WO2020103098A1 - 超声成像方法、设备、存储介质，处理器及计算机设备 - Google Patents

超声成像方法、设备、存储介质，处理器及计算机设备

Info

Publication number
WO2020103098A1
WO2020103098A1 PCT/CN2018/117007 CN2018117007W WO2020103098A1 WO 2020103098 A1 WO2020103098 A1 WO 2020103098A1 CN 2018117007 W CN2018117007 W CN 2018117007W WO 2020103098 A1 WO2020103098 A1 WO 2020103098A1
Authority
WO
WIPO (PCT)
Prior art keywords
key anatomical
anatomical structure
uterus
volume data
dimensional volume
Prior art date
Application number
PCT/CN2018/117007
Other languages
English (en)
French (fr)
Inventor
朱磊
董国豪
邹耀贤
林穆清
胡锦明
Original Assignee
深圳迈瑞生物医疗电子股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳迈瑞生物医疗电子股份有限公司 filed Critical 深圳迈瑞生物医疗电子股份有限公司
Priority to CN201880097331.8A priority Critical patent/CN112654299A/zh
Priority to PCT/CN2018/117007 priority patent/WO2020103098A1/zh
Publication of WO2020103098A1 publication Critical patent/WO2020103098A1/zh

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves

Definitions

  • the present invention relates to the field of ultrasound detection, and in particular, to an ultrasound imaging method, device, storage medium, processor, and computer equipment.
  • the number of fetuses is generally determined in the early pregnancy, the fetus is not mature, the fetus still exists in the form of a gestational sac, its shape is not obvious, it is difficult to identify, and the gestational sac exists in a variety of locations, which can exist in the uterus Multiple locations within.
  • the shape of the uterus which makes it difficult to measure the number of fetuses.
  • Embodiments of the present invention provide an ultrasound imaging method, device, storage medium, processor, and computer device to at least solve the technical problem of low accuracy of fetal quantity detection in related technologies.
  • an ultrasound imaging method including: transmitting ultrasound waves to the uterus by covering the entire uterine area, and receiving ultrasound echoes to obtain ultrasound echo signals; and according to the ultrasound echo signals Obtain three-dimensional volume data of the uterus; determine the pregnancy period of the uterus; identify the key anatomical structure of the fetus corresponding to the pregnancy period from the three-dimensional volume data; determine the fetus in the uterus according to the key anatomical structure quantity.
  • the pregnancy period includes: an early pregnancy period where the gestational week is less than a predetermined number of weeks, or an early pregnancy period where the gestational week is greater than the predetermined number of weeks.
  • the key anatomical structures corresponding to the early pregnancy period include at least one of the following: gestational sac, yolk sac, and embryo; the key anatomical structures corresponding to the early pregnancy period include at least one of the following: cranial brain, trunk , Femur, spine and limbs.
  • identifying the key anatomical structure of the fetus corresponding to the pregnancy period from the three-dimensional volume data includes: acquiring features that can distinguish whether it is a key anatomical structure; and identifying the three-dimensional volume data according to the characteristics At least one area; from the at least one area, a target area is determined, wherein the target area is determined to have the highest probability of the key anatomical structure; and the target area is determined to be the key anatomical structure.
  • the features include at least one of the following: two-dimensional features and three-dimensional features.
  • obtaining features that can distinguish whether it is a key anatomical structure includes: collecting a positive sample determined to be the key anatomical structure and determining a negative sample not being the key anatomical structure; based on machine learning, correcting the positive The samples and the negative samples are trained to obtain features that can distinguish whether they are key anatomical structures.
  • identifying the key anatomical structure of the fetus corresponding to the pregnancy period from the three-dimensional volume data includes: classifying pixel points in the image of the three-dimensional volume data to obtain a classification result; according to the classification The results identified key anatomical structures.
  • identifying the key anatomical structure of the fetus corresponding to the pregnancy period from the three-dimensional volume data includes: determining a structure template, wherein the structure template includes multiple real key anatomical structures; according to the The structure template identifies a target area from the three-dimensional volume data, wherein the target area is the area with the highest matching degree with the key anatomical structure in the structure template; it is determined that the target area is the key anatomical structure.
  • identifying the critical anatomical structure of the fetus corresponding to the pregnancy period from the three-dimensional volume data includes: identifying the to-be-determined key anatomical structure from the three-dimensional volume data; The key anatomical structure to be determined is adjusted to obtain the key anatomical structure.
  • determining the number of fetuses in the uterus according to the key anatomical structure includes: in the case where there are multiple key anatomical structures, acquiring the intrauterine fetuses determined according to the multiple key anatomical structures, respectively The number; determine the number of the most consistent is the number of fetuses in the womb.
  • an ultrasound imaging method comprising: displaying three-dimensional volume data of the uterus, wherein the three-dimensional volume data is to scan the uterus by ultrasound to cover the entire uterine area Data obtained afterwards; showing the pregnancy period of the uterus, and showing the key anatomical structure of the fetus corresponding to the pregnancy period; showing the number of fetuses in the uterus determined according to the key anatomical structure.
  • displaying the key anatomical structure of the fetus corresponding to the pregnancy period includes: displaying a feature that can distinguish whether it is a key anatomical structure; displaying at least one region identified from the three-dimensional volume data according to the feature; highlighting A target area is displayed, wherein the target area is determined to have the highest probability of being the key anatomical structure.
  • displaying the key anatomical structure of the fetus corresponding to the pregnancy period includes: displaying the pixel contour obtained after classifying the pixel points in the image of the three-dimensional volume data, wherein the pixel contour is used to distinguish Key anatomical structures and non-key anatomical structures.
  • displaying the key anatomical structure of the fetus corresponding to the pregnancy period includes: displaying the key anatomical structure in the structure template, wherein the structure template includes a variety of real key anatomical structures; displaying according to the structure template A target area identified from the three-dimensional volume data, wherein the target area is the area with the highest degree of matching with the key anatomical structure in the structural template.
  • displaying the key anatomical structure of the fetus corresponding to the pregnancy period includes: displaying the pending key anatomical structure identified from the three-dimensional volume data; displaying the input operation; displaying the key to the pending key according to the operation After the anatomical structure is adjusted, the key anatomical structure is obtained.
  • another ultrasound imaging method including: acquiring three-dimensional volume data of the uterus, wherein the three-dimensional volume data is data obtained by scanning the uterus with ultrasound; A key anatomical structure is identified from the three-dimensional volume data; according to the key anatomical structure, the number of fetuses in the uterus is determined.
  • an ultrasound imaging apparatus including: a probe; a transmitting circuit that excites the probe to transmit ultrasonic waves to the uterus; a receiving circuit that receives the probe through the probe Receiving an ultrasound echo returned from the uterus to obtain an ultrasound echo signal; a processor that processes the ultrasound echo signal to obtain three-dimensional volume data of the uterus; a display that displays the three-dimensional volume Volume data; wherein, the processor further performs the following steps: identifying key anatomical structures from the three-dimensional volume data; and determining the number of fetuses in the uterus according to the key anatomical structures.
  • the display is further used to display at least one of the following: the key anatomical structure, the number of fetuses in the uterus, and the pregnancy period of the uterus.
  • the storage medium includes a stored program, wherein, when the program is running, the device where the storage medium is located is controlled to execute any one of the above Ultrasound imaging method.
  • a processor for running a program wherein the ultrasound imaging method described in any one of the above is executed when the program is executed.
  • a computer device including: a memory and a processor, the memory stores a computer program; the processor is configured to execute the computer program stored in the memory, When the computer program runs, it executes any of the ultrasound imaging methods described above.
  • an ultrasound wave is transmitted to the uterus by covering the entire uterine area, and an ultrasound echo signal is received to obtain an ultrasound echo signal; three-dimensional volume data of the uterus is obtained according to the ultrasound echo signal; and the uterus is determined
  • the detection of three-dimensional volume data achieves the purpose of effectively detecting and calculating the number of fetuses in the uterus by identifying the key anatomical structure that can accurately represent the number of fetuses, thereby achieving the technical effect of improving the accuracy of fetal number detection, and thus solving the related The technical problem of low accuracy of fetal quantity detection in technology.
  • FIG. 1 is a schematic structural block diagram of an ultrasound imaging device 10 in an embodiment of the present application
  • FIG. 2 is a flowchart of an ultrasound imaging method according to an embodiment of the present invention.
  • FIG. 3 is a flowchart of another ultrasound imaging method according to an embodiment of the present invention.
  • FIG. 4 is a flowchart of another ultrasound imaging method according to an embodiment of the present invention.
  • FIG. 5 is a flowchart of a method for measuring the number of fetuses according to an embodiment of the present invention
  • FIG. 6 is a schematic structural diagram of an ultrasound imaging apparatus according to an embodiment of the present invention.
  • FIG. 1 is a schematic structural block diagram of an ultrasound imaging device 10 in an embodiment of the present application.
  • the ultrasound imaging apparatus 10 may include a probe 100, a transmission circuit 101, a transmission / reception selection switch 102, a reception circuit 103, a beam synthesis circuit 104, a processor 105, and a display 106.
  • the transmitting circuit 101 may excite the probe 100 to transmit ultrasonic waves to the target object.
  • the receiving circuit 103 may receive the ultrasonic echo returned from the target object through the probe 100, thereby obtaining an ultrasonic echo signal.
  • the processor 105 processes the ultrasound echo signal to obtain an ultrasound image of the target object.
  • the ultrasound image obtained by the processor 105 may be stored in the memory 107. These ultrasound images can be displayed on the display 106.
  • a method embodiment of an ultrasound imaging method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings may be executed in a computer system such as a set of computer-executable instructions, and, Although a logical sequence is shown in the flowchart, in some cases, the steps shown or described may be performed in an order different from here.
  • FIG. 2 is a flowchart of an ultrasound imaging method according to an embodiment of the present invention. As shown in FIG. 2, the method includes the following steps:
  • Step S202 transmitting ultrasound waves to the uterus by covering the entire uterine area, and receiving ultrasound echoes to obtain ultrasound echo signals;
  • Step S204 obtaining three-dimensional volume data of the uterus according to the ultrasound echo signal
  • Step S206 determining the pregnancy period of the uterus
  • Step S208 identifying the key anatomical structure of the fetus corresponding to the pregnancy period from the three-dimensional volume data
  • Step S210 Determine the number of fetuses in the uterus according to the key anatomical structure.
  • the method of determining the number of fetuses in the uterus according to the key anatomical structures identified from the three-dimensional volume data the key anatomical structures can more accurately reflect the number of fetuses in the uterus, therefore, the key anatomy
  • the identification of the structure can realize the identification of the fetus in the uterus, and determine the number of fetuses in the uterus according to the identified key anatomical structure, so as to effectively and accurately detect the number of fetuses in the uterus, thereby improving the accuracy of the detection of the number of fetuses
  • the technical effect further solves the technical problem of low accuracy of fetal quantity detection in related technologies.
  • the ultrasound imaging area of the above ultrasound covers the entire uterine area.
  • the ultrasound imaging area detected by ultrasound covers the entire uterine area. Since the fetus exists in the form of a gestational sac during early pregnancy, the gestational sac can be located in various positions in the uterus. For example, uterine fundus, anterior uterine wall, posterior uterine wall, upper uterus or middle uterus, etc.
  • the ultrasound imaging area is extended to the entire uterine area. It should be noted that the ultrasound imaging area can be less than or equal to the ultrasound detection area.
  • the ultrasound detection area should also cover at least the entire uterine area, so that the determined number of fetuses is more accurate, and leakage due to the hidden position of the fetus is avoided. Number of problems.
  • the entire uterine area is used as the ultrasound detection area, which may be an ultrasound wave transmitted to the uterus by covering the entire uterine area, and receiving ultrasound echoes to obtain ultrasound echo signals.
  • the ultrasonic echo signal may be obtained by sending and receiving the ultrasonic wave once. It is also possible to obtain multiple ultrasonic echo signals by sending and receiving multiple ultrasonic waves. The number of times the ultrasound is sent can be determined according to actual needs. If the ultrasound cannot determine the three-dimensional volume data of the entire uterus area or the obtained three-dimensional volume data is difficult to know the number of fetuses, you can send multiple ultrasound waves, and Receive multiple ultrasound echo signals to determine the three-dimensional volume data of the entire uterus area.
  • the three-dimensional volume data of the uterus obtained based on the ultrasonic echo signal may include the three-dimensional coordinates of the measurement point in the uterus in the spatial three-dimensional coordinate system, and may also include the position function of the uterus in the three-dimensional coordinate system.
  • the three-dimensional volume data may further include the three-dimensional size of the uterus, and the three-dimensional size may be length, width, and height.
  • the above three-dimensional volume data may be a three-dimensional array obtained by scanning with ultrasound, that is, the internal environment of the uterus is reflected by means of the array.
  • the three-dimensional volume data can determine the three-dimensional size of the uterus.
  • the above three-dimensional volume data can be determined in various ways.
  • the three-dimensional volume data is acquired through ultrasonic detection.
  • the three-dimensional volume data obtained above may be obtained by real-time scanning, or may be scanned and stored in advance, and read from the memory when the number of fetuses in the uterus needs to be determined.
  • the three-dimensional volume data of the uterus may include the internal structure of the uterus, for example, the position where the gestational sac landed in the uterus and the morphological data of the gestational sac.
  • the above determination of the gestation period of the uterus is performed before the key anatomical structure of the fetus is identified based on the three-dimensional volume data.
  • the above-mentioned determination of the pregnancy period of the uterus can be determined in various ways.
  • the pregnancy period of the above-mentioned uterus can be obtained through human-computer interaction with the subject to which the above-mentioned uterus belongs; ) The testee, by calling the historical test data in the case (or test record) to determine the pregnancy period of the uterus, or other commonly used methods to determine the pregnancy period of the testee to which the uterus belongs. Since the anatomical structure of the uterus differs greatly during different gestation periods, in this embodiment, in order to accurately determine the corresponding number of fetuses corresponding to the gestation period, the step of identifying the key anatomical structure from the three-dimensional volume data may be The pregnancy period of the uterus.
  • the step of determining the pregnancy period of the uterus may include transmitting ultrasonic waves to the uterus by covering the entire uterine area and receiving ultrasonic echoes to obtain ultrasonic echo signals. prior to.
  • the pregnancy period can be determined by the above-mentioned way of determining the pregnancy period of the uterus.
  • the pregnancy period may be determined in the above-described manner of determining the pregnancy period of the uterus.
  • the pregnancy period of the uterus can also be determined by the three-dimensional volume data of the uterus.
  • the pregnancy period can be determined according to the morphological characteristics of different pregnancy periods in the ultrasound detection image corresponding to the three-dimensional volume data.
  • the morphological characteristics include the uterine morphological characteristics and the fetal morphological characteristics.
  • the detection of the number of fetuses is generally in the pre-pregnancy period when the fetus is immature. Since the morphology of the fetuses in the early and second trimesters is not perfect, it is necessary to determine the number of fetuses by combining ultrasound detection images with key anatomical structures. After the second and third trimesters, the fetus develops gradually, and the fetal morphology is relatively complete, with obvious morphological characteristics such as the skull and limbs, the aura of the skull is clear, and key anatomical structures such as the spine and limbs have appeared, which can be used as the basis for identifying the fetus.
  • the pregnancy period is generally early pregnancy period, or early pregnancy period.
  • the above-mentioned pregnancy period includes: early pregnancy period whose gestational week is less than the predetermined number of weeks, or early and middle pregnancy period whose gestational week is greater than the predetermined number of weeks.
  • the first trimester of pregnancy is less than 8 weeks
  • the second trimester of pregnancy is greater than 8 weeks. It should be noted that the 8 weeks listed here are only a reference week number. Due to different uterine individuals, the reference week number used may also be different, depending on the specific circumstances.
  • the key anatomical structures may be the overall structure of the uterus and the structure of fetal development in the uterus. It can be various key anatomical structures such as the anatomical structure in the first trimester, the anatomical structure in the first and second trimester, or the anatomical structure in the second trimester, or other anatomical structures during pregnancy.
  • the anatomical structure in the early pregnancy period may include at least one of the following: amniotic membrane, body pedicle, plexiform chorion, blastoderm, yolk sac, chorion, etc.
  • the anatomical structure during early pregnancy can also include at least one of the following: primitive streak, yolk sac, chorionic cavity, etc.
  • Early and middle pregnancy can include at least one of the following: chorionic cavity, amniotic cavity, intestine, umbilical cord, yolk sac.
  • Early and middle pregnancy can also include at least one of the following: placenta, yolk sac traces, amniotic membrane, chorionic sac.
  • the number of fetuses is determined according to the key anatomical structures identified above. Since the fetus has different morphologies in different gestation periods, the key anatomical structures in different gestation periods are also different. For example, in the early pregnancy period, the shape of the fetus can be expressed as the blastoderm. Therefore, the key anatomical structure in the early pregnancy period can determine the number of fetuses according to the number of blastoderms. In other pregnancy periods, the number of fetuses can be determined according to the key anatomical structures corresponding to other pregnancy periods.
  • the key anatomical structures corresponding to the early pregnancy period include at least one of the following: gestational sac, yolk sac, and embryo; the key anatomical structures corresponding to the early and middle pregnancy period include at least one of the following: cranial brain, trunk, femur, Spine and limbs.
  • the fetus in the early pregnancy, the fetus exists in the form of a gestational sac, and the number of fetuses can be determined according to the gestational sac and the yolk sac, or embryo. In the early and middle pregnancy, the fetus gradually develops, and key anatomical structures such as the brain and limbs appear. The number of fetuses can be determined according to the brain, trunk, femur, spine, and limbs.
  • the key anatomical structure may be identified according to the characteristics of the key anatomical structure. For example, identifying the key anatomical structure from the three-dimensional volume data may include: acquiring features that can distinguish whether it is a key anatomical structure; At least one area is identified; from at least one area, a target area is determined, wherein the target area is determined to be the key anatomical structure with the highest probability; the target area is determined to be the key anatomical structure.
  • the key anatomical structures identified from the three-dimensional volume data can be fully automatic or semi-automatic.
  • the fully automatic method can be monitored according to machine learning or deep learning methods.
  • the semi-automatic method can be determined based on machine learning or deep learning.
  • the features corresponding to the pregnancy period are combined with manual identification.
  • the above-mentioned semi-automatic method may be through machine learning, the area with the highest probability of determining the joint anatomy from the at least one area is the target area, and then manually identifying the target area is the key anatomy, so as to determine the key anatomy Number of fetuses.
  • the above-mentioned area may be an area suspected of having a key anatomical structure. For example, in the early pregnancy period, the area where the gestational sac appears is relatively high. Even if the structure of the gestational sac is not recognized, there is still a certain probability that the sac is present in the area.
  • the aforementioned area may also be an area having a structure similar to the key anatomical structure.
  • the target area is determined from the above at least one area, where the target area is determined to have the highest probability of being a key anatomical structure.
  • the target area can be determined according to the method of machine learning or deep learning. By identifying whether multiple areas are key anatomical structures, a machine learning model or deep learning model is used for training. According to the trained machine learning model or deep learning model, determine the probability that the region is determined to be a key anatomical structure. The probability of determining whether a region is a key anatomical structure can also be determined based on experience.
  • the target area is the area most likely to be a key anatomical structure among the at least one area. Since there are many key anatomical structures, not every key anatomical structure can be used to identify the number of fetuses. Therefore, in this embodiment, the key anatomical structure can be identified from the three-dimensional volume data according to the above method. The identification area of key anatomical structures can be effectively reduced, thereby improving the identification efficiency.
  • the features include at least one of the following: two-dimensional features and three-dimensional features.
  • the above features can be two-dimensional features, which are easy to obtain, quick and easy to handle. It can also be a three-dimensional feature with high accuracy. It can also be a combination of two-dimensional features and three-dimensional features, which is not only easy to obtain and easy to process, but also can guarantee a certain accuracy.
  • a variety of methods may also be used. For example, to efficiently, quickly and accurately obtain the characteristics of a key anatomical structure, the following methods may be used: Positive samples, and negative samples that are not identified as critical anatomical structures; based on machine learning, train positive and negative samples to obtain features that can distinguish whether they are critical anatomical structures.
  • the above-mentioned key anatomical structure may be a fully automatic method for monitoring according to the method of machine learning or deep learning. It may be that a positive sample determined as a critical anatomical structure is collected first, and a negative sample determined as not a critical anatomical structure. Based on machine learning, the positive and negative samples are trained to obtain features that can distinguish whether they are key anatomical structures. For example, in the early pregnancy period, the embryo can be used as a feature to distinguish whether it is a key anatomical structure, and the location of the embryo can be determined as a key anatomical feature. When performing machine learning based on true samples and negative samples, the learning model may be trained based on true samples and negative samples. The positive samples may be germ images, and the negative samples may be non-germ images.
  • the key anatomical structure can be identified according to the way of classifying the pixels in the 3D volume data image. For example, identifying the key anatomical structure from the 3D volume data includes: classifying the pixels in the 3D volume data image To get classification results; identify key anatomical structures based on the classification results.
  • the above-mentioned key anatomical structures are identified and segmented by a recognition algorithm.
  • the above recognition algorithm can be recognized and segmented in various ways.
  • the pixels in the influence of the three-dimensional volume data can be classified to obtain the classification result, and then the key anatomical structure can be identified according to the classification result.
  • the pixels of the fetal trunk, head, and limbs are generally the same type of pixel, and the above-mentioned key anatomical structure can be determined according to the classification of the pixels of this type.
  • the key anatomical structure can be identified according to the structure template.
  • identifying the key anatomical structure from the three-dimensional volume data includes: determining the structure template, wherein the structure template includes multiple real key Anatomical structure; the target area is identified from the three-dimensional volume data according to the structural template, wherein the target area is the area with the highest matching degree with the key anatomical structure in the structural template.
  • the structure of embryos in early and middle pregnancy is relatively fixed, you can collect some early embryo data in advance to create a template, traverse all possible regions in the volume data during detection, and match the template with similarity, select the region with the highest similarity as target area.
  • identifying the key anatomical structure from the three-dimensional volume data may further include: identifying the key anatomical structure to be determined from the three-dimensional volume data; To adjust key anatomical structures to be determined to obtain key anatomical structures.
  • the adjustment and correction processing method can avoid the situation that some three-dimensional volume data can not reflect the key anatomical structure more realistically.
  • the above-mentioned key anatomical structures that will die elsewhere in the three-dimensional volume data may also be in a semi-automatic manner, and the location of a specific anatomical structure in the volume data may be identified through the above method.
  • the user can also manually add, delete, modify, etc. the detection structure through a certain workflow through tools such as the keyboard and mouse, to realize semi-automatic anatomical structure detection. For example, use mouse.
  • determining the number of fetuses in the uterus according to the key anatomical structure includes: when there are multiple key anatomical structures, obtaining the number of fetuses in the uterus determined according to the multiple key anatomical structures respectively; determining the highest consistency The number is the number of fetuses in the womb.
  • the number of fetuses when determining the number of fetuses according to the key anatomical structure, since the number of the above key anatomical structures can be multiple, for example, in the early pregnancy period, the gestational sac, yolk sac, and embryo can be used as the key anatomical structure.
  • the number with the highest consistency is selected as the number of fetuses in the womb.
  • the consistency of the number of fetuses may be that, among the multiple key anatomical structures, the number of identified key anatomical structures with the same number of fetuses accounts for a proportion of the total number of all key anatomical structures.
  • gestational sac for example, in the above-mentioned early pregnancy period, there may be three key anatomical structures: gestational sac, yolk sac, and embryo.
  • gestational sac for example, one fetus is determined according to the gestational sac, one fetus is determined according to the yolk sac, and two are determined according to the embryo
  • the consistency of one fetus is two-thirds, and the consistency of two fetuses is one-third.
  • the number that determines the highest consistency is the number of fetuses in the womb, which means that one fetus is the number of fetuses in the womb. , That is to determine the number of fetuses in the uterus as one.
  • FIG. 3 is a flowchart of another ultrasound imaging method according to an embodiment of the present invention. As shown in FIG. 3, according to another aspect of the embodiment of the present invention, another ultrasound imaging method is provided. The method includes the following steps :
  • Step S302 displaying three-dimensional volume data of the uterus, wherein the three-dimensional volume data is data obtained by scanning the uterus in a manner that covers the entire uterine area by ultrasound;
  • Step S304 displaying the pregnancy period of the uterus, and displaying the key anatomical structure of the fetus corresponding to the pregnancy period;
  • step S306 the number of fetuses in the uterus determined according to the key anatomical structure is displayed.
  • the execution subject of the above steps may be a display device.
  • the above display steps according to the key anatomical structure identified from the three-dimensional volume data, determine the number of fetuses in the uterus.
  • the key anatomical structure can more accurately reflect the number of fetuses in the uterus.
  • the identification of the structure can realize the identification of the fetus in the uterus, and determine the number of fetuses in the uterus according to the identified key anatomical structure, so as to effectively and accurately detect the number of fetuses in the uterus, thereby improving the accuracy of the detection of the number of fetuses
  • the technical effect further solves the technical problem of low accuracy of fetal quantity detection in related technologies.
  • the processor of the display device can perform data processing and acquisition, and the display device can display it. It is also possible to receive and process data according to the processing device, and the processing device sends the displayed data to the display device for display by the display device.
  • displaying the key anatomical structure of the fetus corresponding to the pregnancy period includes: displaying a feature that can distinguish whether it is a key anatomical structure; displaying at least one region identified from the three-dimensional volume data according to the feature; Highlight the target area, where the target area is determined to have the highest probability of being the key anatomical structure.
  • displaying the key anatomical structure of the fetus corresponding to the pregnancy period includes: displaying the pixel contour obtained after classifying the pixel points in the image of the three-dimensional volume data, wherein the pixel contour is used to Distinguish between critical and non-critical anatomical structures.
  • displaying the key anatomical structure of the fetus corresponding to the pregnancy period includes: displaying the key anatomical structure in the structure template, wherein the structure template includes multiple real key anatomical structures; displaying according to the structure The template identifies the target area from the three-dimensional volume data, wherein the target area is the area with the highest degree of matching with the key anatomical structure in the structural template.
  • displaying the key anatomical structure of the fetus corresponding to the pregnancy period includes: displaying the to-be-determined key anatomical structure identified from the three-dimensional volume data; displaying the input operation; displaying the to-be-determined according to the operation After adjusting the key anatomical structure, the key anatomical structure is obtained.
  • the pregnancy period includes: an early pregnancy period in which the gestational week is less than a predetermined number of weeks, or an early pregnancy period in which the gestational week is greater than the predetermined number of weeks.
  • the display device displays the gestation period of the above-mentioned display uterus: including the early pregnancy period whose gestational week is less than the predetermined number of weeks, or the early and middle pregnancy period whose gestational period is greater than the predetermined number of weeks.
  • the above display device can prompt the doctor or the tester to facilitate Doctors or testers make reference when they need to make reasonable guesses or judgments based on the test situation.
  • FIG. 4 is a flowchart of another ultrasound imaging method according to an embodiment of the present invention. As shown in FIG. 4, according to another aspect of the embodiment of the present invention, another ultrasound imaging method is provided. The method includes the following steps :
  • Step S402 acquiring three-dimensional volume data of the uterus, wherein the three-dimensional volume data is data obtained by scanning the uterus through ultrasound;
  • Step S404 identifying key anatomical structures from the three-dimensional volume data
  • Step S406 Determine the number of fetuses in the uterus according to the key anatomical structure.
  • the key anatomical structure identified from the three-dimensional volume data determine the number of fetuses in the uterus.
  • the key anatomical structure can more accurately reflect the number of fetuses in the uterus. Therefore, the key anatomical structure
  • the identification can realize the identification of the fetus in the womb, determine the number of fetuses in the womb according to the identified key anatomical structure, and achieve the purpose of effectively and accurately detecting the number of fetuses in the womb, thereby realizing the technology of improving the accuracy of fetal number detection The effect further solves the technical problem of low accuracy of fetal quantity detection in related technologies.
  • an embodiment of the present invention also provides a method for detecting the number of fetuses in the uterus. This detection method can be used as a preferred implementation of this embodiment, which will be described in detail below.
  • Ultrasound technology has become the most widely used, most frequently used, and fastest new technology in medical imaging examinations due to its advantages of safety, reliability, fast and convenient, and repeatable examinations.
  • the development of ultrasound technology and artificial intelligence technology has further promoted the progress of clinical diagnosis and treatment technology.
  • China has a large population base.
  • the intelligentization of ultrasound equipment can enable hospitals to obtain more shared resources and technical support, systematically reduce costs; help doctors improve examination efficiency and reduce misdiagnosis rates; and provide patients with more accurate diagnosis suggestions and personalized treatment plans. Therefore, the research and development of intelligent ultrasound products are of great importance and necessity for all levels of society.
  • Obstetric ultrasound is one of the most widely used areas for ultrasound diagnosis. In the obstetric ultrasound examination of early and middle pregnancy, determining the number of viable fetuses is the basis of all other examinations. Misdiagnosis will bring a series of serious problems.
  • This embodiment provides a method and device for automatically counting the number of fetuses. After the doctor completes the 3D ultrasound data collection, the method and device can automatically identify the anatomical structure of different pregnancy periods, count the number of fetuses, and solve the problem of multiple births in ultrasound examinations. It is prone to the problem of quantitative errors, and it can save the time of prenatal examination and reduce the technical dependence on ultrasound clinicians.
  • two-dimensional ultrasound can only obtain single-sided information of the inspection object, which can easily lead to misdiagnosis and missed diagnosis of multiple births, and has certain limitations.
  • Three-dimensional ultrasound makes up for the shortcomings of two-dimensional ultrasound spatial imaging. It can display the three-dimensional shape, internal structure of the anatomical part and its spatial position relationship with the surrounding tissue through multiple imaging modes.
  • the accurate statistics of twins or multiple births are time-consuming and laborious, especially when the internal environment of the uterus is relatively complex and the position of the fetus is difficult to observe, the experience of ultrasound clinicians is relatively high.
  • This embodiment is based on pregnant woman's uterine body data, through pattern recognition or machine learning algorithm, to process the three-dimensional volume data of the entire uterus range, automatically identify the key anatomical parts of the fetus, can accurately and quickly count the number of fetuses.
  • FIG. 5 is a flowchart of a method for measuring the number of fetuses according to an embodiment of the present invention.
  • the implementation process of the technical solution of this embodiment is divided into three steps, namely: obtaining pregnant women Three-dimensional volume data of the uterus; automatically identify the key anatomical structures of the fetus in the three-dimensional volume data; and determine the number of fetuses based on the number of identified key structures.
  • the specific details of the three steps are as follows:
  • Step 1 Obtain the three-dimensional volume data of the pregnant woman's uterus
  • the ROI (Region of Interest) and the fan scan angle can be set to be large enough so that the scanning range covers the entire Uterus area. Due to the small uterine area in the first and second trimester, three-dimensional ultrasound can usually scan the entire uterine area.
  • Step 2 Identify the key anatomical structure of the fetus in the three-dimensional volume data
  • the system After acquiring the three-dimensional volume data of the uterus, the system needs to identify key anatomical structures, which can be distinguished and identified according to the two situations of early pregnancy and early and middle pregnancy.
  • the identification method can be semi-automatic or fully automatic.
  • the key anatomical structures of each pregnancy can be automatically identified.
  • Key anatomical structures such as gestational sac, yolk sac, and embryo can be detected in the early pregnancy stage of less than 8 weeks (probably the gestational week, need not be very accurate, the same below).
  • the first sign of pregnancy found by ultrasound is the gestational sac.
  • Transabdominal ultrasound can usually find the gestational sac 5 to 6 weeks after menopause, while transvaginal ultrasound can see the gestational sac 4 weeks after the last menstruation. After 5 to 6 weeks of pregnancy, through vaginal ultrasound examination, 100% of normal pregnancy can show yolk sac, and at the same time can detect embryo and heart beat.
  • the identification of key anatomical parts can be either fully automatic or semi-automatic. It can detect key anatomical structures as well as the entire fetus. There are many ways to automatically detect key anatomical structures. Machine learning or deep learning methods can be used to detect key anatomical structures in the three-dimensional volume data.
  • a certain number of early embryo images (called positive samples) and a certain number of non-germ images (called negative samples) can be collected in advance, and then an artificial neural network can be designed based on machine learning or deep learning algorithms .
  • an artificial neural network can be designed based on machine learning or deep learning algorithms .
  • Use the multi-layer network structure to automatically learn the features that can distinguish between positive and negative samples, use these features to traverse all possible regions in the three-dimensional volume data during detection, calculate the probability that the region is judged to be a positive sample, and select the most probable The area is the target area.
  • Traditional machine learning algorithms need to perform feature extraction based on certain feature extraction methods (such as grayscale, texture, and spatial information) in advance. Such methods commonly include Adaboost algorithm, support vector machine (SVM), random forest (Random Forest), etc.
  • Deep learning algorithms can directly perform feature extraction and network training based on multi-frame 2D video or 3D volume data for effective anatomical structure detection.
  • Common methods of this type include Convolutional Neural Network Algorithm (CNN) and Recurrent Neural Network Algorithm ( RNN), FastRCNN, YOLO, SSD, etc.
  • the key anatomical structures in the three-dimensional volume data can be accurately segmented through image segmentation methods.
  • Segmentation is to classify which category each pixel in the image belongs to, and can directly obtain the outline and position of the key anatomical structure in the image.
  • Commonly used segmentation methods include LevelSet, Graph Cut, Snake, Random walker, watershed algorithm, threshold segmentation and other methods; in addition to traditional methods, deep learning methods can also achieve key anatomy Structure segmentation, such as FCN, UNet, SegNet, Deeplab, etc.
  • the position of a specific anatomical structure in the three-dimensional volume data can be identified. If the full-automatic method cannot be accurately identified, users can also use the keyboard, mouse and other tools to supplement, delete, and modify the detection structure through a certain workflow to achieve semi-automatic anatomical structure detection, for example, using a mouse.
  • the machine learning and pattern recognition algorithms mentioned above are all algorithms for identifying the critical anatomical structure of the embryo or fetus.
  • the core of this embodiment is to determine the number of fetuses by the number of key anatomical structures, and other methods can also be used to detect anatomical structures. The purpose has not changed the substantive process.
  • Step 3 Determine the number of fetuses according to the number of key anatomical structures
  • the number of fetuses can be counted according to the key structures.
  • the statistics of the number of fetuses in the early pregnancy period can be determined by the number of gestational sacs and embryos identified; in the early and middle pregnancy period, the number of fetuses can be determined according to the number of key anatomical structures identified such as the brain, trunk, femur, and spine.
  • the system can identify the number of fetuses by identifying an anatomical structure. For example, if one fetal brain is detected in the three-dimensional volume data, it means that there is only one fetus in the three-dimensional volume data. However, any detection algorithm has a certain false detection rate. In order to improve the recognition accuracy, multiple key anatomical structures can be identified at the same time in the embodiment of the present invention, and then the voting strategy of the number of multiple key anatomical structures is used in the final count of fetuses. Determine, for example, that the system has detected a total of 5 anatomical structures.
  • the fetal statistics will be displayed on the image interface.
  • the credibility of the structure can also be output. For some volume data with bad images, there may be errors in the structure recognition. At this time, a low credibility value can be used to remind the doctor to pay attention to the review.
  • the key point of this embodiment is: a method for determining the number of fetuses in the uterus by identifying key anatomical structures in the three-dimensional volume data of the uterus or the entire embryonic structure.
  • the volume data acquisition step is used to acquire three-dimensional volume data;
  • the anatomical structure recognition step is used to identify the anatomical structure in the volume data;
  • the step of determining the number of fetuses is used to determine the number of fetuses and displayed on the image interface.
  • FIG. 6 is a schematic structural diagram of an ultrasound imaging device according to an embodiment of the present invention.
  • an ultrasound imaging device is also provided, including: a probe 602, a transmitting circuit 604, and a receiving circuit 606,
  • the processor 608 and the display 610 will be described in detail below.
  • the display 610 is also used to display at least one of the following: key anatomical structures, the number of fetuses in the uterus, and the pregnancy period in which the uterus is located.
  • the storage medium includes a stored program, wherein, when the program is running, the device where the storage medium is located is controlled to perform any one of the ultrasound imaging methods described above.
  • a processor for running a program wherein any one of the above-mentioned ultrasound imaging methods is executed when the program runs.
  • a computer device including: a memory and a processor, the memory stores a computer program; a processor, used to execute the computer program stored in the memory, the computer program executes the above when running Any one of the ultrasound imaging methods.
  • the disclosed technical content may be implemented in other ways.
  • the device embodiments described above are only schematic.
  • the division of the unit may be a logical function division.
  • there may be another division manner for example, multiple units or components may be combined or Integration into another system, or some features can be ignored, or not implemented.
  • the displayed or discussed mutual coupling or direct coupling or communication connection may be indirect coupling or communication connection through some interfaces, units or modules, and may be in electrical or other forms.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit.
  • the above integrated unit may be implemented in the form of hardware or software functional unit.
  • the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it may be stored in a computer-readable storage medium.
  • the technical solution of the present invention essentially or part of the contribution to the existing technology or all or part of the technical solution can be embodied in the form of a software product, the computer software product is stored in a storage medium , Including several instructions to enable a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention.
  • the foregoing storage media include: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program code .

Landscapes

  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Pathology (AREA)
  • Radiology & Medical Imaging (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Ultra Sonic Daignosis Equipment (AREA)

Abstract

一种超声成像方法、设备、存储介质，处理器及计算机设备。其中，该方法包括：通过覆盖整个子宫区域的方式向子宫发射超声波，并接收超声回波，获得超声回波信号（S202）；根据超声回波信号获得子宫的三维体数据（S204）；确定子宫所处的孕期（S206）；从三维体数据中识别出与孕期对应的胎儿的关键解剖结构（S208）；根据关键解剖结构，确定子宫内胎儿的数量（S210）。解决了相关技术中对胎儿数量检测的准确率较低的技术问题。

Description

超声成像方法、设备、存储介质，处理器及计算机设备 技术领域
本发明涉及超声检测领域，具体而言，涉及一种超声成像方法、设备、存储介质，处理器及计算机设备。
背景技术
在超声检查中，确定胎儿数量一般是在怀孕早期，胎儿并未发育成熟，胎儿尚且以孕囊形式存在，其形状特征还不明显，难以识别，而且孕囊存在的位置多样，可以存在于子宫内的多种位置。另外，根据个体差异，子宫的形状存在较大差异，这对测量胎儿数量增加了难度。
现有技术中，在检测胎儿数量时，一般都是由医生或者检验者，根据检测到的超声图像，结合检验者的想象，进行合理推测。存在较大的不准确性，准确率较低。
针对上述的问题，目前尚未提出有效的解决方案。
发明内容
本发明实施例提供了一种超声成像方法、设备、存储介质，处理器及计算机设备，以至少解决相关技术中对胎儿数量检测的准确率较低的技术问题。
根据本发明实施例的一个方面，提供了一种超声成像方法，包括：通过覆盖整个子宫区域的方式向子宫发射超声波，并接收超声回波，获得超声回波信号；根据所述超声回波信号获得子宫的三维体数据；确定所述子宫所处的孕期；从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结构；根据所述关键解剖结构，确定所述子宫内胎儿的数量。
一个实施例中，所述孕期包括：孕周小于预定周数的早孕期，或者孕周大于预定周数的早中孕期。
一个实施例中，与所述早孕期对应的关键解剖结构包括以下至少之一：妊娠囊、卵黄囊及胚芽；与所述早中孕期对应的关键解剖结构包括以下至少之一：颅脑、躯干、股骨、脊柱及四肢。
一个实施例中，从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结 构包括：获取能够区别是否为关键解剖结构的特征；根据所述特征从所述三维体数据中识别出至少一个区域；从所述至少一个区域中，确定出目标区域，其中，所述目标区域被判定为所述关键解剖结构的概率最大；确定所述目标区域为所述关键解剖结构。
一个实施例中，所述特征包括以下至少之一：二维特征和三维特征。
一个实施例中，获取能够区别是否为关键解剖结构的特征包括：收集确定为所述关键解剖结构的正样本，和确定不为所述关键解剖结构的负样本；基于机器学习，对所述正样本和所述负样本进行训练，得到能够区别是否为关键解剖结构的特征。
一个实施例中，从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结构包括：对所述三维体数据的影像中的像素点进行分类，得到分类结果；根据所述分类结果识别出关键解剖结构。
一个实施例中，从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结构包括：确定结构模板，其中，所述结构模板中包括多种真实的关键解剖结构；根据所述结构模板从所述三维体数据中识别出目标区域，其中，所述目标区域为与所述结构模板中的关键解剖结构匹配度最高的区域；确定所述目标区域为所述关键解剖结构。
一个实施例中，从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结构包括：从所述三维体数据中识别出待定关键解剖结构；通过接收输入的操作的方式，对所述待定关键解剖结构进行调整，得到所述关键解剖结构。
一个实施例中，根据所述关键解剖结构，确定所述子宫内胎儿的数量包括：在所述关键解剖结构为多个的情况下，获取分别依据多个关键解剖结构确定的所述子宫内胎儿的数量；确定一致性最高的数量为所述子宫内胎儿的数量。
根据本发明实施例的另一方面，还提供了一种超声成像方法，包括：显示子宫的三维体数据，其中，所述三维体数据是通过超声以覆盖整个子宫区域的方式对子宫进行扫查后得到的数据；显示所述子宫所处的孕期，以及显示与所述孕期对应的胎儿的关键解剖结构；显示根据所述关键解剖结构确定的所述子宫内胎儿的数量。
一个实施例中，显示与所述孕期对应的胎儿的关键解剖结构包括：显示能够区别是否为关键解剖结构的特征；显示根据所述特征从所述三维体数据中识别出的至少一个区域；突出显示目标区域，其中，所述目标区域被判定为所述关键解剖结构的概率最大。
一个实施例中，显示与所述孕期对应的胎儿的关键解剖结构包括：显示对所述三维体数据的影像中的像素点进行分类后，得到的像素轮廓，其中，所述像素轮廓用于 区分关键解剖结构与非关键解剖结构。
一个实施例中，显示与所述孕期对应的胎儿的关键解剖结构包括：显示结构模板中的关键解剖结构，其中，所述结构模板中包括多种真实的关键解剖结构；显示根据所述结构模板从所述三维体数据中识别出的目标区域，其中，所述目标区域为与所述结构模板中的关键解剖结构匹配度最高的区域。
一个实施例中，显示与所述孕期对应的胎儿的关键解剖结构包括：显示从所述三维体数据中识别出的待定关键解剖结构；显示输入的操作；显示根据所述操作对所述待定关键解剖结构进行调整后，得到的所述关键解剖结构。
根据本发明实施例的另一方面，还提供了另一种超声成像方法，包括：获取子宫的三维体数据，其中，所述三维体数据是由经超声对子宫进行扫查后得到的数据；从所述三维体数据中识别出关键解剖结构；根据所述关键解剖结构，确定所述子宫内胎儿的数量。
根据本发明实施例的另一方面，还提供了一种超声成像设备，包括：探头；发射电路，所述发射电路激励所述探头向子宫发射超声波；接收电路，所述接收电路通过所述探头接收从所述子宫返回的超声回波以获得超声回波信号；处理器，所述处理器处理所述超声回波信号以获得所述子宫的三维体数据；显示器，所述显示器显示所述三维体数据；其中，所述处理器还执行如下步骤：从所述三维体数据中识别出关键解剖结构；并根据所述关键解剖结构，确定所述子宫内胎儿的数量。
一个实施例中，所述显示器，还用于显示以下至少之一：所述关键解剖结构，所述子宫内胎儿的数量，子宫所处的孕期。
根据本发明实施例的另一方面，还提供了一种存储介质，所述存储介质包括存储的程序，其中，在所述程序运行时控制所述存储介质所在设备执行上述中任意一项所述的超声成像方法。
根据本发明实施例的另一方面，还提供了一种处理器，所述处理器用于运行程序，其中，所述程序运行时执行上述中任意一项所述的超声成像方法。
根据本发明实施例的另一方面，还提供了一种计算机设备，包括：存储器和处理器，所述存储器存储有计算机程序；所述处理器，用于执行所述存储器中存储的计算机程序，所述计算机程序运行时执行上述中任意一项所述的超声成像方法。
在本发明实施例中，采用通过覆盖整个子宫区域的方式向子宫发射超声波，并接收超声回波，获得超声回波信号；根据所述超声回波信号获得子宫的三维体数据；确 定所述子宫所处的孕期；从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结构；根据所述关键解剖结构，确定所述子宫内胎儿的数量的方式，通过对子宫内胎儿的三维体数据的检测，通过识别能够准确代表胎儿数量的关键解剖结构，达到了有效检测与计算子宫内胎儿数量的目的，从而实现了提高了胎儿数量检测的准确率的技术效果，进而解决了相关技术中对胎儿数量检测的准确率较低的技术问题。
附图说明
此处所说明的附图用来提供对本发明的进一步理解，构成本申请的一部分，本发明的示意性实施例及其说明用于解释本发明，并不构成对本发明的不当限定。在附图中：
图1为本申请实施例中的超声成像设备10的结构框图示意图；
图2是根据本发明实施例的一种超声成像方法的流程图；
图3是根据本发明实施例的另一种超声成像方法的流程图；
图4是根据本发明实施例的另一种超声成像方法的流程图；
图5是根据本发明实施方式的胎儿数量测量方法的流程图；
图6是根据本发明实施例的一种超声成像设备的结构示意图。
具体实施方式
为了使本技术领域的人员更好地理解本发明方案，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分的实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都应当属于本发明保护的范围。
需要说明的是，本发明的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本发明的实施例能够以除了在这里图示或描述的那些以外的顺序实施。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
图1为本申请实施例中的超声成像设备10的结构框图示意图。该超声成像设备10可以包括探头100、发射电路101、发射/接收选择开关102、接收电路103、波束合成电路104、处理器105和显示器106。发射电路101可以激励探头100向目标对象发射超声波。接收电路103可以通过探头100接收从目标对象返回的超声回波，从而获得超声回波信号。该超声回波信号经过波束合成电路104进行波束合成处理后，送入处理器105。处理器105对该超声回波信号进行处理，以获得目标对象的超声图像。处理器105获得的超声图像可以存储于存储器107中。这些超声图像可以在显示器106上显示。
根据本发明实施例，提供了一种超声成像方法的方法实施例，需要说明的是，在附图的流程图示出的步骤可以在诸如一组计算机可执行指令的计算机系统中执行，并且，虽然在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤。
图2是根据本发明实施例的一种超声成像方法的流程图，如图2所示，该方法包括如下步骤：
步骤S202，通过覆盖整个子宫区域的方式向子宫发射超声波，并接收超声回波，获得超声回波信号；
步骤S204，根据超声回波信号获得子宫的三维体数据；
步骤S206，确定子宫所处的孕期；
步骤S208，从三维体数据中识别出与孕期对应的胎儿的关键解剖结构；
步骤S210，根据关键解剖结构，确定子宫内胎儿的数量。
通过上述步骤，采用根据从三维体数据中识别出的关键解剖结构，确定子宫内胎儿的数量的方式，关键解剖结构能在一定程度上较为准确地体现子宫内胎儿的数量，因此，对关键解剖结构的识别能够实现对子宫内胎儿的识别，依据识别出的关键解剖结构确定宫内胎儿的数量，达到了有效准确地检测子宫内胎儿数量的目的，从而实现了提高胎儿数量检测的准确率的技术效果，进而解决了相关技术中对胎儿数量检测的准确率较低的技术问题。
上述超声的超声成像区域覆盖整个子宫区域。在获取子宫的三维体数据时，超声探测的超声成像区域覆盖整个子宫区域。由于在早孕期，胎儿以孕囊的形式存在，孕囊可以位于子宫内的多种位置。例如，子宫宫底、子宫前壁、子宫后壁、子宫上部或者子宫中部等。为了防止发生漏检或者错检，将超声成像区域的范围扩大至整个子宫 区域。需要说明的是，超声成像区域可以小于或者等于超声探测区域，因此，超声探测区域也应该至少覆盖整个子宫区域，从而使得确定出的胎儿数量较为准确，避免由于胎儿所处位置较为隐蔽而出现漏数的问题。在本实施例中，采用整个子宫区域作为超声探测区域，可以是通过覆盖整个子宫区域的方式向子宫发射超声波，并接收超声回波，获得超声回波信号。
在以覆盖整个子宫区域的方式向子宫发射超声波，并接收超声回波，获得超声回波信号时，可以是通过发送一次超声波并接收，获得超声回波信号。还可以是通过发送多次超声波并接收，从而获得超声多次超声回波信号。上述超声波的发送次数可以根据实际需求确定，在发送一次超声波无法确定上述整个子宫区域的三维体数据，或者获得的三维体数据较为困难地获知胎儿数量的情况下，可以通过发送多次超声波，并接收多次超声回波信号，从而确定上述整个子宫区域的三维体数据。
上述根据超声回波信号所获得的子宫的三维体数据，可以包括子宫内的测量点在空间立体坐标系中的三维坐标，还可以包括上述子宫在三维立体坐标系中的位置函数。上述三维体数据还可以包括上述子宫的三维尺寸，上述三维尺寸可以是长度，宽度，高度。上述三维体数据可以是通过超声进行扫描后获得的立体阵列，即通过阵列的方式来体现子宫的内部环境。根据上述三维体数据可以确定子宫的立体尺寸。上述三维体数据可以通过多种方式确定，在本实施例中，通过超声探测获取上述三维体数据。上述获取三维体数据可以是实时扫查获得的，也可以时预先扫描并存储，在需要确定子宫内胎儿的数量时从存储器中读取的。上述子宫的三维体数据可以包括子宫内部结构，例如，孕囊着落在子宫内的位置和上述孕囊的形态数据。
由于胎儿在不同孕期的形态不同，而且在不同的孕期形态差异较大，因此在上述在根据三维体数据识别出胎儿的关键解剖结构之前，执行上述确定上述子宫所处的孕期。上述确定子宫所处的孕期，可以根据多种方式确定，例如，可以通过与上述子宫所属的被检测者进行人机交互的方式，获取上述子宫所处孕期；还可以对于有病例(或者检测记录)的被检测者，通过调用上述病例(或者检测记录)中的历史检测数据确定上述子宫所处孕期，还可以是其他常用的确定上述子宫所属的被检测者的孕期的方式。由于不同孕期的子宫关键解剖结构形态差异较大，因此在本实施例中，为准确确定出对应的孕期对应的胎儿数量，可以从三维体数据中识别出关键解剖结构的步骤之前，先确定该子宫所处的孕期。需要说明的是，在一种实施例中，上述确定上述子宫所处的孕期的步骤，可以在通过覆盖整个子宫区域的方式向子宫发射超声波，并接收超声回波，获得超声回波信号的步骤之前。例如，在向子宫发射超声波之前，可以通过上述所举例的确定子宫孕期的方式确定孕期。还可以在根据超声回波信号获得子宫 的三维体数据的步骤之前，例如，在根据超声回波信号获得子宫的三维体数据之前，可以通过上述确定子宫孕期的方式确定孕期。
在子宫孕期无法通过上述方法确定时，还可以通过上述子宫的三维体数据确定子宫所处孕期。在确定孕期时，可以根据三维体数据对应的超声检测图像中不同孕期的形态特征确定孕期，上述形态特征包括子宫的形态特征和胎儿的形态特征。
在一种施例中，胎儿数量检测一般在胎儿发育未成熟的怀孕前期，由于胎儿在早孕期和中孕期的形态不太完善，因此需要超声检测图像结合关键解剖结构确定胎儿数量。在中晚孕期以后，胎儿逐渐发育，胎儿形态发育较为完全，具有头颅、四肢等明显的形态特征，颅骨光环清晰，脊柱、四肢等关键解剖结构均已出现，均可以作为识别胎儿的依据。另外，胎儿的体积也较大，可以通过超声图像直接进行辨识和统计数量。因此，在本实施例的上述确定子宫孕期的步骤中，其中的孕期一般为早孕期，或者早中孕期。上述孕期包括：孕周小于预定周数的早孕期，或者孕周大于预定周数的早中孕期。例如，孕周小于8周的早孕期，孕周大于8周的中孕期。需要说明的是，此处所列举的8周仅仅为一个参考周数，由于子宫个体的不同，所采用的参考周数也是可以不同的，需要依据具体的情况而定。
从获取的子宫的三维体数据中识别出关键解剖结构，上述关键解剖结构可以是子宫的整体结构，以及子宫内胎儿发育的结构。可以是早孕期解剖结构，早中孕期解剖结构，或者中孕期解剖结构，或者其他孕期的解剖结构等各个关键解剖结构。其中，早孕期的解剖结构，可以包括下列至少之一：羊膜、体蒂、丛密绒毛膜、胚盘、卵黄囊、绒毛膜等。早孕期的解剖结构还可以包括下列至少之一：原条、卵黄囊、绒毛膜腔等。早中孕期可以包括下列至少之一：绒毛膜腔、羊膜腔、肠、脐带、卵黄囊。早中孕期还可以包括下列至少之一：胎盘、卵黄囊遗迹、羊膜、绒毛膜囊。
根据上述识别出的关键解剖结构确定胎儿的数量，由于胎儿在不同的孕期的形态不同，因此，在不同孕期的关键解剖结构也不同。例如，在早孕期，胎儿的形态可以表现为胚盘，因此，在早孕期的关键解剖结构，可以根据胚盘的数量确定胎儿的数量。在其他孕期都可以根据其他孕期相对应的关键解剖结构确定胎儿数量。在一个实施例中，与早孕期对应的关键解剖结构包括以下至少之一：妊娠囊、卵黄囊及胚芽；与早中孕期对应的关键解剖结构包括以下至少之一：颅脑、躯干、股骨、脊柱及四肢。
需要说明的是，在早孕期，胎儿以孕囊的形式存在，可以根据妊娠囊、卵黄囊即胚芽确定胎儿的数量。在早中孕期，胎儿逐渐发育，出现颅脑、四肢等关键解剖结构，可以根据颅脑、躯干、股骨、脊柱及四肢确定胎儿的数量。
在从三维体数据中识别出关键解剖结构时，可以采用多种处理方式，下面举例说明。
例如，可以依据关键解剖结构的特征的方式来识别关键解剖结构，例如，从三维体数据中识别出关键解剖结构可以包括：获取能够区别是否为关键解剖结构的特征；根据特征从三维体数据中识别出至少一个区域；从至少一个区域中，确定出目标区域，其中，目标区域被判定为关键解剖结构的概率最大；确定目标区域为关键解剖结构。
从三维体数据中识别出关键解剖结构可以是全自动方式，也可以是半自动方式，上述全自动方式可以根据机器学习或者深度学习的方法进行监测，上述半自动方式可以是根据机器学习或者深度学习确定该孕期对应的特征结合人工识别的方式。上述半自动方式可以是通过机器学习，从上述至少一个区域中确定关节解剖结构的概率最大的区域为目标区域，然后通过人工识别出目标区域即是关键解剖结构，从而根据识别出的关键解剖结构确定胎儿数量。上述的区域可以是疑似具有关键解剖结构的区域，例如，在早孕期中，孕囊出现概率比较大的区域，即便没有识别出孕囊的结构，但是在该区域内还是存在一定几率存在孕囊。上述区域还可以是具有与关键解剖结构相似的结构的区域。
上述从至少一个区域中，确定出目标区域，其中，目标区域被判定为关键解剖结构的概率最大。确定出目标区域可以根据机器学习或者深度学习的方法进行确定，通过对多个区域是否为关键解剖结构的识别结果，采用机器学习模型或者深度学习模型进行训练。根据训练好的上述机器学习模型或者深度学习模型确定区域被判定为关键解剖结构的概率。还可以根据经验判定确定区域是否为关键解剖结构的概率。上述目标区域为上述至少一个的区域中最有可能为关键解剖结构的区域。由于关键解剖结构中包括的结构较多，并不是每个关键解剖结构都是可以用于识别胎儿数量的，因此在本实施例中，可以根据上述方式从三维体数据中识别出关键解剖结构，可以有效缩小关键解剖结构的识别区域，从而提高识别效率。
在一个实施例中，特征包括以下至少之一：二维特征和三维特征。
上述特征可以是二维特征，获取方便，迅速，便于处理。还可以是三维特征，准确率高。还可以是二维特征与三维特征相结合的方式，不仅获取方便，便于处理，而且能够保证一定的准确率。
在一个实施例中，获取能够区别是否为关键解剖结构的特征时，也可以采用多种方式，例如，为高效快速准确获取关键解剖结构的特征，可以采用以下方式：收集确定为关键解剖结构的正样本，和确定不为关键解剖结构的负样本；基于机器学习，对 正样本和负样本进行训练，得到能够区别是否为关键解剖结构的特征。
在上述从三维体数据中识别出关键解剖结构时，可以是根据机器学习或者深度学习的方法进行监测的全自动方式。可以是先收集确定为关键解剖结构的正样本，以及确定不为关键解剖结构的负样本。基于机器学习，对正样本和负样本进行训练，得到能够区别是否为关键解剖结构的特征。例如，在早孕期，可以将胚芽作为能否区别是否为关键解剖结构的特征，有胚芽的位置，可以确定为关键解剖特征。在根据真样本和负样本进行机器学习时，可以根据真样本和负样本对学习模型进行训练，上述正样本可以是胚芽图像，负样本可以是非胚芽图像。
又例如，可以依据对三维体数据的影像中的像素进行分类的方式来识别关键解剖结构，例如，从三维体数据中识别出关键解剖结构包括：对三维体数据的影像中的像素点进行分类，得到分类结果；根据分类结果识别出关键解剖结构。
从三维体数据中识别关键解剖结构时，由识别算法进行识别和分割上述关键解剖结构。上述识别算法可以采用多种方式进行识别和分割，可以采用对三维体数据影响中的像素点进行分类，得到分类结果，然后根据分类结果识别出关键解剖结构。例如，胎儿躯干、头颅、四肢的像素一般为同一种像素，可以根据对该种像素的分类，确定上述关键解剖结构。
还例如，可以依据结构模板的方式来识别关键解剖结构，例如，在一个实施例中，从三维体数据中识别出关键解剖结构包括：确定结构模板，其中，结构模板中包括多种真实的关键解剖结构；根据结构模板从三维体数据中识别出目标区域，其中，目标区域为与结构模板中的关键解剖结构匹配度最高的区域。
在从三维体数据中识别出关键解剖结构，还可以采用模板匹配的方法在体数据中检测出一些关键解剖结构。例如，早中孕时期的胚胎结构比较固定，可以事先收集一些早中孕胚胎数据建立模板，在检测时遍历体数据中所有可能的区域，和模板进行相似度匹配，选择相似度最高的区域为目标区域。
在一个实施例中，为使得识别出关键解剖结构更为准确，从三维体数据中识别出关键解剖结构还可以包括：从三维体数据中识别出待定关键解剖结构；通过接收输入的操作的方式，对待定关键解剖结构进行调整，得到关键解剖结构。采用调整修正的处理方式，可以在一定程度上避免了有些三维体数据不能较为真实地反映关键解剖结构的情况。
上述从三维体数据中会死别处关键解剖结构还可以是半自动的方式，可以是通过上述方法，可以识别出特定解剖结构在体数据中的位置。在全自动方法无法准确识别 的情况下，也可以让用户手动通过键盘、鼠标等工具，通过一定的工作流对检测结构进行补充、删除、修改等操作，实现半自动的解剖结构检测，例如，用鼠标。
在一个实施例中，根据关键解剖结构，确定子宫内胎儿的数量包括：在关键解剖结构为多个的情况下，获取分别依据多个关键解剖结构确定的子宫内胎儿的数量；确定一致性最高的数量为子宫内胎儿的数量。
在根据关键解剖结构确定胎儿数量时，由于上述关键解剖结构的数量可以为多个，例如，在早孕期，妊娠囊、卵黄囊及胚芽等均可以作为关键解剖结构。在确定胎儿数量时，根据多个关键解剖结构分别确定的胎儿数量，选取一致性最高的数量为子宫内胎儿数量。上述胎儿数量的一致性可以是，上述多个关键解剖结构中，识别出的胎儿数量相同的关键解剖结构的个数，在全部关键解剖结构的总数量中所占比例。例如，在上述早孕期中，可以有三个关键解剖结构分别为妊娠囊、卵黄囊和胚芽，在确定胎儿数量时，根据妊娠囊确定为一个胎儿，根据卵黄囊确定为一个胎儿，根据胚芽确定两个胎儿，，则一个胎儿的一致性为三分之二，两个胎儿的一致性为三分之一，确定一致性最高的数量为子宫内胎儿数量，也即是确定一个胎儿为子宫内胎儿数量，也即是确定子宫内的胎儿数量为一个。
图3是根据本发明实施例的另一种超声成像方法的流程图，如图3所示，根据本发明实施例的另一方面，还提供了另一种超声成像方法，该方法包括以下步骤：
步骤S302，显示子宫的三维体数据，其中，三维体数据是通过超声以覆盖整个子宫区域的方式对子宫进行扫查后得到的数据；
步骤S304，显示子宫所处的孕期，以及显示与孕期对应的胎儿的关键解剖结构；
步骤S306，显示根据关键解剖结构确定的子宫内胎儿的数量。
上述步骤的执行主体可以是显示设备。通过上述显示步骤，根据从三维体数据中识别出的关键解剖结构，确定子宫内胎儿的数量的方式，关键解剖结构能在一定程度上较为准确地体现子宫内胎儿的数量，因此，对关键解剖结构的识别能够实现对子宫内胎儿的识别，依据识别出的关键解剖结构确定宫内胎儿的数量，达到了有效准确地检测子宫内胎儿数量的目的，从而实现了提高胎儿数量检测的准确率的技术效果，进而解决了相关技术中对胎儿数量检测的准确率较低的技术问题。
作为显示设备端，可以由显示设备的处理器执行数据处理和获取，由显示设备进行显示。还可以根据处理装置接收和处理数据，并由处理装置将显示的数据发送给显示设备由显示设备显示。
在一个实施例中，显示与所述孕期对应的胎儿的关键解剖结构包括：显示能够区别是否为关键解剖结构的特征；显示根据所述特征从所述三维体数据中识别出的至少一个区域；突出显示目标区域，其中，所述目标区域被判定为所述关键解剖结构的概率最大。
在一个实施例中，显示与所述孕期对应的胎儿的关键解剖结构包括：显示对所述三维体数据的影像中的像素点进行分类后，得到的像素轮廓，其中，所述像素轮廓用于区分关键解剖结构与非关键解剖结构。
在一个实施例中，显示与所述孕期对应的胎儿的关键解剖结构包括：显示结构模板中的关键解剖结构，其中，所述结构模板中包括多种真实的关键解剖结构；显示根据所述结构模板从所述三维体数据中识别出的目标区域，其中，所述目标区域为与所述结构模板中的关键解剖结构匹配度最高的区域。
在一个实施例中，显示与所述孕期对应的胎儿的关键解剖结构包括：显示从所述三维体数据中识别出的待定关键解剖结构；显示输入的操作；显示根据所述操作对所述待定关键解剖结构进行调整后，得到的所述关键解剖结构。
在一个实施例中，孕期包括：孕周小于预定周数的早孕期，或者孕周大于预定周数的早中孕期。
显示设备显示上述显示子宫所处的孕期：包括孕周小于预定周数的早孕期，或者孕周大于预定周数的早中孕期，通过上述显示设备，可以对医生或者检测者，进行提示，便于医生或者检测者在需要根据检测情况进行合理推测或者判断时，进行参考。
图4是根据本发明实施例的另一种超声成像方法的流程图，如图4所示，根据本发明实施例的另一方面，还提供了另一种超声成像方法，该方法包括以下步骤：
步骤S402，获取子宫的三维体数据，其中，三维体数据是由经超声对子宫进行扫查后得到的数据；
步骤S404，从三维体数据中识别出关键解剖结构；
步骤S406，根据关键解剖结构，确定子宫内胎儿的数量。
通过上述步骤，根据从三维体数据中识别出的关键解剖结构，确定子宫内胎儿的数量的方式，关键解剖结构能在一定程度上较为准确地体现子宫内胎儿的数量，因此，对关键解剖结构的识别能够实现对子宫内胎儿的识别，依据识别出的关键解剖结构确定宫内胎儿的数量，达到了有效准确地检测子宫内胎儿数量的目的，从而实现了提高胎儿数量检测的准确率的技术效果，进而解决了相关技术中对胎儿数量检测的准确率 较低的技术问题。
需要说明的是，本发明实施例中，还提供了一种子宫内胎儿数量的检测方法，。该检测方法可以作为本实施例的一种优选实施方式，下面对该实施方式进行详细说明。
超声技术由于其具有安全可靠、快速便捷、可重复检查等优点，已经成为医学影像检查中应用最广、使用频率最高、新技术普及速度最快的检查手段。超声技术和人工智能技术的发展进一步推动了临床诊疗技术的进步。但是，我国人口基数大，随着二胎政策的全面开放，医院的超声检查面临更加严峻的考验，需求越来越旺盛。同时，医疗资源分配不平衡、基层医生专业技术能力还有很大提升空间。超声设备的智能化，可以让医院获得更多的共享资源和技术支持，系统性地降低成本；帮助医生提高检查效率、降低误诊率；给患者提供更精准的诊断建议及个性化治疗方案。因此，研制和开发智能化超声产品于社会各个阶层都具有极大重要性和必要性。
产科超声是超声诊断应用最广泛的领域之一，在产科早中孕超声检查中，确定存活胎儿数目是其他所有检查的基础，误诊会带来一系列严重问题。
在早孕期，胚胎还未发育好，以孕囊的形态存在，其位置可以在子宫宫底、前壁、后壁、上部和中部等；二维超声检查无法获取子宫的空间位置信息，要准确判断胎儿数目，需要医生具有一定的抽象空间想象力。对于缺乏经验的超声临床医生，在子宫环境比较复杂的情况下容易出现漏算和误算。例如，早孕期因孕囊着床常伴有宫腔内少量出血、粘液积贮，致子宫包蜕膜与壁蜕膜分离而显示“双囊征”，超声检查中较容易与真正的双妊娠囊误检，也容易将真正的双妊娠囊误检为一个妊娠囊，另一个解释为出血；在中晚孕期，有时会在妊娠中出现双胎输血综合征，一胎会因为羊水过少而“粘连”在子宫壁上，很容易将其遗漏而仅发现另一羊水过多的胎儿；或者在多胎妊娠中一胎在较早时期死亡，形成“纸样胎儿”，在超声检查过程中，很可能将其漏诊或误认为是胎盘囊肿或脐带囊肿等。
本实施方式提供了一种自动统计胎儿数量的方法和装置，在医生完成3D超声数据采集后，该方法和装置可以自动识别不同孕期的解剖结构，统计胎儿数目，解决了超声检查中多胞胎易出现数量统计错误的问题，同时可以节省产前检查时间，降低对超声临床医生的技术依赖性。
在超声产前检查过程中，二维超声只能获取检查对象的单一切面信息，容易导致多胞胎的误诊和漏诊，具有一定的局限性。三维超声弥补了二维超声空间显像的不足，能够通过多种成像模式完整显示解剖部位的立体形态、内部结构及其与周围组织的空间位置关系。但是双胞胎或多胞胎的准确统计费时费力，特别是在子宫内部环境比较 复杂和胎儿体位不易观察的情况下，对超声临床医生的经验要求较高。本实施方式基于孕妇子宫体数据，通过模式识别或机器学习算法，对整个子宫范围的三维体数据进行处理，自动识别胎儿关键解剖部位，可以准确快速统计出胎儿数量。
图5是根据本发明实施方式的胎儿数量测量方法的流程图，如图5所示，为实现胎儿数量的自动统计，本实施方式技术方案的实现过程分为三个步骤，分别是：获取孕妇子宫的三维体数据；自动识别三维体数据中的胎儿关键解剖结构；以及根据识别出的关键结构数量确定胎儿数量。三个步骤的具体细节如下：
步骤1，获取孕妇子宫三维体数据；
要实现全子宫的胎儿数量统计，需要获得整个子宫区域的三维超声数据，在扫查过程中可将ROI(Region of Interest，超声成像区域)及扇扫角度设置成足够大，使得扫描范围覆盖整个子宫区域。由于早中孕阶段的子宫区域较小，因此三维超声通常是可以扫描到整个子宫区域的。
步骤2，识别三维体数据中的胎儿关键解剖结构；
获取子宫三维体数据后，系统需要识别关键解剖结构，可以按照早孕和早中孕两种情况进行区分识别。识别方法可以是半自动，也可以是全自动的方式。
本实施方式中，各孕期的关键解剖结构均可以进行自动识别。对于小于8周(大概孕周，不需要很精确，下同)的早孕阶段可以检测到妊娠囊、卵黄囊及胚芽等关键解剖结构。超声首先发现的妊娠标志就是妊娠囊，经腹超声一般在停经后5～6周可发现妊娠囊，而经阴道超声在末次月经后4周可见妊娠囊。妊娠5～6周后，经阴道超声检查，正常妊娠100％可显示卵黄囊，同时可以检测到胚芽和心脏搏动。
对于大于8周的早中孕阶段，可以检测到胎儿的颅脑、躯干、股骨、脊柱等关键解剖结构。腹部探头可在孕9周左右观察到胎儿轮廓，胎盘雏形已经明显；孕周12周以后，胎儿逐步发育完全，颅脑光环清晰，脊柱，四肢等关键解剖结构都已经出现，均可作为识别胎儿的依据。
对于关键解剖部位的识别，可以是全自动方式，也可以是半自动方式。可以检测关键解剖结构，也可检测整个胎儿。自动检测关键解剖结构的方法有很多。可以采用机器学习或深度学习的方法在三维体数据中检测出关键解剖结构。例如，早孕中的胚芽，可事先收集一定数量的早孕胚芽图像(称为正样本)，以及一定数量的非胚芽图像(称为负样本)，然后基于机器学习或者深度学习算法，设计人工神经网络，利用多层网络结构自动学习出能够区分正样本和负样本的特征，利用这些特征在检测时遍历三 维体数据中所有可能的区域，计算该区域被判断为正样本的概率，选择概率最大的区域为目标区域。传统机器学习算法需要事先基于一定的特征提取方法进行特征提取(如灰度、纹理和空间信息等)，此类方法常见的有Adaboost算法、支持向量机(SVM)、随机森林(Random Forest)等；深度学习算法可以基于多帧二维视频或者三维体数据直接进行特征提取和网络训练，进行有效的解剖结构检测，此类方法常见的有卷积神经网络算法(CNN)、递归神经网络算法(RNN)、FastRCNN、YOLO、SSD等等。
除了模式识别的方法，还可以通过图像分割方法，精确分割出三维体数据中的关键解剖结构。
分割是对影像中每个像素点属于哪一个类别进行分类，能够直接得到图像中关键解剖结构的轮廓和位置。常用的分割方法包括水平集(LevelSet)、图割(Graph Cut)、Snake、随机游走(Random walker)、分水岭算法、阈值分割等方法；除了传统方法外，深度学习的方法也能实现关键解剖结构的分割，例如FCN，UNet，SegNet，Deeplab等。
也可以采用模板匹配的方法在三维体数据中检测出一些关键解剖结构，例如，早中孕时期的胚胎结构比较固定，可以事先收集一些早中孕胚胎数据建立模板，在检测时遍历三维体数据中所有可能的区域，和模板进行相似度匹配，选择相似度最高的区域为目标区域。
通过以上一种或多种方法，可以识别出特定解剖结构在三维体数据中的位置。如果全自动方法无法准确识别，也可以让用户通过键盘、鼠标等工具，通过一定的工作流对检测结构进行补充、删除、修改等操作，实现半自动的解剖结构检测，例如，用鼠标。
以上提到的机器学习和模式识别算法都是一些用于识别胚胎或胎儿关键解剖结构的算法，本实施方式的核心是通过关键解剖结构数量确定胎儿数量，使用其他一些方法也可以达到检测解剖结构的目的，并没有改变实质过程。
步骤3，根据关键解剖结构的数量确定胎儿数量；
识别出三维体数据中的关键结构后，就可以根据关键结构统计胎儿数量。早孕阶段的胎儿数量统计可通过识别到的孕囊和胚芽个数确定的；在早中孕阶段可以根据识别到的颅脑、躯干、股骨、脊柱等关键解剖结构的数量来确定胎儿数量。
系统识别一种解剖结构就可以判断胎儿数量，例如，在三维体数据中检测到1个胎儿颅脑，即说明三维体数据中只有1个胎儿。但是，任何检测算法都存在一定的误 检率，为提高识别准确率，本发实施方式中可同时识别多个关键解剖结构实现，再最终统计胎儿数量阶段使用多个关键解剖结构数量的投票策略确定，例如，系统一共检测了5个解剖结构，通过4个解剖结构可推断出三维体数据中有2个胎儿，另一个结构推断出体数据中只有1个探头，根据投票原则，可以最终推断三维体数据有2个胎儿。
最终将胎儿统计数量在图像界面进行显示。同时也可输出结构的可信度，对于一些图像不好的体数据，可能结构识别存在误差，这时可以一个低可信度值，用以提醒医生注意复核。
本实施方式的关键点是：通过识别子宫三维体数据中关键解剖结构或整个胚胎结构来确定子宫内胎儿数量的方法。体数据获取步骤，用于获取三维体数据；解剖结构识别步骤，用于识别体数据中的解剖结构；确定胎儿数量步骤，用于确定胎儿数量，并在图像界面进行显示。
图6是根据本发明实施例的一种超声成像设备的结构示意图，根据本发明实施例的另一方面，还提供了一种超声成像设备，包括：探头602，发射电路604，接收电路606，处理器608和显示器610，下面对该设备进行详细说明。
探头602；发射电路604，与上述探头602相连，发射电路激励探头向子宫发射超声波；接收电路606，与上述探头602相连，接收电路通过探头接收从子宫返回的超声回波以获得超声回波信号；处理器608，与上述接收电路606相连，处理器处理超声回波信号以获得子宫的三维体数据；显示器610，与上述处理器608相连，显示器显示三维体数据；其中，处理器608还执行如下步骤：从三维体数据中识别出关键解剖结构；并根据关键解剖结构，确定子宫内胎儿的数量。
在一个实施例中，显示器610，还用于显示以下至少之一：关键解剖结构，子宫内胎儿的数量，子宫所处的孕期。
根据本发明实施例的另一方面，还提供了一种存储介质，存储介质包括存储的程序，其中，在程序运行时控制存储介质所在设备执行上述中任意一项的超声成像方法。
根据本发明实施例的另一方面，还提供了一种处理器，处理器用于运行程序，其中，程序运行时执行上述中任意一项的超声成像方法。
根据本发明实施例的另一方面，还提供了一种计算机设备，包括：存储器和处理器，存储器存储有计算机程序；处理器，用于执行存储器中存储的计算机程序，计算机程序运行时执行上述中任意一项的超声成像方法。
上述本发明实施例序号仅仅为了描述，不代表实施例的优劣。
在本发明的上述实施例中，对各个实施例的描述都各有侧重，某个实施例中没有详述的部分，可以参见其他实施例的相关描述。
在本申请所提供的几个实施例中，应该理解到，所揭露的技术内容，可通过其它的方式实现。其中，以上所描述的装置实施例仅仅是示意性的，例如所述单元的划分，可以为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，单元或模块的间接耦合或通信连接，可以是电性或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本发明各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本发明的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可为个人计算机、服务器或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、移动硬盘、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述仅是本发明的优选实施方式，应当指出，对于本技术领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。

Claims (21)

  1. 一种超声成像方法，其特征在于，包括：
    通过覆盖整个子宫区域的方式向子宫发射超声波，并接收超声回波，获得超声回波信号；
    根据所述超声回波信号获得子宫的三维体数据；
    确定所述子宫所处的孕期；
    从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结构；
    根据所述关键解剖结构，确定所述子宫内胎儿的数量。
  2. 根据权利要求1所述的方法，其特征在于，所述孕期包括：孕周小于预定周数的早孕期，或者孕周大于预定周数的早中孕期。
  3. 根据权利要求2所述的方法，其特征在于，与所述早孕期对应的关键解剖结构包括以下至少之一：妊娠囊、卵黄囊及胚芽；与所述早中孕期对应的关键解剖结构包括以下至少之一：颅脑、躯干、股骨、脊柱及四肢。
  4. 根据权利要求1所述的方法，其特征在于，从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结构包括：
    获取能够区别是否为关键解剖结构的特征；
    根据所述特征从所述三维体数据中识别出至少一个区域；
    从所述至少一个区域中，确定出目标区域，其中，所述目标区域被判定为所述关键解剖结构的概率最大；
    确定所述目标区域为所述关键解剖结构。
  5. 根据权利要求4所述的方法，其特征在于，所述特征包括以下至少之一：二维特征和三维特征。
  6. 根据权利要求5所述的方法，其特征在于，获取能够区别是否为关键解剖结构的特征包括：
    收集确定为所述关键解剖结构的正样本，和确定不为所述关键解剖结构的负样本；
    基于机器学习，对所述正样本和所述负样本进行训练，得到能够区别是否为关键解剖结构的特征。
  7. 根据权利要求1所述的方法，其特征在于，从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结构包括：
    对所述三维体数据的影像中的像素点进行分类，得到分类结果；
    根据所述分类结果识别出关键解剖结构。
  8. 根据权利要求1所述的方法，其特征在于，从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结构包括：
    确定结构模板，其中，所述结构模板中包括多种真实的关键解剖结构；
    根据所述结构模板从所述三维体数据中识别出目标区域，其中，所述目标区域为与所述结构模板中的关键解剖结构匹配度最高的区域；
    确定所述目标区域为所述关键解剖结构。
  9. 根据权利要求1至8中任一项所述的方法，其特征在于，从所述三维体数据中识别出与所述孕期对应的胎儿的关键解剖结构包括：
    从所述三维体数据中识别出待定关键解剖结构；
    通过接收输入的操作的方式，对所述待定关键解剖结构进行调整，得到所述关键解剖结构。
  10. 根据权利要求1至8中任一项所述的方法，其特征在于，根据所述关键解剖结构，确定所述子宫内胎儿的数量包括：
    在所述关键解剖结构为多个的情况下，获取分别依据多个关键解剖结构确定的所述子宫内胎儿的数量；
    确定一致性最高的数量为所述子宫内胎儿的数量。
  11. 一种超声成像方法，其特征在于，包括：
    显示子宫的三维体数据，其中，所述三维体数据是通过超声以覆盖整个子宫区域的方式对子宫进行扫查后得到的数据；
    显示所述子宫所处的孕期，以及显示与所述孕期对应的胎儿的关键解剖结构；
    显示根据所述关键解剖结构确定的所述子宫内胎儿的数量。
  12. 根据权利要求11所述的方法，其特征在于，显示与所述孕期对应的胎儿的关键解剖结构包括：
    显示能够区别是否为关键解剖结构的特征；
    显示根据所述特征从所述三维体数据中识别出的至少一个区域；
    突出显示目标区域，其中，所述目标区域被判定为所述关键解剖结构的概率最大。
  13. 根据权利要求11所述的方法，其特征在于，显示与所述孕期对应的胎儿的关键解剖结构包括：
    显示对所述三维体数据的影像中的像素点进行分类后，得到的像素轮廓，其中，所述像素轮廓用于区分关键解剖结构与非关键解剖结构。
  14. 根据权利要求11所述的方法，其特征在于，显示与所述孕期对应的胎儿的关键解剖结构包括：
    显示结构模板中的关键解剖结构，其中，所述结构模板中包括多种真实的关键解剖结构；
    显示根据所述结构模板从所述三维体数据中识别出的目标区域，其中，所述目标区域为与所述结构模板中的关键解剖结构匹配度最高的区域。
  15. 根据权利要求11至14中任一项所述的方法，其特征在于，显示与所述孕期对应的胎儿的关键解剖结构包括：
    显示从所述三维体数据中识别出的待定关键解剖结构；
    显示输入的操作；
    显示根据所述操作对所述待定关键解剖结构进行调整后，得到的所述关键解剖结构。
  16. 一种超声成像方法，其特征在于，包括：
    获取子宫的三维体数据，其中，所述三维体数据是由经超声对子宫进行扫查后得到的数据；
    从所述三维体数据中识别出关键解剖结构；
    根据所述关键解剖结构，确定所述子宫内胎儿的数量。
  17. 一种超声成像设备，其特征在于，包括：
    探头；
    发射电路，所述发射电路激励所述探头向子宫发射超声波；
    接收电路，所述接收电路通过所述探头接收从所述子宫返回的超声回波以获得超声回波信号；
    处理器，所述处理器处理所述超声回波信号以获得所述子宫的三维体数据；
    显示器，所述显示器显示所述三维体数据；
    其中，所述处理器还执行如下步骤：从所述三维体数据中识别出关键解剖结构；并根据所述关键解剖结构，确定所述子宫内胎儿的数量。
  18. 根据权利要求17所述的设备，其特征在于，
    所述显示器，还用于显示以下至少之一：所述关键解剖结构，所述子宫内胎儿的数量，子宫所处的孕期。
  19. 一种存储介质，其特征在于，所述存储介质包括存储的程序，其中，在所述程序运行时控制所述存储介质所在设备执行权利要求1至17中任意一项所述的超声成像方法。
  20. 一种处理器，其特征在于，所述处理器用于运行程序，其中，所述程序运行时执行权利要求1至16中任意一项所述的超声成像方法。
  21. 一种计算机设备，其特征在于，包括：存储器和处理器，
    所述存储器存储有计算机程序；
    所述处理器，用于执行所述存储器中存储的计算机程序，所述计算机程序运行时执行权利要求1至16中任意一项所述的超声成像方法。
PCT/CN2018/117007 2018-11-22 2018-11-22 超声成像方法、设备、存储介质，处理器及计算机设备 WO2020103098A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201880097331.8A CN112654299A (zh) 2018-11-22 2018-11-22 超声成像方法、设备、存储介质,处理器及计算机设备
PCT/CN2018/117007 WO2020103098A1 (zh) 2018-11-22 2018-11-22 超声成像方法、设备、存储介质，处理器及计算机设备

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/117007 WO2020103098A1 (zh) 2018-11-22 2018-11-22 超声成像方法、设备、存储介质，处理器及计算机设备

Publications (1)

Publication Number Publication Date
WO2020103098A1 true WO2020103098A1 (zh) 2020-05-28

Family

ID=70773452

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/117007 WO2020103098A1 (zh) 2018-11-22 2018-11-22 超声成像方法、设备、存储介质，处理器及计算机设备

Country Status (2)

Country Link
CN (1) CN112654299A (zh)
WO (1) WO2020103098A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116392166A (zh) * 2023-03-07 2023-07-07 广州市妇女儿童医疗中心 一种孕囊及胚芽自动测量方法、系统、装置及介质

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115919367A (zh) * 2022-12-09 2023-04-07 开立生物医疗科技(武汉)有限公司 超声图像处理方法及装置、电子设备和存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101732070A (zh) * 2008-11-18 2010-06-16 株式会社东芝 超声波图像取得装置
US20110201935A1 (en) * 2008-10-22 2011-08-18 Koninklijke Philips Electronics N.V. 3-d ultrasound imaging
CN104394771A (zh) * 2012-06-04 2015-03-04 泰尔哈绍梅尔医学研究基础设施和服务有限公司 超声图像处理
CN104414680A (zh) * 2013-08-21 2015-03-18 深圳迈瑞生物医疗电子股份有限公司 一种三维超声成像方法及系统

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7783095B2 (en) * 2005-08-11 2010-08-24 Siemens Medical Solutions Usa, Inc. System and method for fetal biometric measurements from ultrasound data and fusion of same for estimation of fetal gestational age
US8556814B2 (en) * 2007-10-04 2013-10-15 Siemens Medical Solutions Usa, Inc. Automated fetal measurement from three-dimensional ultrasound data
CN102949206B (zh) * 2011-08-26 2015-12-02 深圳迈瑞生物医疗电子股份有限公司 一种三维超声成像的方法及装置
US20170238907A1 (en) * 2016-02-22 2017-08-24 General Electric Company Methods and systems for generating an ultrasound image

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110201935A1 (en) * 2008-10-22 2011-08-18 Koninklijke Philips Electronics N.V. 3-d ultrasound imaging
CN101732070A (zh) * 2008-11-18 2010-06-16 株式会社东芝 超声波图像取得装置
CN104394771A (zh) * 2012-06-04 2015-03-04 泰尔哈绍梅尔医学研究基础设施和服务有限公司 超声图像处理
CN104414680A (zh) * 2013-08-21 2015-03-18 深圳迈瑞生物医疗电子股份有限公司 一种三维超声成像方法及系统

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116392166A (zh) * 2023-03-07 2023-07-07 广州市妇女儿童医疗中心 一种孕囊及胚芽自动测量方法、系统、装置及介质
CN116392166B (zh) * 2023-03-07 2024-02-23 广州市妇女儿童医疗中心 一种孕囊及胚芽自动测量方法、系统、装置及介质

Also Published As

Publication number Publication date
CN112654299A (zh) 2021-04-13

Similar Documents

Publication Publication Date Title
Torkildsen et al. Prediction of delivery mode with transperineal ultrasound in women with prolonged first stage of labor
Bega et al. Three‐dimensional ultrasonography in gynecology: technical aspects and clinical applications
Chen et al. Artificial intelligence in obstetric ultrasound: an update and future applications
US11622744B2 (en) System and method for automated ovarian follicular monitoring
KR20170053685A (ko) 초음파 측정치 및 그래픽 모델을 산출하여 표시함에 의한 태아 영상화 방법 및 시스템
Cobo et al. Feasibility and reproducibility of fetal lung texture analysis by automatic quantitative ultrasound analysis and correlation with gestational age
Hu et al. Automated placenta segmentation with a convolutional neural network weighted by acoustic shadow detection
CN110279433A (zh) 一种基于卷积神经网络的胎儿头围自动精确测量方法
Martínez-Payo et al. Detection of single umbilical artery in the first trimester ultrasound: its value as a marker of fetal malformation
Sainz et al. Intra and interobserver variability of intrapartum transperineal ultrasound measurements with contraction and pushing
WO2020103098A1 (zh) 超声成像方法、设备、存储介质，处理器及计算机设备
Du et al. Quantitative elastography of cervical stiffness during the three trimesters of pregnancy with a semiautomatic measurement program: a longitudinal prospective pilot study
Conversano et al. Automatic ultrasound technique to measure angle of progression during labor
WO2020215485A1 (zh) 胎儿生长参数测量方法、系统及超声设备
Saavedra et al. Automatic ultrasound assessment of placenta previa during the third trimester for rural areas
CN113520317A (zh) 基于oct的子宫内膜检测分析方法、装置、设备及存储介质
KR20140140712A (ko) 초음파 영상 처리 장치 및 방법
CN116763347A (zh) 基于超声图像的胎头方向角测量方法以及相关装置
Shiney et al. A Review on techniques for computer aided diagnosis of soft markers for detection of down syndrome in ultrasound fetal images
EP3689248B1 (en) Ultrasound imaging apparatus and method of generating ultrasound image
WO2018040293A1 (zh) 一种b超图像的处理方法及其装置
WO1994014132A1 (en) Non-invasive medical scanning
CN115919367A (zh) 超声图像处理方法及装置、电子设备和存储介质
Ramphul et al. Role of ultrasound on the labor ward
Vásquez et al. Labor induction failure prediction based on B-mode ultrasound image processing using multiscale local binary patterns

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18940580

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18940580

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 18940580

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 01/02/2022)