WO2020199694A1 - 一种脊柱Cobb角测量方法、装置、可读存储介质及终端设备 - Google Patents

一种脊柱Cobb角测量方法、装置、可读存储介质及终端设备 Download PDF

Info

Publication number
WO2020199694A1
WO2020199694A1 PCT/CN2019/130900 CN2019130900W WO2020199694A1 WO 2020199694 A1 WO2020199694 A1 WO 2020199694A1 CN 2019130900 W CN2019130900 W CN 2019130900W WO 2020199694 A1 WO2020199694 A1 WO 2020199694A1
Authority
WO
WIPO (PCT)
Prior art keywords
spine
cobb angle
pixel
area
block
Prior art date
Application number
PCT/CN2019/130900
Other languages
English (en)
French (fr)
Inventor
谭志强
苗燕茹
孙宇
李猛
胡颖
徐艳雯
Original Assignee
中国科学院深圳先进技术研究院
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 中国科学院深圳先进技术研究院 filed Critical 中国科学院深圳先进技术研究院
Publication of WO2020199694A1 publication Critical patent/WO2020199694A1/zh

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/107—Measuring physical dimensions, e.g. size of the entire body or parts thereof
    • A61B5/1071—Measuring physical dimensions, e.g. size of the entire body or parts thereof measuring angles, e.g. using goniometers
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B5/00—Measuring arrangements characterised by the use of mechanical techniques
    • G01B5/24—Measuring arrangements characterised by the use of mechanical techniques for measuring angles or tapers; for testing the alignment of axes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands

Definitions

  • the invention belongs to the technical field of image analysis, and in particular relates to a method, a device, a computer-readable storage medium and a terminal device for measuring the Cobb angle of the spine.
  • the traditional method of measuring the Cobb angle of the spine is to manually measure it on X-rays with a pencil and a protractor.
  • the staff finds out the end vertebra with the largest inclination through experience. Scoliosis will roughly show three bends, and the largest bend is the Cobb angle. Because this method is performed manually, there are certain errors, and the reliability within and between observers is also poor.
  • This method requires manual drawing of the extension line of the end plate and then making the vertical line and measuring with a protractor. These manual operations have brought various errors and caused differences between and within observers.
  • the application of different pencils and protractors, the different widths of the lines drawn, the quality of the X-ray film and other factors will cause measurement errors.
  • the embodiments of the present invention provide a method, device, computer-readable storage medium, and terminal device for measuring the Cobb angle of the spine to solve the reliability of the existing method for measuring the Cobb angle of the spine by manual or semi-manual methods. Poor problem.
  • the first aspect of the embodiments of the present invention provides a method for measuring the Cobb angle of the spine, which may include:
  • segmentation result map being a binary image of the spine region and the non-spine region
  • the traverse calculation is performed on the straight lines of the upper and lower endplates of each spine block, and the angle with the largest value is determined as the Cobb angle of the spine.
  • the identifying each spine block from the spine area of the segmentation result map includes:
  • the method of using the smallest outer envelope rectangle separately identifies each spine block from the spine area of the segmentation result map.
  • determining the upper and lower endplate straight lines of each spine block respectively includes:
  • the pixel value of each spine block is detected, and the pixel in the area where the first type of pixel mutation occurs is determined as the upper endplate point, and the pixel in the second type of pixel mutation area is determined as the lower endplate point.
  • the first type of pixel mutation area is an area that changes from a preset first pixel value to a preset second pixel value
  • the second type of pixel mutation area is an area that changes from the second pixel value to the first pixel value.
  • the least square method is used to perform function fitting on the upper endplate points of each spine block to obtain the upper endplate straight line of each spine block;
  • the least square method was used to perform function fitting on the lower endplate points of each spine block, and the straight line of the lower endplate of each spine block was obtained.
  • the pre-trained deep learning network model before using the pre-trained deep learning network model to segment the spine X-ray image, it also includes:
  • Training the deep learning network model by using the sample image and the truth map of the sample image to obtain a trained deep learning network model.
  • the deep learning network model is a U-net segmentation network.
  • a second aspect of the embodiments of the present invention provides a spine Cobb angle measuring device, which may include:
  • the image segmentation module is used to segment the spine X-ray image using a pre-trained deep learning network model to obtain a segmentation result map, the segmentation result map being a binary image of the spine region and the non-spine region;
  • a spine block identification module which is used to identify each spine block from the spine area of the segmentation result map
  • the upper and lower endplate determining module is used to determine the straight lines of the upper and lower endplates of each spine block;
  • the spine Cobb angle determination module is used to traverse the upper and lower endplate lines of each spine block, and determine the angle with the largest value as the spine Cobb angle.
  • the spine block recognition module is specifically configured to use the smallest outer envelope rectangle method to recognize each spine block from the spine area of the segmentation result map.
  • the upper and lower endplate determining module may include:
  • the pixel detection unit is used to detect the pixel value of each spine block, and determine the pixel in the first type of pixel mutation area as the upper end plate point, and determine the pixel in the second type of pixel mutation area Is a lower end plate point, the first type of pixel mutation area is an area that changes from a preset first pixel value to a preset second pixel value, and the second type of pixel mutation area is from the second pixel The area where the value becomes the first pixel value;
  • the upper endplate straight line fitting unit is used to perform function fitting on the upper endplate points of each spine block by using the least square method to obtain the upper endplate straight line of each spine block;
  • the lower endplate straight line fitting unit is used to perform function fitting on the lower endplate points of each spine block by using the least square method to obtain the lower endplate straight line of each spine block.
  • the spine Cobb angle measuring device may further include:
  • the data labeling module is used for data labeling of the sample image and making a truth map of the sample image
  • the model training module is used to train the deep learning network model using the sample image and the truth map of the sample image to obtain a trained deep learning network model.
  • a third aspect of the embodiments of the present invention provides a computer-readable storage medium, the computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, any of the above-mentioned spinal Cobb Steps of angle measurement method.
  • the fourth aspect of the embodiments of the present invention provides a terminal device, including a memory, a processor, and computer-readable instructions stored in the memory and running on the processor, and the processor executes the computer
  • the steps of any of the above-mentioned methods for measuring the Cobb angle of the spine are realized when the instructions are readable.
  • the embodiment of the present invention has the beneficial effect that: the embodiment of the present invention uses a pre-trained deep learning network model to segment the spine X-ray image to obtain a segmentation result map, and the segmentation result map is the spine Binary images of regions and non-spine regions; respectively identify each spine block from the spine region of the segmentation result map; determine the upper and lower endplate lines of each spine block respectively; traverse the upper and lower endplate lines of each spine block , The angle with the largest value is determined as the Cobb angle of the spine.
  • Figure 1 is a schematic diagram of the traditional measurement method of the Cobb angle of the spine
  • FIG. 2 is a flowchart of an embodiment of a method for measuring Cobb angle of the spine in an embodiment of the present invention
  • Figure 3 is a schematic diagram of making a truth map
  • Fig. 5 is a schematic diagram of a specific example of measuring the Cobb angle of the spine
  • FIG. 6 is a structural diagram of an embodiment of a device for measuring Cobb angle of the spine in an embodiment of the present invention
  • Fig. 7 is a schematic block diagram of a terminal device in an embodiment of the present invention.
  • an embodiment of a method for measuring Cobb angle of the spine in an embodiment of the present invention may include:
  • Step S201 Use the pre-trained deep learning network model to perform segmentation processing on the spine X-ray image to obtain a segmentation result map.
  • the deep learning network model may be a U-net segmentation network.
  • the spine needs to be extracted from the X-ray image, and the X-ray image is a grayscale image.
  • the task can be simplified to a two-classification problem, that is, the image is divided into spine and non-spine areas.
  • the deep learning network model is trained to obtain a trained deep learning network model. That is to say, supervise the learning of the segmentation network by making a good truth map.
  • the truth map provides a "gold standard" in the deep learning segmentation network, that is, when the network learns features, it refers to the truth map for parameter learning, so , The quality of the truth map directly affects the result of segmentation. Because there is a lot of redundant information in the X-ray image of the spine, experienced staff are required to make the data set.
  • Figure 3 is a schematic diagram of making a truth map
  • the staff marked the spine area the pixel value of the marked spine area was changed to 0, and the background area, that is, the non-spine area, was changed to 1.
  • the input data pixels are only 0 and 1, and the data is processed as above to complete the process of making training data.
  • the upper and lower endplates of the spine are extracted through the staff’s anatomical experience, instead of simply extracting the entire spine block area.
  • Using such a truth map production method can improve segmentation The accuracy is so that the later Cobb angle measurement can accurately calculate the angle size.
  • the original image data of the spine X-ray image is relatively small, and deep learning requires a lot of data for training. Therefore, after the data labeling work is completed, the data set needs to be expanded, and the original image is transformed by rotation, offset, etc. Expand to an ideal state.
  • the data enhancement code that comes with Keras is used.
  • Keras is a framework based on tensorflow developed by Google, which is very convenient to use, and the network construction work can be completed by directly using its modules.
  • network training can be performed.
  • the entire training process is based on the Keras framework, the programming language used is Pyhton, the version is 3.6, and the compiler is Pycharm.
  • the experimental platform is based on Windows 7 system, and the graphics card used is With NVIDIA GeForce GTX 1080 Ti, a segmentation result map can be obtained after a period of training, and the segmentation result map is a binary image of the spine region and the non-spine region.
  • Step S202 Identify each spine block from the spine area of the segmentation result map.
  • the shape of the spine block is similar to a rectangle, in this embodiment, it is preferable to use the method of the smallest outer envelope rectangle to identify each spine block from the spine region of the segmentation result map, and to compare each segment in the segmentation result map as shown.
  • the spine blocks are all enveloped so that each spine block can be manipulated individually.
  • Step S203 Determine the upper and lower endplate straight lines of each spine block respectively.
  • the pixel value of each spine block is detected first, and the pixel in the area where the first type of pixel mutation occurs is determined as the upper end plate point, and the pixel in the second type of pixel mutation area is determined as the bottom end. Board point.
  • the first type of pixel mutation area is an area that changes from a preset first pixel value to a preset second pixel value
  • the second type of pixel mutation area is an area that changes from the second pixel value to the first pixel value.
  • the first pixel value is the pixel value of the non-spine area
  • the second pixel value is the pixel value of the spine area.
  • the least square method is used to perform function fitting on the upper endplate points of each spine block to obtain the upper endplate straight line of each spine block.
  • the least square method is used to perform function fitting on the inferior endplate points of each spine block to obtain the inferior endplate straight line of each spine block.
  • the least square method is used Fitting the red point group to the function can get the upper endplate straight line of the spine block.
  • using the least squares method to fit the blue point group to the function fitting can get the lower endplate straight line of the spine block, and then the same way Operate each spine block to obtain the upper and lower endplate straight lines of each spine block.
  • the fitted upper endplate line and the lower endplate line can be stored separately for subsequent use.
  • Step S204 Perform traversal calculation on the upper and lower endplate straight lines of each spine block, and determine the angle with the largest value as the spine Cobb angle.
  • this embodiment can also give the specific position of the Cobb angle of the spine at the same time.
  • every spine block has a corresponding name.
  • the spine Cobb angle measurement here does not involve the cervical spine, so it is not included here.
  • the upper and lower endplates of each spine block are given corresponding names to realize the vertebral counting function.
  • Figure 4 shows a schematic diagram of the complete process of the method for measuring the Cobb angle of the spine in this embodiment.
  • Figure 5 shows a specific example of measuring the Cobb angle of the spine. In this example, the measured angle is 40 degrees.
  • the upper vertebra is T11, and the lower vertebra is L3.
  • the entire automatic counting Cobb angle code is based on the computer vision library Opencv 3.3.1 version
  • the writing language is C++ language
  • the compilation platform is Visual Stdio 2013, and the experiment platform is based on Windows 7 system
  • the graphics card used is NVIDIA GeForce GTX 1080 Ti.
  • the embodiment of the present invention uses a pre-trained deep learning network model to segment the spine X-ray image to obtain a segmentation result map, which is a binary image of the spine region and the non-spine region;
  • a segmentation result map which is a binary image of the spine region and the non-spine region;
  • Each spine block is identified in the spine region of the segmentation result map; the upper and lower endplate straight lines of each spine block are respectively determined; the upper and lower endplate straight lines of each spine block are traversed and the angle with the largest value is determined as the spine Cobb corner.
  • FIG. 6 shows a structural diagram of an embodiment of a device for measuring the Cobb angle of the spine provided by an embodiment of the present invention.
  • a device for measuring Cobb angle of the spine may include:
  • the image segmentation module 601 is configured to perform segmentation processing on the spine X-ray image using a pre-trained deep learning network model to obtain a segmentation result map, which is a binary image of the spine region and the non-spine region;
  • the spine block identification module 602 is used to identify each spine block from the spine area of the segmentation result map
  • the upper and lower endplate determining module 603 is used to separately determine the upper and lower endplate straight lines of each spine block;
  • the spine Cobb angle determination module 604 is used to traverse the upper and lower endplate lines of each spine block, and determine the angle with the largest value as the spine Cobb angle.
  • the spine block recognition module is specifically configured to use the smallest outer envelope rectangle method to recognize each spine block from the spine area of the segmentation result map.
  • the upper and lower endplate determining module may include:
  • the pixel detection unit is used to detect the pixel value of each spine block, and determine the pixel in the first type of pixel mutation area as the upper end plate point, and determine the pixel in the second type of pixel mutation area Is a lower end plate point, the first type of pixel mutation area is an area that changes from a preset first pixel value to a preset second pixel value, and the second type of pixel mutation area is from the second pixel The area where the value becomes the first pixel value;
  • the upper endplate straight line fitting unit is used to perform function fitting on the upper endplate points of each spine block by using the least square method to obtain the upper endplate straight line of each spine block;
  • the lower endplate straight line fitting unit is used to perform function fitting on the lower endplate points of each spine block by using the least square method to obtain the lower endplate straight line of each spine block.
  • the spine Cobb angle measuring device may further include:
  • the data labeling module is used for data labeling of the sample image and making a truth map of the sample image
  • the model training module is used to train the deep learning network model using the sample image and the truth map of the sample image to obtain a trained deep learning network model.
  • FIG. 7 shows a schematic block diagram of a terminal device according to an embodiment of the present invention. For ease of description, only parts related to the embodiment of the present invention are shown.
  • the spine Cobb angle measurement terminal device 7 of this embodiment includes a processor 70, a memory 71, and a computer program 72 stored in the memory 71 and running on the processor 70.
  • the processor 70 executes the computer program 72, the steps in the foregoing embodiments of the method for measuring the Cobb angle of the spine are implemented, such as steps S201 to S204 shown in FIG. 2.
  • the processor 70 executes the computer program 72, the functions of the modules/units in the foregoing device embodiments, for example, the functions of the modules 601 to 604 shown in FIG. 6 are realized.
  • the computer program 72 may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 71 and executed by the processor 70 to complete this invention.
  • the one or more modules/units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 72 in the spine Cobb angle measurement terminal device 7.
  • the spine Cobb angle measurement terminal device 7 may be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. Those skilled in the art can understand that FIG. 7 is only an example of the spine Cobb angle measurement terminal device 7 and does not constitute a limitation on the spine Cobb angle measurement terminal device 7. It may include more or less components than shown in the figure, or a combination Certain components, or different components, for example, the spine Cobb angle measurement terminal device 7 may also include input and output devices, network access devices, buses, and the like.
  • the processor 70 may be a central processing unit (Central Processing Unit, CPU), other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like.
  • the memory 71 may be an internal storage unit of the spine Cobb angle measurement terminal device 7, for example, a hard disk or a memory of the spine Cobb angle measurement terminal device 7.
  • the memory 71 may also be an external storage device of the spine Cobb angle measurement terminal device 7, for example, a plug-in hard disk or a smart media card (SMC) equipped on the spine Cobb angle measurement terminal device 7, Secure Digital (SD) card, flash memory card (Flash Card) etc.
  • the memory 71 may also include both an internal storage unit of the spine Cobb angle measurement terminal device 7 and an external storage device.
  • the memory 71 is used to store the computer program and other programs and data required by the spine Cobb angle measurement terminal device 7.
  • the memory 71 can also be used to temporarily store data that has been output or will be output.
  • the disclosed device/terminal device and method may be implemented in other ways.
  • the device/terminal device embodiments described above are only illustrative.
  • the division of the modules or units is only a logical function division, and there may be other divisions in actual implementation, such as multiple units.
  • components can be combined or integrated into another system, or some features can be omitted 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, devices or units, and may be in electrical, mechanical 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, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solutions of the embodiments.
  • the functional units in the various embodiments 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-mentioned integrated unit can be implemented in the form of hardware or software functional unit.
  • the integrated module/unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium.
  • the present invention implements all or part of the processes in the above-mentioned embodiments and methods, and can also be completed by instructing relevant hardware through a computer program.
  • the computer program can be stored in a computer-readable storage medium. When the program is executed by the processor, the steps of the foregoing method embodiments can be implemented.
  • the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file, or some intermediate forms.
  • the computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signals, telecommunications signals, and software distribution media.
  • ROM Read-Only Memory
  • RAM Random Access Memory
  • electrical carrier signals telecommunications signals
  • software distribution media any entity or device capable of carrying the computer program code
  • recording medium U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signals, telecommunications signals, and software distribution media.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Biophysics (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Pathology (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Dentistry (AREA)
  • Image Analysis (AREA)

Abstract

一种脊柱Cobb角测量方法、装置、计算机可读存储介质及终端设备，涉及图像分析技术领域，所述方法使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理，得到分割结果图（S201），所述分割结果图为脊柱区域与非脊柱区域的二值图像；从所述分割结果图的脊柱区域中分别识别出各个脊柱块（S202）；分别确定各个脊柱块的上下终板直线（S203）；对各个脊柱块的上下终板直线进行遍历计算，将取值最大的夹角确定为脊柱Cobb角（S204）。所述方法实现了真正意义上的脊柱Cobb角自动测量，无需工作人员再进行额外的操作就能得到所需要的脊柱Cobb角，避免了手工或者半手工的方式进行脊柱Cobb角测量所引入的误差，具有更好的可靠性。

Description

一种脊柱Cobb角测量方法、装置、可读存储介质及终端设备 技术领域
本发明属于图像分析技术领域，尤其涉及一种脊柱Cobb角测量方法、装置、计算机可读存储介质及终端设备。
 背景技术
脊柱Cobb角的传统测量方法是工作人员利用铅笔和量角器在X线片上手工测量。如图1所示，工作人员在进行Cobb角测量的过程中，通过经验找出倾斜度最大的端椎，脊柱侧弯大致会呈现出三个弯，而弯曲程度最大的即为Cobb角。因为这个方法是手工执行的，所以存在一定的误差，其观察者内和观察者间的可靠性也较差。这个方法需要手工绘画终板的延长线再作垂线并用量角器测量，这些手工操作都带来了各种误差，引起观察者内和观察者间的差异。应用不同的铅笔和量角器、所绘直线的不同宽度、X线片的质量等因素均会导致测量的误差。在1979年，Barry F Jefferies提出侧弯测量的5°左右差别通常是测量的误差，而不是侧弯程度本身的差异。在1994年，J E H Prujis等人发现Cobb角的测量误差在3.2°左右。近年来，出现了计算机辅助的Cobb角测量系统，例如使用较多的Surgimap，该软件提供给工作人员一个快捷的测量Cobb角的平台，其主要用法为将X光图像导入平台中，工作人员通过在片子上画点或者画线的方式将上下终板标识出来，计算机通过标识的点或者线来计算夹角。从而得出Cobb角度的大小。但使用Surgimap只是将测量角度的平台从现实中换到了电脑中，使用过程还是需要工作人员凭经验去寻找Cobb角存在的范围，然后在这个范围通过手画点或者线的方式去确定角度大小。这实际上只是换了一个场景，而且在手画线或点的过程中还是没有消除手工测量角度存在的差异性这个问题。
 技术问题
有鉴于此，本发明实施例提供了一种脊柱Cobb角测量方法、装置、计算机可读存储介质及终端设备，以解决现有的通过手工或者半手工的方式进行脊柱Cobb角测量的方法可靠性较差的问题。
 技术解决方案
本发明实施例的第一方面提供了一种脊柱Cobb角测量方法，可以包括：
使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理，得到分割结果图，所述分割结果图为脊柱区域与非脊柱区域的二值图像；
从所述分割结果图的脊柱区域中分别识别出各个脊柱块；
分别确定各个脊柱块的上下终板直线；
对各个脊柱块的上下终板直线进行遍历计算，将取值最大的夹角确定为脊柱Cobb角。
优选地，所述从所述分割结果图的脊柱区域中分别识别出各个脊柱块包括：
使用最小外包络矩形的方法从所述分割结果图的脊柱区域中分别识别出各个脊柱块。
进一步地，所述分别确定各个脊柱块的上下终板直线包括：
对各个脊柱块的像素点进行像素值检测，并将发生第一类像素突变区域的像素点确定为上终板点，将发生第二类像素突变区域的像素点确定为下终板点，所述第一类像素突变区域为从预设的第一像素值变为预设的第二像素值的区域，所述第二类像素突变区域为从所述第二像素值变为所述第一像素值的区域；
采用最小二乘法分别对各个脊柱块的上终板点进行函数拟合，得到各个脊柱块的上终板直线；
采用最小二乘法分别对各个脊柱块的下终板点进行函数拟合，得到各个脊柱块的下终板直线。
进一步地，在使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理之前，还包括：
对样本图像进行数据标注，制作所述样本图像的真值图；
使用所述样本图像以及所述样本图像的真值图对所述深度学习网络模型进行训练，得到训练好的深度学习网络模型。
优选地，所述深度学习网络模型为U-net分割网络。
本发明实施例的第二方面提供了一种脊柱Cobb角测量装置，可以包括：
图像分割模块，用于使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理，得到分割结果图，所述分割结果图为脊柱区域与非脊柱区域的二值图像；
脊柱块识别模块，用于从所述分割结果图的脊柱区域中分别识别出各个脊柱块；
上下终板确定模块，用于分别确定各个脊柱块的上下终板直线；
脊柱Cobb角确定模块，用于对各个脊柱块的上下终板直线进行遍历计算，将取值最大的夹角确定为脊柱Cobb角。
优选地，所述脊柱块识别模块具体用于使用最小外包络矩形的方法从所述分割结果图的脊柱区域中分别识别出各个脊柱块。
进一步地，所述上下终板确定模块可以包括：
像素点检测单元，用于对各个脊柱块的像素点进行像素值检测，并将发生第一类像素突变区域的像素点确定为上终板点，将发生第二类像素突变区域的像素点确定为下终板点，所述第一类像素突变区域为从预设的第一像素值变为预设的第二像素值的区域，所述第二类像素突变区域为从所述第二像素值变为所述第一像素值的区域；
上终板直线拟合单元，用于采用最小二乘法分别对各个脊柱块的上终板点进行函数拟合，得到各个脊柱块的上终板直线；
下终板直线拟合单元，用于采用最小二乘法分别对各个脊柱块的下终板点进行函数拟合，得到各个脊柱块的下终板直线。
进一步地，所述脊柱Cobb角测量装置还可以包括：
数据标注模块，用于对样本图像进行数据标注，制作所述样本图像的真值图；
模型训练模块，用于使用所述样本图像以及所述样本图像的真值图对所述深度学习网络模型进行训练，得到训练好的深度学习网络模型。
本发明实施例的第三方面提供了一种计算机可读存储介质，所述计算机可读存储介质存储有计算机可读指令，所述计算机可读指令被处理器执行时实现上述任意一种脊柱Cobb角测量方法的步骤。
本发明实施例的第四方面提供了一种终端设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令，所述处理器执行所述计算机可读指令时实现上述任意一种脊柱Cobb角测量方法的步骤。
 有益效果
本发明实施例与现有技术相比存在的有益效果是：本发明实施例使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理，得到分割结果图，所述分割结果图为脊柱区域与非脊柱区域的二值图像；从所述分割结果图的脊柱区域中分别识别出各个脊柱块；分别确定各个脊柱块的上下终板直线；对各个脊柱块的上下终板直线进行遍历计算，将取值最大的夹角确定为脊柱Cobb角。通过本发明实施例，实现了真正意义上的脊柱Cobb角自动测量，无需工作人员再进行额外的操作就能得到所需要的脊柱Cobb角，避免了手工或者半手工的方式进行脊柱Cobb角测量所引入的误差，具有更好的可靠性。
 附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其它的附图。
图1为脊柱Cobb角的传统测量方法的示意图；
图2为本发明实施例中一种脊柱Cobb角测量方法的一个实施例流程图；
图3为制作真值图的示意图；
图4为本实施例中的脊柱Cobb角测量方法的完整过程的示意图；
图5为进行脊柱Cobb角测量的一个具体实例的示意图；
图6为本发明实施例中一种脊柱Cobb角测量装置的一个实施例结构图；
图7为本发明实施例中一种终端设备的示意框图。
 本发明的实施方式
为使得本发明的发明目的、特征、优点能够更加的明显和易懂，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，下面所描述的实施例仅仅是本发明一部分实施例，而非全部的实施例。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它实施例，都属于本发明保护的范围。
请参阅图2，本发明实施例中一种脊柱Cobb角测量方法的一个实施例可以包括：
步骤S201、使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理，得到分割结果图。
优选地，所述深度学习网络模型可以为U-net分割网络。
使用所述深度学习网络模型对脊柱X光图像进行自动分割，数据是关键，针对本实施例中的具体问题，需要将脊柱从X光图像中提取出来，而X光图像是灰度图像，所以任务可以简化为二分类问题，即将图像分为脊柱区域与非脊柱区域。
在使用所述深度学习网络模型之前，首先需要对样本图像进行数据标注，制作所述样本图像的真值图（Ground truth），并使用所述样本图像以及所述样本图像的真值图对所述深度学习网络模型进行训练，得到训练好的深度学习网络模型。也即通过制作良好的真值图来监督分割网络学习，真值图是在深度学习分割网络中提供一个“金标准”，即网络学习特征的时候是参照真值图来进行参数学习的，所以，真值图制作的好坏直接影响分割的结果。由于脊柱X光图像存在许多冗杂信息，这里需要有经验的工作人员进行数据集制作。
如图3所示，即为制作真值图的示意图，工作人员将脊柱区域标注出来，标注出来的脊柱区域像素值改为0，背景区域即非脊柱区域改为1。进行训练时输入的数据像素点只有0和1，将数据进行如上处理，完成训练数据的制作过程。需要注意的是，在制作脊柱数据集时，通过工作人员的解剖经验将脊柱的上下终板提取出来，而不是简单地将整个脊柱块区域提取出来，采用这样的真值图制作方法能够提高分割的精确度以便后期的Cobb角测量能够精确地计算出角度大小。
相对来说，脊柱X光图像的原始图像数据偏少，而深度学习需要大量的数据进行训练，因此完成数据的标注工作后需要对数据集进行数据扩充，通过旋转，偏移等方式将原始图像扩充到一个理想的状态。本实施例中使用的是Keras中自带的数据增强代码，其中，Keras是基于谷歌开发的tensorflow提出的一种框架，使用十分方便，直接使用其模块即可完成网络的搭建工作。
完成上述工作就可以进行网络训练。在本实施例的一种具体实现中，整个训练过程基于Keras框架，使用的编程语言为Pyhton，版本为3.6，编译器为Pycharm。实验平台基于 Windows 7系统，使用的显卡为 NVIDIA GeForce GTX 1080 Ti，经过一段时间的训练就可以得到分割结果图，所述分割结果图为脊柱区域与非脊柱区域的二值图像。
步骤S202、从所述分割结果图的脊柱区域中分别识别出各个脊柱块。
由于脊柱块的形状类似于矩形，因此，本实施例中优选使用最小外包络矩形的方法从所述分割结果图的脊柱区域中分别识别出各个脊柱块，将所示分割结果图中的各个脊柱块均包络起来，这样就可以对每一个脊柱块单独进行操作。
步骤S203、分别确定各个脊柱块的上下终板直线。
具体地，首先对各个脊柱块的像素点进行像素值检测，并将发生第一类像素突变区域的像素点确定为上终板点，将发生第二类像素突变区域的像素点确定为下终板点。
所述第一类像素突变区域为从预设的第一像素值变为预设的第二像素值的区域，所述第二类像素突变区域为从所述第二像素值变为所述第一像素值的区域，所述第一像素值为非脊柱区域的像素值，所述第二像素值为脊柱区域的像素值。
然后，采用最小二乘法分别对各个脊柱块的上终板点进行函数拟合，得到各个脊柱块的上终板直线。类似地，采用最小二乘法分别对各个脊柱块的下终板点进行函数拟合，得到各个脊柱块的下终板直线。
如图3（b）中所示，由于上下终板的边界像素会发生突变，根据局部逼近的思想，不断地逼近某一脊柱块的上下终板，在分割结果图中，从上到下来进行检测，当像素值从所述第一像素值变为所述第二像素值时代表检测到了上终板，当像素值从所述第二像素值变为所述第一像素值时代表检测到了下终板，这里需要进行二重遍历操作，例如，当程序检测到上终板时标记为红色点，当检测到下终板时标记为蓝色点，遍历完一个脊柱块之后使用最小二乘法对红色点群进行函数拟合可以得到脊柱块的上终板直线，同理使用最小二乘法对蓝色点群进行函数拟合可以得到脊柱块的下终板直线，然后再以同样的方式对每一个脊柱块进行操作，得到每一个脊柱块的上下终板直线。将需要处理的分割结果图中所有的脊柱块检测完以后，可以将拟合的上终板直线与下终板直线分别进行存储，以便后续使用。
步骤S204、对各个脊柱块的上下终板直线进行遍历计算，将取值最大的夹角确定为脊柱Cobb角。
在得到了各个脊柱块的上下终板直线以后，只需要对所有脊柱块进行遍历就可以比较出脊柱Cobb角，即对存储的直线进行穷举，从中找出两条直线间的最大的角度，即为脊柱Cobb角。
进一步地，本实施例还可以同时给出脊柱Cobb角所在的具体位置。在医学中每一个脊柱块均有相应的命名。从下往上计数，分别为5个胸椎（L5-L1），12个腰椎（T12-T1）。此处的脊柱Cobb角测量不涉及颈椎，所以此处不包括在内。在计算并比较角度的时候将每一个脊柱块的上下终板给予相应的命名即可实现椎体计数功能。
图4所示为本实施例中的脊柱Cobb角测量方法的完整过程的示意图，图5所示即为进行脊柱Cobb角测量的一个具体实例，在该实例中，测量的角度结果为40度，上端椎为T11，下端椎为L3。
在本实施例的一种具体实现中，整个自动计数Cobb角度代码基于计算机视觉库Opencv 3.3.1版本，编写语言为C++语言，编译平台为Visual Stdio 2013, 实验平台基于 Windows 7系统，使用的显卡为 NVIDIA GeForce GTX 1080 Ti。
综上所述，本发明实施例使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理，得到分割结果图，所述分割结果图为脊柱区域与非脊柱区域的二值图像；从所述分割结果图的脊柱区域中分别识别出各个脊柱块；分别确定各个脊柱块的上下终板直线；对各个脊柱块的上下终板直线进行遍历计算，将取值最大的夹角确定为脊柱Cobb角。通过本发明实施例，实现了真正意义上的脊柱Cobb角自动测量，无需工作人员再进行额外的操作就能得到所需要的脊柱Cobb角，避免了手工或者半手工的方式进行脊柱Cobb角测量所引入的误差，具有更好的可靠性。
应理解，上述实施例中各步骤的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本发明实施例的实施过程构成任何限定。
对应于上文实施例所述的一种脊柱Cobb角测量方法，图6示出了本发明实施例提供的一种脊柱Cobb角测量装置的一个实施例结构图。
本实施例中，一种脊柱Cobb角测量装置可以包括：
图像分割模块601，用于使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理，得到分割结果图，所述分割结果图为脊柱区域与非脊柱区域的二值图像；
脊柱块识别模块602，用于从所述分割结果图的脊柱区域中分别识别出各个脊柱块；
上下终板确定模块603，用于分别确定各个脊柱块的上下终板直线；
脊柱Cobb角确定模块604，用于对各个脊柱块的上下终板直线进行遍历计算，将取值最大的夹角确定为脊柱Cobb角。
优选地，所述脊柱块识别模块具体用于使用最小外包络矩形的方法从所述分割结果图的脊柱区域中分别识别出各个脊柱块。
进一步地，所述上下终板确定模块可以包括：
像素点检测单元，用于对各个脊柱块的像素点进行像素值检测，并将发生第一类像素突变区域的像素点确定为上终板点，将发生第二类像素突变区域的像素点确定为下终板点，所述第一类像素突变区域为从预设的第一像素值变为预设的第二像素值的区域，所述第二类像素突变区域为从所述第二像素值变为所述第一像素值的区域；
上终板直线拟合单元，用于采用最小二乘法分别对各个脊柱块的上终板点进行函数拟合，得到各个脊柱块的上终板直线；
下终板直线拟合单元，用于采用最小二乘法分别对各个脊柱块的下终板点进行函数拟合，得到各个脊柱块的下终板直线。
进一步地，所述脊柱Cobb角测量装置还可以包括：
数据标注模块，用于对样本图像进行数据标注，制作所述样本图像的真值图；
模型训练模块，用于使用所述样本图像以及所述样本图像的真值图对所述深度学习网络模型进行训练，得到训练好的深度学习网络模型。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的装置，模块和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在上述实施例中，对各个实施例的描述都各有侧重，某个实施例中没有详述或记载的部分，可以参见其它实施例的相关描述。
图7示出了本发明实施例提供的一种终端设备的示意框图，为了便于说明，仅示出了与本发明实施例相关的部分。
如图7所示，该实施例的脊柱Cobb角测量终端设备7包括：处理器70、存储器71以及存储在所述存储器71中并可在所述处理器70上运行的计算机程序72。所述处理器70执行所述计算机程序72时实现上述各个脊柱Cobb角测量方法实施例中的步骤，例如图2所示的步骤S201至步骤S204。或者，所述处理器70执行所述计算机程序72时实现上述各装置实施例中各模块/单元的功能，例如图6所示模块601至模块604的功能。
示例性的，所述计算机程序72可以被分割成一个或多个模块/单元，所述一个或者多个模块/单元被存储在所述存储器71中，并由所述处理器70执行，以完成本发明。所述一个或多个模块/单元可以是能够完成特定功能的一系列计算机程序指令段，该指令段用于描述所述计算机程序72在所述脊柱Cobb角测量终端设备7中的执行过程。
所述脊柱Cobb角测量终端设备7可以是桌上型计算机、笔记本、掌上电脑及云端服务器等计算设备。本领域技术人员可以理解，图7仅仅是脊柱Cobb角测量终端设备7的示例，并不构成对脊柱Cobb角测量终端设备7的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件，例如所述脊柱Cobb角测量终端设备7还可以包括输入输出设备、网络接入设备、总线等。
所述处理器70可以是中央处理单元(Central Processing Unit，CPU)，还可以是其它通用处理器、数字信号处理器 (Digital Signal Processor，DSP)、专用集成电路 (Application Specific Integrated Circuit，ASIC)、现场可编程门阵列(Field-Programmable Gate Array，FPGA) 或者其它可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
所述存储器71可以是所述脊柱Cobb角测量终端设备7的内部存储单元，例如脊柱Cobb角测量终端设备7的硬盘或内存。所述存储器71也可以是所述脊柱Cobb角测量终端设备7的外部存储设备，例如所述脊柱Cobb角测量终端设备7上配备的插接式硬盘，智能存储卡（Smart Media Card, SMC），安全数字（Secure Digital, SD）卡，闪存卡（Flash Card）等。进一步地，所述存储器71还可以既包括所述脊柱Cobb角测量终端设备7的内部存储单元也包括外部存储设备。所述存储器71用于存储所述计算机程序以及所述脊柱Cobb角测量终端设备7所需的其它程序和数据。所述存储器71还可以用于暂时地存储已经输出或者将要输出的数据。
所属领域的技术人员可以清楚地了解到，为了描述的方便和简洁，仅以上述各功能单元、模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能单元、模块完成，即将所述装置的内部结构划分成不同的功能单元或模块，以完成以上描述的全部或者部分功能。实施例中的各功能单元、模块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。另外，各功能单元、模块的具体名称也只是为了便于相互区分，并不用于限制本申请的保护范围。上述系统中单元、模块的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在上述实施例中，对各个实施例的描述都各有侧重，某个实施例中没有详述或记载的部分，可以参见其它实施例的相关描述。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本发明的范围。
在本发明所提供的实施例中，应该理解到，所揭露的装置/终端设备和方法，可以通过其它的方式实现。例如，以上所描述的装置/终端设备实施例仅仅是示意性的，例如，所述模块或单元的划分，仅仅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讯连接可以是通过一些接口，装置或单元的间接耦合或通讯连接，可以是电性，机械或其它的形式。
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网络单元上。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。
所述集成的模块/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，本发明实现上述实施例方法中的全部或部分流程，也可以通过计算机程序来指令相关的硬件来完成，所述的计算机程序可存储于一计算机可读存储介质中，该计算机程序在被处理器执行时，可实现上述各个方法实施例的步骤。其中，所述计算机程序包括计算机程序代码，所述计算机程序代码可以为源代码形式、对象代码形式、可执行文件或某些中间形式等。所述计算机可读介质可以包括：能够携带所述计算机程序代码的任何实体或装置、记录介质、U盘、移动硬盘、磁碟、光盘、计算机存储器、只读存储器（ROM，Read-Only Memory）、随机存取存储器（RAM，Random Access Memory）、电载波信号、电信信号以及软件分发介质等。需要说明的是，所述计算机可读介质包含的内容可以根据司法管辖区内立法和专利实践的要求进行适当的增减，例如在某些司法管辖区，根据立法和专利实践，计算机可读介质不包括电载波信号和电信信号。
以上所述实施例仅用以说明本发明的技术方案，而非对其限制；尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的精神和范围，均应包含在本发明的保护范围之内。

Claims (10)

  1. 一种脊柱Cobb角测量方法，其特征在于，包括：
    使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理，得到分割结果图，所述分割结果图为脊柱区域与非脊柱区域的二值图像；
    从所述分割结果图的脊柱区域中分别识别出各个脊柱块；
    分别确定各个脊柱块的上下终板直线；
    对各个脊柱块的上下终板直线进行遍历计算，将取值最大的夹角确定为脊柱Cobb角。
  2. 根据权利要求1所述的脊柱Cobb角测量方法，其特征在于，所述从所述分割结果图的脊柱区域中分别识别出各个脊柱块包括：
    使用最小外包络矩形的方法从所述分割结果图的脊柱区域中分别识别出各个脊柱块。
  3. 根据权利要求1所述的脊柱Cobb角测量方法，其特征在于，所述分别确定各个脊柱块的上下终板直线包括：
    对各个脊柱块的像素点进行像素值检测，并将发生第一类像素突变区域的像素点确定为上终板点，将发生第二类像素突变区域的像素点确定为下终板点，所述第一类像素突变区域为从预设的第一像素值变为预设的第二像素值的区域，所述第二类像素突变区域为从所述第二像素值变为所述第一像素值的区域；
    采用最小二乘法分别对各个脊柱块的上终板点进行函数拟合，得到各个脊柱块的上终板直线；
    采用最小二乘法分别对各个脊柱块的下终板点进行函数拟合，得到各个脊柱块的下终板直线。
  4. 根据权利要求1所述的脊柱Cobb角测量方法，其特征在于，在使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理之前，还包括：
    对样本图像进行数据标注，制作所述样本图像的真值图；
    使用所述样本图像以及所述样本图像的真值图对所述深度学习网络模型进行训练，得到训练好的深度学习网络模型。
  5. 根据权利要求1至4中任一项所述的脊柱Cobb角测量方法，其特征在于，所述深度学习网络模型为U-net分割网络。
  6. 一种脊柱Cobb角测量装置，其特征在于，包括：
    图像分割模块，用于使用预先训练好的深度学习网络模型对脊柱X光图像进行分割处理，得到分割结果图，所述分割结果图为脊柱区域与非脊柱区域的二值图像；
    脊柱块识别模块，用于从所述分割结果图的脊柱区域中分别识别出各个脊柱块；
    上下终板确定模块，用于分别确定各个脊柱块的上下终板直线；
    脊柱Cobb角确定模块，用于对各个脊柱块的上下终板直线进行遍历计算，将取值最大的夹角确定为脊柱Cobb角。
  7. 根据权利要求6所述的脊柱Cobb角测量装置，其特征在于，所述上下终板确定模块包括：
    像素点检测单元，用于对各个脊柱块的像素点进行像素值检测，并将发生第一类像素突变区域的像素点确定为上终板点，将发生第二类像素突变区域的像素点确定为下终板点，所述第一类像素突变区域为从预设的第一像素值变为预设的第二像素值的区域，所述第二类像素突变区域为从所述第二像素值变为所述第一像素值的区域；
    上终板直线拟合单元，用于采用最小二乘法分别对各个脊柱块的上终板点进行函数拟合，得到各个脊柱块的上终板直线；
    下终板直线拟合单元，用于采用最小二乘法分别对各个脊柱块的下终板点进行函数拟合，得到各个脊柱块的下终板直线。
  8. 根据权利要求6所述的脊柱Cobb角测量装置，其特征在于，所述脊柱Cobb角测量装置还包括：
    数据标注模块，用于对样本图像进行数据标注，制作所述样本图像的真值图；
    模型训练模块，用于使用所述样本图像以及所述样本图像的真值图对所述深度学习网络模型进行训练，得到训练好的深度学习网络模型。
  9. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机可读指令，其特征在于，所述计算机可读指令被处理器执行时实现如权利要求1至5中任一项所述的脊柱Cobb角测量方法的步骤。
  10. 一种终端设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令，其特征在于，所述处理器执行所述计算机可读指令时实现如权利要求1至5中任一项所述的脊柱Cobb角测量方法的步骤。
PCT/CN2019/130900 2019-04-03 2019-12-31 一种脊柱Cobb角测量方法、装置、可读存储介质及终端设备 WO2020199694A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910267097.5 2019-04-03
CN201910267097.5A CN110175502A (zh) 2019-04-03 2019-04-03 一种脊柱Cobb角测量方法、装置、可读存储介质及终端设备

Publications (1)

Publication Number Publication Date
WO2020199694A1 true WO2020199694A1 (zh) 2020-10-08

Family

ID=67689391

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/130900 WO2020199694A1 (zh) 2019-04-03 2019-12-31 一种脊柱Cobb角测量方法、装置、可读存储介质及终端设备

Country Status (2)

Country Link
CN (1) CN110175502A (zh)
WO (1) WO2020199694A1 (zh)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112529863A (zh) * 2020-12-04 2021-03-19 推想医疗科技股份有限公司 测量骨密度的方法及装置
CN112535489A (zh) * 2020-12-25 2021-03-23 四川大学华西医院 一种利用波形图测量脊柱侧弯Cobb角的方法及其系统
CN113781453A (zh) * 2021-09-15 2021-12-10 南京大学 一种基于x线片的脊柱侧凸进展预测及方法和装置
CN114983459A (zh) * 2022-06-21 2022-09-02 济南大学 一种三维脊柱模型Cobb角的交互式测量系统及方法
CN115153509A (zh) * 2022-08-05 2022-10-11 芙索特(上海)医疗科技有限公司 一种脊柱空间曲线生成方法、系统及脊柱测量仪
CN115187606A (zh) * 2022-09-14 2022-10-14 中国医学科学院北京协和医院 一种青少年特发性脊柱侧凸pumc分型方法
EP4152247A1 (en) * 2021-09-16 2023-03-22 Siemens Healthcare GmbH Method and device for automatically determining spine deformation from an image

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110175502A (zh) * 2019-04-03 2019-08-27 中国科学院深圳先进技术研究院 一种脊柱Cobb角测量方法、装置、可读存储介质及终端设备
KR102383857B1 (ko) * 2020-05-14 2022-04-06 가천대학교 산학협력단 콥 각도 측정 방법 및 시스템, 콥 각도 측정 프로그램
CN112274164B (zh) * 2020-07-29 2023-02-21 深圳市智影医疗科技有限公司 脊柱侧弯预测方法、装置、电子设备和存储介质
CN112258516B (zh) * 2020-09-04 2023-04-07 温州医科大学附属第二医院、温州医科大学附属育英儿童医院 一种脊柱侧弯图像检测模型的生成方法
CN112381869B (zh) * 2020-10-09 2022-12-13 温州医科大学附属第二医院、温州医科大学附属育英儿童医院 脊柱全长x线片测算脊柱侧弯测量和计算装置
CN112381757A (zh) * 2020-10-09 2021-02-19 温州医科大学附属第二医院、温州医科大学附属育英儿童医院 一种基于人工智能-图像识别的脊柱全长X线片测算脊柱侧弯Cobb角的系统及方法
CN112529860A (zh) * 2020-12-04 2021-03-19 中国科学院深圳先进技术研究院 脊柱图像处理方法、装置、电子设备及存储介质
CN112472111B (zh) * 2020-12-14 2023-08-29 上海交通大学医学院附属新华医院 一种脊柱侧凸早期自检与康复云平台
CN112686854B (zh) * 2020-12-25 2023-01-24 四川大学华西医院 一种自动测量脊柱侧弯Cobb角的方法及其系统
CN112734757B (zh) * 2021-03-29 2021-06-25 成都成电金盘健康数据技术有限公司 一种脊柱X光图像cobb角测量方法
CN112802019B (zh) * 2021-04-01 2021-06-29 成都成电金盘健康数据技术有限公司 一种基于脊柱AIS影像的lenke分型方法
CN114078120B (zh) * 2021-11-22 2022-05-20 北京欧应信息技术有限公司 用于检测脊柱侧弯的方法、设备和介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20160092666A (ko) * 2015-01-28 2016-08-05 울산대학교 산학협력단 콥 각도 측정장치 및 그 방법
US20170340268A1 (en) * 2016-05-30 2017-11-30 Prismatic Sensors Ab X-ray imaging for enabling assessment of scoliosis
CN108573502A (zh) * 2018-03-06 2018-09-25 安徽大学 一种自动测量Cobb角的方法
CN109493334A (zh) * 2018-11-12 2019-03-19 深圳码隆科技有限公司 测量脊柱弯曲的方法及装置
CN110175502A (zh) * 2019-04-03 2019-08-27 中国科学院深圳先进技术研究院 一种脊柱Cobb角测量方法、装置、可读存储介质及终端设备

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20160092666A (ko) * 2015-01-28 2016-08-05 울산대학교 산학협력단 콥 각도 측정장치 및 그 방법
US20170340268A1 (en) * 2016-05-30 2017-11-30 Prismatic Sensors Ab X-ray imaging for enabling assessment of scoliosis
CN108573502A (zh) * 2018-03-06 2018-09-25 安徽大学 一种自动测量Cobb角的方法
CN109493334A (zh) * 2018-11-12 2019-03-19 深圳码隆科技有限公司 测量脊柱弯曲的方法及装置
CN110175502A (zh) * 2019-04-03 2019-08-27 中国科学院深圳先进技术研究院 一种脊柱Cobb角测量方法、装置、可读存储介质及终端设备

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112529863A (zh) * 2020-12-04 2021-03-19 推想医疗科技股份有限公司 测量骨密度的方法及装置
CN112529863B (zh) * 2020-12-04 2024-01-23 推想医疗科技股份有限公司 测量骨密度的方法及装置
CN112535489A (zh) * 2020-12-25 2021-03-23 四川大学华西医院 一种利用波形图测量脊柱侧弯Cobb角的方法及其系统
CN112535489B (zh) * 2020-12-25 2023-12-05 四川大学华西医院 一种利用波形图测量脊柱侧弯Cobb角的方法及其系统
CN113781453A (zh) * 2021-09-15 2021-12-10 南京大学 一种基于x线片的脊柱侧凸进展预测及方法和装置
CN113781453B (zh) * 2021-09-15 2024-04-02 南京大学 一种基于x线片的脊柱侧凸进展预测及方法和装置
EP4152247A1 (en) * 2021-09-16 2023-03-22 Siemens Healthcare GmbH Method and device for automatically determining spine deformation from an image
CN114983459A (zh) * 2022-06-21 2022-09-02 济南大学 一种三维脊柱模型Cobb角的交互式测量系统及方法
CN115153509A (zh) * 2022-08-05 2022-10-11 芙索特(上海)医疗科技有限公司 一种脊柱空间曲线生成方法、系统及脊柱测量仪
CN115153509B (zh) * 2022-08-05 2023-05-05 芙索特(上海)医疗科技有限公司 一种脊柱空间曲线生成方法、系统及脊柱测量仪
CN115187606A (zh) * 2022-09-14 2022-10-14 中国医学科学院北京协和医院 一种青少年特发性脊柱侧凸pumc分型方法
CN115187606B (zh) * 2022-09-14 2022-12-13 中国医学科学院北京协和医院 一种青少年特发性脊柱侧凸pumc分型方法

Also Published As

Publication number Publication date
CN110175502A (zh) 2019-08-27

Similar Documents

Publication Publication Date Title
WO2020199694A1 (zh) 一种脊柱Cobb角测量方法、装置、可读存储介质及终端设备
WO2020207190A1 (zh) 一种三维信息确定方法、三维信息确定装置及终端设备
WO2021056705A1 (zh) 基于语义分割网络的人体外表损伤检测方法及相关设备
CN110111248B (zh) 一种基于特征点的图像拼接方法、虚拟现实系统、照相机
CN110781885A (zh) 基于图像处理的文本检测方法、装置、介质及电子设备
JP7489732B2 (ja) 深層学習に基づく脊椎ｍｒｉ映像キーポイント検出方法
WO2020151340A1 (zh) 一种目标细胞标记方法、装置、存储介质及终端设备
CN111488874A (zh) 一种指针式仪表倾斜校正方法及系统
CN112529860A (zh) 脊柱图像处理方法、装置、电子设备及存储介质
US20190138840A1 (en) Automatic ruler detection
CN110930414A (zh) 医学影像的肺部区域阴影标记方法、装置、服务器及存储介质
CN114121232A (zh) 分类模型训练方法、装置、电子设备及存储介质
CN113706562A (zh) 图像分割方法、装置、系统及细胞分割方法
CN111126268A (zh) 关键点检测模型训练方法、装置、电子设备及存储介质
CN108992033B (zh) 一种视觉测试的评分装置、设备和存储介质
WO2024087910A1 (zh) 正畸治疗监测方法、装置、设备及存储介质
WO2019109410A1 (zh) 用于分割 mri 图像中异常信号区的全卷积网络模型训练方法
CN110782390A (zh) 图像的矫正处理方法及装置、电子设备
CN107742316B (zh) 图像拼接点获取方法及获取装置
CN113255756B (zh) 图像融合方法、装置、电子设备及存储介质
CN117745989B (zh) 基于椎管构造的神经根阻滞靶点注入路径规划方法及系统
TWI765250B (zh) 深度學習演算法的選擇方法及深度學習演算法的選擇裝置
CN110764642B (zh) 一种视觉投影校准的方法和装置
CN113128486B (zh) 手写数学公式样本库的构建方法、装置及终端设备
CN112950582B (zh) 一种基于深度学习的3d肺部病灶分割方法、装置

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: 19923596

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: 19923596

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 30.03.2022)

122 Ep: pct application non-entry in european phase

Ref document number: 19923596

Country of ref document: EP

Kind code of ref document: A1