocular distance parameters in association with it. It is also possible to present the three-dimensional eye region of the fetus in conjunction with the presentation of the fetus's 3D face, and to present the determined interocular distance parameters in association with the presented three-dimensional eye region.

[0049] The methods for associated presentation of binocular distance parameters and VR images of the face and eyeballs according to various embodiments of this disclosure can be combined in this way, and will not be elaborated here.

[0050] Figure 5 An illustration shows the interface displayed on a monitor using the ultrasound imaging method of the second example. (See diagram.) Figure 5 As shown, the face may not be displayed; only the three-dimensional eye region 502 of the fetus, i.e., a pair of eyeballs, may be presented. Presenting the three-dimensional eye region of the fetus allows doctors to freely change the position, angle, and zoom in / out of the three-dimensional eye region, so that doctors can focus on observing the anatomical details of the three-dimensional eye region in three-dimensional space.

[0051] like Figure 5 As shown, the intraocular distance 502b and lateral interocular distance 502a can be presented in association with the three-dimensional eye region 502, thus providing a useful reference for users to screen for facial abnormalities. In some embodiments, the ratio of the intraocular distance 502b to the lateral interocular distance 502a, i.e., the intraocular-to-interocular distance ratio 502e, can also be presented, together with the intraocular distance 502b and the lateral interocular distance 502a, as a parameter related to interocular distance for users to use as a reference to screen for facial abnormalities. In this way, users can compare the absolute value of the intraocular and lateral interocular distances and the relative ratio of the intraocular and lateral interocular distances to more accurately and efficiently determine whether the fetus's face has abnormalities.

[0052] Similar to Figure 3 Various methods can be used to visualize the determined interocular distance parameters in relation to the three-dimensional eye region 502. Specifically, such as... Figure 5 As shown, a first line 502c connecting the paired medial boundary points of both eyes and a second line 502d connecting the paired lateral boundary points of both eyes can be presented on the three-dimensional eye region 502. The intraocular distance 502b can be presented in association with the first line 502c, and the extraocular distance 302a can be presented in association with the second line 502d. Through the first line 502c and the second line 502d, the intraocular and extraocular distances are intuitively associated with the corresponding parts of the three-dimensional eye region 501, facilitating doctors to view the corresponding presentation in the three-dimensional eye region 502 (e.g., but not limited to VR images) while examining the intraocular and extraocular distances, thus enabling more efficient and accurate assessment of morphological abnormalities. Specifically, the association between each distance and its corresponding line can be achieved in various ways, for example, as shown in... Figure 5 As shown, this relationship is represented by the connection of guide lines. Alternatively, this relationship can also be represented by using the same color. Figure 5 In this example, various interocular distance (IAD) related parameters are displayed in blank areas of the interface. This avoids obscuring the anatomical details in the three-dimensional eye region 502. This is merely an example; the IAD related parameters can also be presented as floating windows. For instance, they can initially be displayed adjacent to the corresponding anatomical structures, allowing doctors to intuitively associate the anatomical structures and IAD related parameters. After this intuitive association, the floating window for the IAD related parameters can be freely moved elsewhere (e.g., to a blank area) to view the anatomical details without obstruction, thereby more accurately and efficiently determining whether the fetus's facial features are abnormal. In some embodiments, the various IAD related parameters are presented in a proximate manner so that doctors can simultaneously compare and view various IAD related parameters without shifting their gaze, reducing attentional fatigue and thus avoiding cognitive errors caused by fatigue.

[0053] Figure 6 A flowchart illustrating a third example of an ultrasound imaging method for a fetal face according to an embodiment of the present disclosure is provided. In step 601, three-dimensional volumetric data of the fetal face can be acquired. In step 602, the three-dimensional eye region of the fetus can be determined based on the acquired three-dimensional volumetric data of the fetal face. In step 603, binocular distance-related parameters, including intraocular distance and lateral distance, can be determined based on the three-dimensional eye region of the fetus. Steps 601, 602, and 603 are respectively similar to... Figure 4 Steps 401, 402, and 403 in the present disclosure, various related embodiments and descriptions (e.g., but not limited to) Figure 4The contents described herein are all applicable and will not be repeated here. The definitions and calculation methods of the interocular distance parameters according to the various embodiments of this disclosure are also applicable and will not be repeated here.

[0054] In step 604, a cross-section passing through the center of both eyeballs can be extracted based on the three-dimensional eyeball region of the fetus.

[0055] In step 605, the determined interocular distance parameters can be presented in association with the extracted profile passing through the center point of each eyeball. For example... Figure 7 As shown, a cross-section of the center point of each eyeball can be displayed, and the eyeball region can be marked (e.g., outlined with an anchor frame) on this cross-section. The lateral interocular distance (i.e., interocular distance) and medial interocular distance (i.e., interocular distance) are presented in association with the eyeball region. In some embodiments, the ratio of medial to lateral interocular distance can also be presented along with the lateral and medial interocular distances. By extracting a cross-section passing through the center of both eyeballs and presenting the determined interocular distance parameters in association with it, doctors can easily observe anatomical details on the cross-section. Compared to 3D rendering of the three-dimensional eyeball region or its combination with 3D rendering of the face, this further reduces workload, especially in ultrasound imaging devices with lower computational performance, where clear anatomical details can be presented without loss of detail due to rendering stuttering caused by 3D calculations. Furthermore, the cross-section passing through the center of both eyeballs more closely matches the actual medial and lateral interocular distances compared to other offset cross-sections, providing richer anatomical details. Note that the associated presentation of the determined interocular distance parameters with the extracted cross-sections passing through the center points of each eyeball can be combined with the method of associated presentation of interocular distance parameters with VR images of the face and eyeballs according to various embodiments of this disclosure, which will not be elaborated here. Specifically, the associated presentation of the determined interocular distance parameters with the extracted cross-sections passing through the center points of each eyeball may include: presenting a first line connecting the paired medial boundary points of the eyes and a second line connecting the paired lateral boundary points of the eyes on the cross-sections passing through the center points of each eyeball; the intraocular distance is presented in association with the first line, and the interocular distance is presented in association with the second line.

[0056] The following details how to determine the three-dimensional eye region and the three-dimensional facial region of the fetus based on the acquired three-dimensional volumetric data of the fetal face. Determining the three-dimensional eye region and the three-dimensional facial region of the fetus based on the acquired three-dimensional volumetric data of the fetal face can be achieved through any one or a combination of the following: For example, image features can be extracted based on the three-dimensional volumetric data of the fetal face, and a trained regression model can be used based on the extracted image features to determine the three-dimensional eye region and / or the three-dimensional facial region of the fetus. Alternatively, a trained segmentation model can be used based on the three-dimensional volumetric data of the fetal face to determine the three-dimensional eye region and / or the three-dimensional facial region of the fetus. Yet another example is that the three-dimensional eye region and / or the three-dimensional facial region of the fetus can be determined by matching representative data of the three-dimensional volumetric data of the fetal face with a reference template of that representative data.

[0057] The above methods will be illustrated below using the determination of the fetal three-dimensional eye region based on the acquired three-dimensional volumetric data of the fetal face as an example. However, it should be noted that those skilled in the art, upon learning of the methods for determining the fetal three-dimensional eye region, can also apply them to determining the fetal facial region, which will not be elaborated upon here.

[0058] Regression method

[0059] Image features can be extracted based on the three-dimensional volume data of the fetal face, and the three-dimensional eye region of the fetus can be determined using a trained regression model based on the extracted image features.

[0060] In some embodiments, traditional image processing or deep learning methods can be used first to extract image features from the three-dimensional data of the fetal face; then, based on the extracted image features, traditional machine learning or deep learning methods can be used to regress the position and orientation of the eye region. Here, regression refers to learning an optimal mapping function from the image features of the three-dimensional volume data of the fetal face to the three-dimensional eye region, minimizing the error between the position of the three-dimensional eye region obtained from the three-dimensional volume data of the fetal face via this mapping function and the actual position of the three-dimensional eye region.

[0061] Various methods can be used to extract image features from three-dimensional volumetric data of the fetal face, including, but not limited to, traditional image processing methods and deep learning methods. Traditional image processing methods include image feature extraction, such as SIFT features, gradient features, texture features like LBP, PCA, LDA, Haar features, HOG and LOG features, as well as image edge extraction, such as edge extraction using the Canny operator. Deep learning methods, for example, involve training a neural network model to complete one or more specific tasks, such as regressing the location of the eye region, identifying key anatomical structures and / or key points of the fetal face, and then extracting the output of one or more intermediate network nodes from the trained neural network as the extracted image features.

[0062] In some embodiments, the regression method for the location of the three-dimensional eye region can include traditional machine learning methods and deep learning methods. Before using image features from the three-dimensional volumetric data of the fetal face to regress the location of the fetal three-dimensional eye region, a fetal facial ultrasound database can be established, where each data entry can include the three-dimensional volumetric data of the fetal face and / or its image features, as well as the location of the three-dimensional eye region. During the training of the regression model, an optimal mapping function is sought from the image features of the three-dimensional volumetric data of the fetal face to the location of the fetal three-dimensional eye region, minimizing the error between the location of the three-dimensional eye region obtained by mapping the image features of the fetal face and the actual location of the three-dimensional eye region. Using this mapping function, the location of the fetal three-dimensional eye region can be predicted based on the image features of the fetal face. Among various regression methods, traditional machine learning methods can include Support Vector Machine (SVM), least squares method, logistic regression, etc.; their mapping functions can include linear functions, polynomial functions, logistic functions, etc. In deep learning methods, deep neural networks can be used as the mapping function, including Convolutional Neural Networks (CNN), Multilayer Perceptrons (MLP), Recurrent Neural Networks (RNN), etc.

[0063] Segmentation methods

[0064] Based on the three-dimensional volumetric data of the fetal face, a trained segmentation model can be used to determine the three-dimensional eye region of the fetus.

[0065] In some embodiments, the segmentation method may include deep learning-based methods and machine learning methods combined with traditional image processing methods, etc. First, it may be necessary to construct a database of ultrasound images, where each image precisely marks the boundary range of the anatomical structure of the three-dimensional eye region in the three-dimensional volume data.

[0066] Deep learning-based image segmentation methods include learning features from a constructed database and the boundaries of the anatomical structures of the three-dimensional eye region by stacking convolutional and deconvolutional layers. For example, for an input image, a deep learning network can directly generate an image mask of the same size, where pixel values ​​can represent whether it is a three-dimensional eye region, thus representing the specific boundary range of the anatomical structures of the three-dimensional eye region. In some embodiments, the deep learning network used for segmentation may include FCN (Fully Convolutional Neural Network), Unet, SegNet, DeepLab, Mask RCNN, etc.

[0067] Other machine learning-based image segmentation methods include: first, pre-segmenting the image using traditional image processing methods such as thresholding, Snake, level set, GraphCut, ASM, and AAM to obtain a set of candidate eyeball anatomical structure boundary ranges in the ultrasound image; then, extracting features from the boundary range of each candidate eyeball, using traditional features such as PCA, LDA, HOG, Haar, and LBP, or features extracted by neural networks; finally, matching the extracted features with the features extracted from the labeled eyeball anatomical structure boundary ranges in the database, and classifying them using discriminators such as KNN, SVM, random forest, or neural networks to determine whether the current candidate eyeball boundary range contains key anatomical structures.

[0068] Matching methods

[0069] The three-dimensional eye region of the fetus can be determined by matching representative data of the three-dimensional volumetric data of the fetal face with a reference template of the representative data.

[0070] In some embodiments, representative data of the three-dimensional volumetric data of the fetal face may include at least one of the three-dimensional volumetric data of the fetal face itself, data of the two-dimensional cross-section of the fetal face, and data of key parts of the fetal face.

[0071] In some embodiments, the two-dimensional cross-section of the fetal face is a two-dimensional cross-section passing through the center of each eyeball. Using a transcentral two-dimensional cross-section is more suitable for measuring interocular distance when matching templates.

[0072] Additionally or alternatively, the key parts of the fetal face may include key anatomical structures and / or key points of the fetal face. Template matching of key anatomical structures and / or key points can significantly reduce the computational load compared to template matching of surfaces and volumes, while still obtaining a relatively accurate three-dimensional eye region.

[0073] In this implementation scheme, the reference template for representative data may include a template for three-dimensional volumetric data of the fetal face, a template for two-dimensional cross-sections of the fetal face, reference data for key anatomical structures of the fetal face, and reference data for key points of the fetal face. The template for the two-dimensional cross-section of the fetal face can be selected from the cross-section most suitable for measuring interocular distance, such as, but not limited to, a two-dimensional cross-section passing through the center of each eyeball. Matching the representative data of the acquired three-dimensional volumetric data of the fetal face with the reference template may include matching with the various templates / reference data mentioned above.

[0074] The following example illustrates the matching of acquired 3D volumetric data of the fetal face with a reference template of the same data. An optimal 3D spatial transformation can be sought to maximize or minimize the similarity between the acquired 3D volumetric data of the fetal face and the reference template. Alternatively, image features (such as gradient features, LBP texture features, Haar features, HOG / LOG features, etc.) can be extracted from both the acquired 3D volumetric data of the fetal face and the reference template, respectively. Then, an optimal 3D spatial transformation can be found to maximize or minimize the similarity between the extracted image features. After matching, the position of the fetal eye region in the 3D volumetric data of the fetal face can be obtained, for example, through the inverse transformation of the aforementioned 3D spatial transformation, based on the position of the fetal eye region in the reference template of the 3D volumetric data of the fetal face.

[0075] In some embodiments, the acquired three-dimensional volumetric data of the fetal face can be matched with a reference template of a two-dimensional cross-section of the fetal face. An optimal two-dimensional cross-section can be found in the acquired three-dimensional volumetric data of the fetal face, such that the two-dimensional cross-section has the highest similarity or lowest difference with the reference template of the two-dimensional cross-section of the fetal face, or that the two-dimensional cross-section has the highest similarity or lowest difference in the image features (such as gradient features, LBP texture features, Haar features, HOG / LOG features, etc.) extracted from the reference template of the two-dimensional cross-section of the fetal face. After matching, the corresponding position of the fetal eye region in the three-dimensional volumetric data of the fetal face can be obtained based on the position of the fetal eye region in the reference template of the two-dimensional cross-section of the fetal face.

[0076] In some embodiments, the acquired three-dimensional volumetric data of the fetal face can also be matched with baseline data of key anatomical structures of the fetal face (e.g., but not limited to, nose, bridge of the nose, jaw, eyeballs, etc.). An optimal image patch can be found in the acquired three-dimensional volumetric data of the fetal face, such that the similarity between the image patch and the baseline data of the key anatomical structures of the fetal face is the highest or the difference is the lowest, or the similarity between the image features extracted from each patch is the highest or the difference is the lowest. Specifically, matching the acquired three-dimensional volumetric data of the fetal face with the baseline data of the key anatomical structures of the fetal face also includes using object detection methods such as Faster R-CNN, Mask R-CNN, SSD, YOLO, Retinanet, Efficientnet, Cornernet, Centernet, FCOS, etc., to detect candidate regions of key anatomical structures of the fetal face in the acquired three-dimensional volumetric data of the fetal face, and then matching the candidate regions with the baseline data of the key anatomical structures of the fetal face. The matching method may include finding an optimal candidate key anatomical structure for the fetal face that has the highest similarity or lowest difference with the baseline data of key anatomical structures for the fetal face; or extracting image features from both the candidate key anatomical structure and the baseline data of key anatomical structures for the fetal face, and then finding an optimal candidate key anatomical structure that has the highest similarity or lowest difference with the image features of the baseline data of key anatomical structures for the fetal face. In some embodiments, an optimal candidate key anatomical structure and an optimal spatial transformation may also be sought, such that the optimal candidate key anatomical structure, under the optimal spatial transformation, has the lowest spatial difference from the baseline data of key anatomical structures for the fetal face. After matching is completed, a region position relative to these key anatomical structures can be determined based on the relative position of the eye region of the fetal face relative to these key anatomical structures, as the detection result of the eye region of the fetal face. Alternatively, the corresponding position of the eye region can be obtained based on the position of the eye region in the baseline template of the three-dimensional volume data of the fetal face, according to the obtained optimal spatial transformation.

[0077] In some embodiments, the acquired 3D volumetric data of the fetal face can also be matched with reference data of key points in the eye region of the fetal face. An optimal point can be found in the acquired 3D volumetric data of the fetal face, such that the image features near that point have the highest similarity or lowest difference with the image features of the reference data of key points in the eye region of the fetal face. Matching the acquired 3D volumetric data of the fetal face with the reference data of key points in the eye region of the fetal face can also be done using feature point extraction methods (such as SIFT), corner detection methods (such as Harris method), or neural network methods to predict point coordinates or the region where the point is located. Candidate key points in the eye region can then be detected in the acquired 3D volumetric data of the fetal face and matched with the reference data of key points in the eye region of the fetal face. The matching method can include finding an optimal candidate key point such that the image features near it have the highest similarity or lowest difference with the image features of the reference data of key points in the eye region. Matching methods can also include finding optimal candidate keypoints for the fetal face and an optimal spatial transformation, such that the optimal candidate keypoints, under the influence of this optimal spatial transformation, have the smallest spatial difference between their positions and the reference positions of keypoints in the eye region of the fetal face. After matching, a region position relative to these keypoints can be determined based on their positions, serving as the detection result for the eye region of the face; alternatively, the position of the eye region in the 3D volumetric data of the face can be obtained by using the obtained optimal spatial transformation (e.g., inverse transformation) based on the position of the eye region in the reference template of the 3D volumetric data of the face.

[0078] Furthermore, although illustrative embodiments are described herein, the scope includes any and all embodiments having equivalent elements, modifications, omissions, combinations (e.g., combinations of schemes across various embodiments), adjustments, or variations based on this disclosure. Elements in the claims are to be interpreted broadly based on the language used in the claims, and not limited to the examples described in this specification or during the duration of this application, which are to be interpreted as non-exclusive. Moreover, the steps of the disclosed methods can be modified in any way, including by reordering steps or inserting or deleting steps. Therefore, the description is intended to be merely illustrative, and the true scope is indicated by the following claims and their full equivalents.

[0079] The above description is intended to be illustrative and not restrictive. For example, the above examples (or one or more of them) can be used in combination with each other. Other embodiments can be used by those skilled in the art when reviewing the above description. Moreover, in the detailed description above, various features can be combined together to simplify this disclosure. This should not be construed as an intention to make unclaimed disclosed features essential to any claim. Rather, the subject matter of the invention may lie in a combination of fewer features than all the features of a disclosed embodiment. Therefore, the following claims are thus incorporated into the detailed description as examples or embodiments, wherein each claim is an independent, separate embodiment, and it is contemplated that these embodiments can be combined with each other in various combinations or substitutions. The scope of the invention should be determined by reference to the appended claims and the full scope of the equivalents given to these claims.

Claims

1. A method for ultrasound imaging of the fetal face, characterized in that, include: Obtain three-dimensional volumetric data of the fetal face; The processor determines the three-dimensional eye region and the three-dimensional facial region of the fetus based on the acquired three-dimensional volumetric data of the fetal face. The processor generates a VR image of the face and eyes based on the three-dimensional eye region and the three-dimensional facial region of the fetus. The processor automatically performs three-dimensional spatial analysis based directly on the three-dimensional eye region of the fetus to obtain binocular distance-related parameters, including the intraocular distance and the lateral distance. The automatic three-dimensional spatial analysis based directly on the three-dimensional eye region of the fetus to obtain binocular distance-related parameters, including the intraocular distance and the lateral distance, includes: determining paired medial boundary points and / or paired lateral boundary points of the two eyeballs based on the three-dimensional eye region of the fetus, determining the distance between the paired medial boundary points of the two eyeballs as the intraocular distance, and / or determining the distance between the paired lateral boundary points of the two eyeballs as the lateral distance. as well as The processor then presents the determined interpupillary distance parameters in association with a VR image of the face and eyeballs.

2. The ultrasound imaging method according to claim 1, characterized in that, Determining paired medial boundary points and / or paired lateral boundary points of both eyeballs based on the three-dimensional eye region of the fetus specifically includes: The center of each eyeball is determined based on the three-dimensional eye region of the fetus; and Determine the boundary points on the boundary of the three-dimensional eye region that intersect the lines connecting the centers of each eyeball, and use these points as paired inner boundary points and paired outer boundary points; or The two closest boundary points on the boundary of the three-dimensional eye region are determined as a pair of inner boundary points, and the two farthest boundary points on the boundary of the three-dimensional eye region are determined as a pair of outer boundary points.

3. The ultrasound imaging method as described in claim 2, characterized in that, in: The distance between boundary points on the respective boundaries of the three-dimensional eye regions of both eyes is obtained by traversing them, thereby determining the two closest boundary points as the inner boundary points and determining the two farthest boundary points as the outer boundary points.

4. The ultrasound imaging method according to claim 1, characterized in that, Other parameters related to interocular distance include the ratio of the intraocular distance to the lateral distance between the two eyes.

5. The ultrasound imaging method according to claim 1, characterized in that, The determined interpupillary distance parameters, presented in association with the VR image of the face and eyes, specifically include: The VR image of the face and eyeballs shows a first line connecting the paired inner boundary points of the eyes and a second line connecting the paired outer boundary points of the eyes. The first line represents the intraocular distance, and the second line represents the lateral distance.

6. The ultrasound imaging method according to claim 1, characterized in that, The three-dimensional eye region and the three-dimensional facial region of the fetus are determined based on the acquired three-dimensional volumetric data of the fetal face through any one or a combination of the following: Image features are extracted based on the three-dimensional volumetric data of the fetal face, and based on the extracted image features, a trained regression model is used to determine the three-dimensional eye region and / or the three-dimensional facial region of the fetus; and / or Based on the three-dimensional volumetric data of the fetal face, a trained segmentation model is used to determine the three-dimensional eye region and / or the three-dimensional facial region of the fetus; and / or The three-dimensional eye region and / or the three-dimensional facial region of the fetus are determined by matching representative data of the three-dimensional volumetric data of the fetus's face with a reference template of the representative data.

7. The ultrasound imaging method according to claim 6, characterized in that, The representative data of the three-dimensional volumetric data of the fetal face includes at least one of the three-dimensional volumetric data of the fetal face itself, the data of the two-dimensional cross-section of the fetal face, and the data of the key parts of the fetal face.

8. The ultrasound imaging method according to claim 7, characterized in that, The two-dimensional cross-section of the fetal face is a two-dimensional cross-section passing through the center of each eyeball; and / or the key parts of the fetal face include the key anatomical structures and / or key points of the fetal face.

9. A method for ultrasound imaging of the fetal face, characterized in that, include: Obtain three-dimensional volumetric data of the fetal face; The processor determines the three-dimensional eye region of the fetus based on the acquired three-dimensional volumetric data of the fetal face. The processor automatically performs three-dimensional spatial analysis based directly on the three-dimensional eye region of the fetus to obtain binocular distance-related parameters, including the intraocular distance and the lateral distance. The automatic three-dimensional spatial analysis based directly on the three-dimensional eye region of the fetus to obtain binocular distance-related parameters, including the intraocular distance and the lateral distance, includes: determining paired medial boundary points and / or paired lateral boundary points of the two eyeballs based on the three-dimensional eye region of the fetus, determining the distance between the paired medial boundary points of the two eyeballs as the intraocular distance, and / or determining the distance between the paired lateral boundary points of the two eyeballs as the lateral distance. as well as The processor presents the determined interocular distance parameters in association with the three-dimensional eye region of the fetus.

10. The ultrasound imaging method according to claim 9, characterized in that, Determining the paired medial boundary points and paired lateral boundary points of both eyeballs based on the three-dimensional eye region of the fetus specifically includes: The center of each eyeball is determined based on the three-dimensional eye region of the fetus; and Determine the boundary points on the boundary of the three-dimensional eye region that intersect the lines connecting the centers of each eyeball, and use these points as paired inner boundary points and paired outer boundary points; or The two closest boundary points on the boundary of the three-dimensional eye region are determined as a pair of inner boundary points, and the two farthest boundary points on the boundary of the three-dimensional eye region are determined as a pair of outer boundary points.

11. The ultrasound imaging method as described in claim 10, characterized in that, in: The distance between boundary points on the respective boundaries of the three-dimensional eye regions of both eyes is obtained by traversing them, thereby determining the two closest boundary points as the inner boundary points and determining the two farthest boundary points as the outer boundary points.

12. The ultrasound imaging method according to claim 9, characterized in that, Other parameters related to interocular distance include the ratio of the intraocular distance to the lateral distance between the two eyes.

13. The ultrasound imaging method according to claim 9, characterized in that, The determined interocular distance parameters, presented in association with the three-dimensional eye region of the fetus, specifically include: Presents the first line connecting the paired inner boundary points of both eyes and the second line connecting the paired outer boundary points of both eyes. The first line represents the intraocular distance, and the second line represents the lateral distance.

14. The ultrasound imaging method according to claim 9, characterized in that, The three-dimensional eye region of the fetus is determined based on the acquired three-dimensional volumetric data of the fetal face through any one or a combination of the following: Image features are extracted based on the three-dimensional volume data of the fetal face, and the three-dimensional eye region of the fetus is determined using a trained regression model based on the extracted image features. and / or Based on the three-dimensional volumetric data of the fetal face, the three-dimensional eye region of the fetus is determined using a trained segmentation model. and / or The three-dimensional eye region of the fetus is determined by matching representative data of the three-dimensional volumetric data of the fetal face with a reference template of the representative data.

15. The ultrasound imaging method according to claim 14, characterized in that, The representative data of the three-dimensional volumetric data of the fetal face includes at least one of the three-dimensional volumetric data of the fetal face itself, the data of the two-dimensional cross-section of the fetal face, and the data of the key parts of the fetal face.

16. The ultrasound imaging method according to claim 15, characterized in that, The two-dimensional cross-section of the fetal face is a two-dimensional cross-section passing through the center of each eyeball; and / or the key parts of the fetal face include the key anatomical structures and / or key points of the fetal face.

17. A method for ultrasound imaging of the fetal face, characterized in that, include: Obtain three-dimensional volumetric data of the fetal face; The processor determines the three-dimensional eye region of the fetus based on the acquired three-dimensional volumetric data of the fetal face. The processor automatically performs three-dimensional spatial analysis based directly on the three-dimensional eye region of the fetus to obtain binocular distance-related parameters, including the intraocular distance and the lateral distance. The automatic three-dimensional spatial analysis based directly on the three-dimensional eye region of the fetus to obtain binocular distance-related parameters, including the intraocular distance and the lateral distance, includes: determining paired medial boundary points and / or paired lateral boundary points of the two eyeballs based on the three-dimensional eye region of the fetus, determining the distance between the paired medial boundary points of the two eyeballs as the intraocular distance, and / or determining the distance between the paired lateral boundary points of the two eyeballs as the lateral distance. The processor extracts a cross-section passing through the center of both eyeballs based on the three-dimensional eye region of the fetus; and The processor then presents the determined interocular distance parameters in association with the extracted cross-sections passing through the center points of each eyeball.

18. The ultrasound imaging method according to claim 17, characterized in that, Determining the paired medial boundary points and paired lateral boundary points of both eyeballs based on the three-dimensional eye region of the fetus specifically includes: The center of each eyeball is determined based on the three-dimensional eye region of the fetus; and Determine the boundary points on the boundary of the three-dimensional eye region that intersect the lines connecting the centers of each eyeball, and use these points as paired inner boundary points and paired outer boundary points; or The two closest boundary points on the boundary of the three-dimensional eye region are determined as a pair of inner boundary points, and the two farthest boundary points on the boundary of the three-dimensional eye region are determined as a pair of outer boundary points.

19. The ultrasound imaging method as described in claim 18, characterized in that, in: The distance between boundary points on the respective boundaries of the three-dimensional eye regions of both eyes is obtained by traversing them, thereby determining the two closest boundary points as the inner boundary points and determining the two farthest boundary points as the outer boundary points.

20. The ultrasound imaging method according to claim 17, characterized in that, Other parameters related to interocular distance include the ratio of the intraocular distance to the lateral distance between the two eyes.

21. The ultrasound imaging method according to claim 17, characterized in that, The determined interocular distance parameters, presented in association with the extracted profiles passing through the center points of each eyeball, specifically include: On a cross-section passing through the center point of each eyeball, a first line connecting the paired inner boundary points of both eyes and a second line connecting the paired outer boundary points of both eyes are presented. The first line represents the intraocular distance, and the second line represents the lateral distance.

22. The ultrasound imaging method according to claim 17, characterized in that, The three-dimensional eye region of the fetus is determined based on the acquired three-dimensional volumetric data of the fetal face through any one or a combination of the following: Image features are extracted based on the three-dimensional volume data of the fetal face, and the three-dimensional eye region of the fetus is determined using a trained regression model based on the extracted image features. and / or Based on the three-dimensional volumetric data of the fetal face, the three-dimensional eye region of the fetus is determined using a trained segmentation model. and / or The three-dimensional eye region of the fetus is determined by matching representative data of the three-dimensional volumetric data of the fetal face with a reference template of the representative data.

23. The ultrasound imaging method according to claim 22, characterized in that, The representative data of the three-dimensional volumetric data of the fetal face includes at least one of the three-dimensional volumetric data of the fetal face itself, the data of the two-dimensional cross-section of the fetal face, and the data of the key parts of the fetal face.

24. The ultrasound imaging method according to claim 23, characterized in that, The two-dimensional cross-section of the fetal face is a two-dimensional cross-section passing through the center of each eyeball; and / or the key parts of the fetal face include the key anatomical structures and / or key points of the fetal face.

25. An ultrasound imaging apparatus for a fetal face, comprising a processor configured to perform an ultrasound imaging method for a fetal face according to any one of claims 1-24.

26. A computer-readable storage medium having stored thereon computer-executable instructions which, when executed by a processor, implement the ultrasound imaging method for a fetal face according to any one of 1-24.

Images