WO2020024127A1 - 骨龄评估与身高预测模型、其系统及其预测方法 - Google Patents

骨龄评估与身高预测模型、其系统及其预测方法 Download PDF

Info

Publication number
WO2020024127A1
WO2020024127A1 PCT/CN2018/097915 CN2018097915W WO2020024127A1 WO 2020024127 A1 WO2020024127 A1 WO 2020024127A1 CN 2018097915 W CN2018097915 W CN 2018097915W WO 2020024127 A1 WO2020024127 A1 WO 2020024127A1
Authority
WO
WIPO (PCT)
Prior art keywords
bone
image data
ray image
height
bone age
Prior art date
Application number
PCT/CN2018/097915
Other languages
English (en)
French (fr)
Inventor
蔡辅仁
黄宗祺
廖英凯
游家鑫
Original Assignee
中国医药大学附设医院
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 中国医药大学附设医院 filed Critical 中国医药大学附设医院
Priority to PCT/CN2018/097915 priority Critical patent/WO2020024127A1/zh
Priority to JP2020528374A priority patent/JP6999812B2/ja
Priority to US17/259,685 priority patent/US20210142477A1/en
Publication of WO2020024127A1 publication Critical patent/WO2020024127A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
    • A61B6/50—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment specially adapted for specific body parts; specially adapted for specific clinical applications
    • A61B6/505—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment specially adapted for specific body parts; specially adapted for specific clinical applications for diagnosis of bone
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
    • A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
    • A61B6/5211—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data
    • A61B6/5217—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data extracting a diagnostic or physiological parameter from medical diagnostic data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G06T7/0014—Biomedical image inspection using an image reference approach
    • 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
    • G06V10/443—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 by matching or filtering
    • G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
    • G06V10/451—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters with interaction between the filter responses, e.g. cortical complex cells
    • G06V10/454—Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H70/00—ICT specially adapted for the handling or processing of medical references
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10116—X-ray image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30008—Bone
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/03—Recognition of patterns in medical or anatomical images
    • G06V2201/033—Recognition of patterns in medical or anatomical images of skeletal patterns

Definitions

  • the invention relates to a medical information analysis model, system and method, in particular to a bone age assessment and height prediction model, a bone age assessment and height prediction system, and a bone age assessment and height prediction method.
  • Bone age is one of the important indicators of the physiological age of the human body. It infers the physiological age of the human body through the rules of bone growth, development, maturity, and aging. Bone age assessment (BAA) is a routine examination commonly used by pediatricians to interpret children's growth and development. It analyzes the different types of bones at different growth stages, and refers to the continuity and staged developmental status of human bones. Then accurately assess the growth and development level and maturity of the individual, and further evaluate and predict the growth and development potential and sexual maturity of the test children.
  • BAA Bone age assessment
  • the known bone age assessment method uses low-dose X-ray photography to obtain X-ray images of the phalanx, metacarpal, and carpal bones of the left or right hand of the subject, and the aforementioned X-ray images are obtained by Greulich and Pyle (GP) method and Tanner -Whitehouse (TW) method to perform bone age assessment by means of map comparison.
  • the GP method is to perform a one-to-one comparison between the original X-ray image of the subject's hand bone and the X-ray image of the hand bone in the database according to different age ranges.
  • the TW method is to take the left palm.
  • the purpose of the present invention is to provide a bone age evaluation and height prediction model, a system and a prediction method thereof, which can effectively improve the accuracy and sensitivity of the bone age evaluation and height prediction, and can shorten the judgment time of the bone age evaluation and height prediction.
  • One aspect of the present invention is to provide a bone age evaluation and height prediction model, which is established by the following steps: obtaining a reference database, performing an image pre-processing step, performing a feature selection step, and performing a training step.
  • the reference database includes a bone age height map data set and a plurality of reference hand bone X-ray image data, wherein each reference hand bone X-ray image data includes physiological age information and gender information.
  • the image pre-processing step uses an image data editing module to adjust the image size and black and white contrast of each reference hand X-ray image data to obtain a plurality of standardized hand bone X-ray image data.
  • the feature selection step is to use the feature selection module to analyze the standardized hand bone X-ray image data to obtain at least one image feature value.
  • the training step is to converge the image feature values through a convolutional neural network learning classifier to obtain a bone age assessment and height prediction model.
  • the bone age assessment and height prediction model is used to judge the subject's hand bone development status, The subject's bone age and predicted adult height.
  • the convolutional neural network learning classifier may be an Inception-ResNet-v2 convolutional neural network.
  • the image format of the reference hand bone X-ray image data may be an image format of Digital Medical Imaging Standards and Communications (DICOM).
  • DICOM Digital Medical Imaging Standards and Communications
  • the image pre-processing step may further perform image chroma expansion processing on each reference hand bone X-ray image data.
  • the bone age height map data set may include a male bone age height map data sub-set and a female bone age height map data sub-set.
  • the reference hand bone X-ray image data may be reference hand bone X-ray image data of non-habitual hands.
  • Another aspect of the present invention is to provide a bone age assessment and height prediction method, which includes the following steps.
  • Provide the subject's target hand bone X-ray image data wherein the aforementioned target hand bone X-ray image data includes physiological age information and gender information.
  • the target hand bone X-ray image data is pre-processed.
  • the aforementioned image data editing module is used to adjust the image size and black and white contrast of the target hand bone X-ray image data to obtain standardized target hand bone X-ray image data.
  • the at least one image feature value is obtained after analyzing the standardized target hand bone X-ray image data by using the aforementioned feature selection module.
  • Using the aforementioned bone age assessment and height prediction model to analyze image feature values to determine the hand bone development status of the subject, the subject's bone age, and predict the subject's adult height.
  • the image format of the target hand bone X-ray image data may be an image format of a digital medical image storage standard agreement.
  • the image data editing module may further perform image chroma expansion processing on the target hand bone X-ray image data.
  • the aforementioned target hand bone X-ray image data may be target hand bone X-ray image data of non-habitual hands.
  • Another aspect of the present invention is to provide a bone age assessment and height prediction system, which includes an image capture unit and a non-transitory machine-readable medium.
  • the image capturing unit is used to obtain target hand bone X-ray image data of the subject, wherein the aforementioned target hand bone X-ray image data includes physiological age information and gender information.
  • the non-transitory machine-readable medium is signally connected to the aforementioned image acquisition unit, wherein the non-transitory machine-readable medium is used to store a program, and when the foregoing program is executed by the processing unit, it is used to evaluate the development state of the hand bone of the subject ,
  • the bone age of the subject and the predicted adult height of the subject, and the aforementioned program includes a reference database acquisition module, a first image data editing module, a feature selection module, a training module, a second image data editing module, and a target feature selection module And comparison modules.
  • the reference database obtaining module is used to obtain a reference database, and the aforementioned reference database includes a bone age height map data set and a plurality of reference hand bone X-ray image data, wherein each reference hand bone X-ray image data includes physiological age information and gender information.
  • the first image data editing module adjusts the image size and the black and white contrast of each reference hand X-ray image data to obtain a plurality of standardized hand X-ray image data.
  • the feature selection module is used to analyze at least one reference image feature value after analyzing the standardized hand bone X-ray image data.
  • the training module is used to train the aforementioned reference image feature values through a convolutional neural network learning classifier to achieve convergence, so as to obtain a bone age assessment and height prediction model.
  • the second image data editing module adjusts the image size and the black and white contrast of the target hand bone X-ray image data to obtain standardized target hand bone X-ray image data.
  • the target feature selection module is used to analyze at least one target image feature value after analyzing the normalized target hand bone X-ray image data.
  • the comparison module is configured to analyze the at least one target image feature value by using the aforementioned bone age assessment and height prediction model to obtain target image feature value weight data, and compare the at least one target image feature value weight data with a reference database. The comparison is performed to output the judgment result of the hand bone development state of the subject, the judgment result of the subject's bone age, and the prediction result of the subject's adult height.
  • the convolutional neural network learning classifier may be an Inception-ResNet-v2 convolutional neural network.
  • the image format of the target hand bone X-ray image data may be an image format of a digital medical image storage standard agreement
  • the aforementioned reference hand bone X-ray image data and image format may be An image format for digital medical image storage standard agreement.
  • the first image data editing module may further perform image chroma expansion processing on each reference hand bone X-ray image data, and the second image data editing module may further perform target hand bone X-rays.
  • the image data is processed for image chrominance expansion.
  • the bone age height map data set may include a male bone age height map data sub-set and a female bone age height map data sub-set.
  • each reference hand bone X-ray image data is reference hand bone X-ray image data of unaccustomed hands
  • the target hand bone X-ray image data is target hand bones of non-habitual hands X-ray image data.
  • the bone age evaluation and height prediction system may further include a warning module for issuing an active warning notification after the standardized target hand bone X-ray image data is analyzed through the bone age evaluation and height prediction model.
  • the bone age assessment and height prediction model, bone age assessment and height prediction system, and bone age assessment and height prediction method of the present invention perform pre-processing of image standardization by referring to hand X-ray image data and target hand X-ray image data, and After using the feature selection module to analyze and obtain at least one image feature value, the convolutional neural network is used to train the image feature value to analyze and judge the opponent's bone development status, bone age, and adult height, which can not only effectively shorten the bone age assessment and height prediction The required time can also avoid the result error caused by different analysts' feature selection and comparison methods in the known bone age assessment methods.
  • the bone age assessment and height prediction model including a convolutional neural network learning classifier can effectively improve the accuracy and sensitivity of the bone age assessment and height prediction, which not only enables the bone age assessment and height prediction model, bone age assessment, and height of the present invention.
  • the prediction system and bone age assessment and height prediction methods are more efficient in determining bone age and height prediction, and can accurately assess the growth and development level and maturity of different subjects, and predict the future growth and development potential of subjects.
  • FIG. 1 is a flowchart illustrating steps for establishing a bone age assessment and height prediction model according to an embodiment of the present invention
  • FIG. 2 is a flowchart illustrating steps of a bone age assessment and height prediction method according to another embodiment of the present invention.
  • FIG. 3 is a schematic diagram illustrating a bone age assessment and height prediction system according to still another embodiment of the present invention.
  • FIG. 4 is a flowchart showing a part of the steps for establishing a bone age assessment and height prediction model according to the present invention
  • FIG. 5 is a schematic diagram showing a convolutional neural network learning classifier of a bone age assessment and height prediction model of the present invention.
  • FIG. 6 is a schematic diagram showing application results of the bone age assessment and height prediction system of the present invention.
  • FIG. 1 is a flowchart illustrating steps for establishing a bone age assessment and height prediction model 100 according to an embodiment of the present invention.
  • the bone age assessment and height prediction model 100 is used to judge the subject's hand bone development status, the subject's bone age, and predict the subject's adult height.
  • the steps of establishing the bone age assessment and height prediction model 100 include step 110 and step. 120, step 130, and step 140.
  • Step 110 is to obtain a reference database, where the reference database includes a bone age height map data set and a plurality of reference hand bone X-ray image data, wherein each reference hand bone X-ray image data includes physiological age information and gender information.
  • the aforementioned reference hand bone X-ray image data may be reference hand bone X-ray image data of non-accustomed hands, so as to avoid the bone age assessment and height prediction model 100 of the present invention caused by the frequency of use of hand or usage habits. The variation of the bone shape affects the accuracy of its judgment.
  • the image format of the aforementioned reference hand bone X-ray image data may be an image format of Digital Medical Imaging Standards and Communications (DICOM), so as to convert the physiological age of each reference hand bone X-ray image data.
  • Basic data such as information and gender information are stored in a header that refers to hand bone X-ray image data to facilitate subsequent analysis.
  • the bone age assessment and height prediction model 100 of the present invention can further refer to different genders.
  • the bone X-ray image data is subjected to feature selection steps and training steps to judge and predict gendered hand bone development status, bone age, and adult height.
  • the aforementioned bone age height atlas data set may include a male bone age height atlas data sub-set and a female bone age height atlas data sub-set to facilitate analysis of subjects of different genders.
  • Step 120 is an image pre-processing step, which uses an image data editing module to adjust the image size and image black and white contrast of each reference hand bone X-ray image data to obtain a plurality of standardized hand bone X-ray image data.
  • the image data editing module can adjust the image size of different reference hand bone X-ray image data to 256 pixels (pixel) ⁇ 256 pixels, and adjust its black and white contrast to reduce different reference hand bone X-ray images. The difference in black and white between the data and increase the sharpness of the image to facilitate subsequent analysis.
  • the image data editing module may further perform image chroma expansion processing on each reference hand bone X-ray image data.
  • the image data editing module can calculate the image gray level of each reference hand bone X-ray image data, and automatically fill in the image pixel rows and columns of each reference hand bone X-ray image data in order according to the aforementioned calculation results.
  • the color is used to convert each reference hand bone X-ray image data showing grayscale tones into color tones, thereby improving the accuracy of subsequent analysis, but the present invention is not limited to the contents disclosed in the foregoing description and the accompanying drawings.
  • Step 130 is a feature selection step, which uses the feature selection module to analyze the standardized hand bone X-ray image data to obtain at least one image feature value.
  • the bone age assessment and height prediction model 100 of the present invention can automatically analyze image information of standardized hand bone X-ray image data by using a feature selection module, and automatically extract corresponding image feature values, thereby improving the bone age of the present invention. Evaluation and prediction efficiency of the evaluation and height prediction model 100.
  • Step 140 is a training step, in which the aforementioned image feature values are trained by a convolutional neural network learning classifier to achieve convergence to obtain a bone age assessment and height prediction model 100.
  • the aforementioned convolutional neural network learning classifier may be an Inception-ResNet-v2 convolutional neural network.
  • Inception-ResNet-v2 convolutional neural network is a large-scale visual recognition (Large Scale Visual Recognition) convolutional neural network based on the ImageNet visual data database. It can effectively expand the convolution by means of residual connections.
  • the training depth of the neural network makes the Inception-ResNet-v2 convolutional neural network have a fairly high accuracy rate in image classification and recognition.
  • FIG. 2 is a flowchart illustrating steps of a bone age assessment and height prediction method 200 according to another embodiment of the present invention.
  • the bone age assessment and height prediction method 200 includes steps 210, 220, 230, 240, and 250.
  • Step 210 is to provide a bone age assessment and height prediction model, and the bone age assessment and height prediction model are established through the foregoing steps 110 to 140.
  • Step 220 is to provide target hand bone X-ray image data, where the target hand bone X-ray image data includes physiological age information and gender information.
  • the aforementioned target hand bone X-ray image data may be target hand bone X-ray image data of non-habitual hands, so as to prevent the subject from being affected by the frequency of habitual hand use or the skeletal shape variation caused by the use habit. Analysis accuracy of bone age assessment and height prediction method 200.
  • the aforementioned image format of the target hand bone X-ray image data may be an image format of a digital medical image storage standard agreement, so as to store basic data such as physiological age information and gender information of the target hand bone X-ray image data in the target hand.
  • the file of bone X-ray image data is convenient for subsequent analysis.
  • the bone age assessment and height prediction method 200 of the present invention will target target hand bones of different genders, respectively. X-ray image data were used to evaluate and analyze the gendered hand bone development status, bone age, and adult height.
  • Step 230 is to pre-process the target hand bone X-ray image data, which uses the aforementioned image data editing module to adjust the image size and image black and white contrast of the target hand bone X-ray image data to obtain standardized target hand bone X-ray image data.
  • the image data editing module adjusts the image size of the target hand bone X-ray image data to 256 pixels ⁇ 256 pixels, and adjusts the black and white contrast to increase the sharpness of the image to facilitate subsequent analysis.
  • the image data editing module may further perform image chroma expansion processing on the target hand bone X-ray image data.
  • the image data editing module can calculate the image gray level of the target hand bone X-ray image data, and automatically fill colors in the image pixel rows and columns of the target hand bone X-ray image data in order according to the aforementioned calculation results.
  • the target hand bone X-ray image data showing grayscale tones is converted to color tones, thereby improving the accuracy of subsequent analysis.
  • the present invention is not limited to the contents disclosed in the foregoing description and the accompanying drawings.
  • Step 240 is to use the feature selection module to analyze the normalized target hand bone X-ray image data to obtain at least one image feature value.
  • the bone age assessment and height prediction method 200 of the present invention can automatically analyze image information of standardized target hand bone X-ray image data by using a feature selection module, and automatically extract corresponding image feature values, thereby improving the present invention. Evaluation and prediction efficiency of the bone age evaluation and height prediction method 200.
  • Step 250 is to use the aforementioned bone age assessment and height prediction model to analyze image feature values to determine the hand bone development status of the subject, the subject's bone age, and predict the subject's adult height.
  • FIG. 3 is a schematic diagram of a bone age assessment and height prediction system 300 according to another embodiment of the present invention.
  • the bone age assessment and height prediction system 300 includes an image capture unit 400 and a non-transitory machine-readable medium 500.
  • the image capture unit 400 is used to obtain target hand bone X-ray image data, where the target hand bone X-ray image data includes physiological age information and gender information.
  • the image capturing unit 400 may be an X-ray detection apparatus, which irradiates a subject's hand with a low dose of X-rays to obtain target hand bone X-ray image data with an appropriate resolution.
  • the aforementioned target hand bone X-ray image data may be target hand bone X-ray image data of a non-habitual hand, so as to avoid the bone age assessment and height prediction system 300 of the present invention from being caused by the frequency of use of hand or use habits.
  • the variation of the bone shape affects the accuracy of its analysis.
  • the aforementioned image format of the target hand bone X-ray image data may be an image format of a digital medical image storage standard agreement, so as to store basic data such as physiological age information and gender information of the target hand bone X-ray image data in the target hand.
  • the file of bone X-ray image data is convenient for subsequent analysis.
  • the non-transitory machine-readable medium 500 is signal-connected to the image capturing unit 400.
  • the non-transitory machine-readable medium 500 is used to store a program (not shown).
  • a program (not shown)
  • Time is used to evaluate the hand bone development status of the subject, the subject's bone age, and predict the subject's adult height.
  • the aforementioned program includes a reference database acquisition module 510, a first image data editing module 520, and a feature selection module. 530, a training module 540, a second image data editing module 550, a target feature selection module 560, and a comparison module 570.
  • the reference database obtaining module 510 is used to obtain a reference database, and the aforementioned reference database includes a bone age height map data set and a plurality of reference hand bone X-ray image data, wherein each reference hand bone X-ray image data includes physiological age information and gender information.
  • each reference hand bone X-ray image data may be reference hand bone X-ray image data of an unaccustomed hand
  • the aforementioned bone age height map data set may include a male bone age height map data sub-set and a female bone age height map data sub-set. Collection to facilitate analysis of subjects of different genders.
  • the aforementioned image format of the reference hand bone X-ray image data may be an image format of a digital medical image storage standard agreement, so that basic data such as physiological age information and gender information of each reference hand bone X-ray image data are stored in the reference. In the file of the hand bone X-ray image data, it is convenient for subsequent analysis.
  • the first image data editing module 520 adjusts the image size and image black-and-white contrast of each reference hand bone X-ray image data to obtain a plurality of standardized hand bone X-ray image data.
  • the first image data editing module 520 adjusts the image size of different reference hand bone X-ray image data to 256 pixels ⁇ 256 pixels, and adjusts its black and white contrast to reduce different reference hand bone X-ray image data. The difference in black and whiteness as well as the sharpness of the image.
  • the first image data editing module 520 may further perform image chroma expansion processing on each reference hand bone X-ray image data.
  • the first image data editing module 520 may calculate an image gray level of each reference hand bone X-ray image data, and sequentially perform image pixel rows, The columns are automatically filled with colors to convert the reference hand bone X-ray image data showing grayscale tones to color tones to improve the accuracy of subsequent analysis, but the present invention is not limited to the content disclosed in the foregoing description and drawings.
  • the feature selection module 530 is configured to analyze at least one reference image feature value after analyzing the standardized hand bone X-ray image data.
  • the bone age assessment and height prediction system 300 of the present invention can use the feature selection module 530 to automatically analyze the image information of the standardized hand bone X-ray image data and automatically extract the corresponding image feature values.
  • the training module 540 is used to train the aforementioned reference image feature values through a convolutional neural network learning classifier to achieve convergence, so as to obtain a bone age assessment and height prediction model.
  • the aforementioned convolutional neural network learning classifier may be an Inception-ResNet-v2 convolutional neural network to effectively extend the training depth of the convolutional neural network, thereby improving the image classification and recognition capabilities of the training module 540.
  • the second image data editing module 550 adjusts the image size and the black and white contrast of the target hand bone X-ray image data to obtain standardized target hand bone X-ray image data.
  • the second image data editing module 550 adjusts the image size of the target hand bone X-ray image data to 256 pixels ⁇ 256 pixels, and adjusts the black and white contrast to increase the sharpness of the image, thereby obtaining the aforementioned standardized target.
  • X-ray image data of hand bone is a predefined range
  • the second image data editing module 550 may further perform image chroma expansion processing on the target hand bone X-ray image data, which is to calculate the image gray level of the target hand bone X-ray image data, and according to the foregoing calculation result,
  • the image pixel rows and columns of the target hand bone X-ray image data are automatically filled in order in order to convert the target hand bone X-ray image data showing grayscale tones into color tones, but the present invention does not use the foregoing description and drawings. Public content is limited.
  • the target feature selection module 560 is used for analyzing at least one target image feature value after analyzing the normalized target hand bone X-ray image data.
  • the target feature selection module 560 can automatically analyze the image information of the standardized target hand bone X-ray image data, and automatically extract the corresponding image feature values.
  • the target feature selection module 560 can automatically cut the palm region and the background region in the target hand bone X-ray image data, and use the image of the palm region as a positive sample, and the image outside the palm region as a negative sample. Then, the aforementioned positive samples and negative samples are processed by the target feature selection module 560 to obtain the target image feature values of the normalized target hand bone X-ray image data for subsequent analysis.
  • the comparison module 570 is configured to analyze the aforementioned target image feature values using the aforementioned bone age assessment and height prediction model to obtain target image feature value weight data, and perform the aforementioned target image feature value weight data and the aforementioned reference database The comparison is to output the judgment result of the hand bone development state of the subject, the judgment result of the subject's bone age, and the prediction result of the subject's adult height.
  • the comparison module 570 can further standardize target hand bone X-rays of subjects of different genders.
  • the image data was compared with the male bone age height atlas data subset or female bone age height atlas data subset to analyze and predict the hand bone development status, bone age, and adult height of subjects of different genders.
  • the bone age assessment and height prediction system 300 of the present invention may further include a warning module (not shown).
  • a warning module may issue an active alert notification as soon as possible, To facilitate the implementation of subsequent treatment or other relevant countermeasures.
  • the reference database used in the present invention is retrospective pediatric bone age X-ray image data collected by the affiliated hospital of China Medical University, and it is a clinical trial plan approved by the Research Ethics Committee of China Medical University Hospital & Hospital Research, Ethics and Committee. The number is: CMUH 107-REC2-097.
  • the aforementioned reference database contains reference hand X-ray data of 2,758 male subjects and 4,462 female subjects, totaling 7,220 subjects. The age range of the subjects was between 2 and 16 years old.
  • the aforementioned image formats of the reference hand bone X-ray image data are all image formats of the Digital Medical Image Storage Standard Agreement, in order to correlate relevant data such as the physiological age information, gender information, medical record number, and test number of each subject. Stored in the file header of the image data to facilitate subsequent analysis.
  • the aforementioned reference database also includes a bone age height atlas data set.
  • the aforementioned reference hand bone X-ray image data is the reference hand bone X-ray image data of the subject's non-accustomed hands, to avoid the influence of the frequency of habitual hand use or the skeletal shape variation caused by the use habit.
  • the credibility of the reference database, and the bone age height atlas data set contains reference data such as bone growth atlas and growth curve atlas.
  • the bone age height atlas data set of this test example may include a male bone age height atlas data sub-set and a female bone age height atlas data sub-set to analyze reference subjects of different genders.
  • FIG. 4 is a flowchart illustrating a part of the establishment steps of the bone age assessment and height prediction model (not shown) of the present invention.
  • reference hand bone X-ray image data 611 a, reference hand bone X-ray image data 611 b, and reference hand bone X-ray image data 611 c are taken as examples to illustrate the bone age evaluation and height prediction model of the present invention. Operation method and analysis method.
  • the reference hand bone X-ray image data 611a, the reference hand bone X-ray image data 611b, and the reference hand bone X-ray image data 611c will be subjected to image pre-processing step 620 to size them. And chroma standardization to obtain standardized hand bone X-ray image data 621a, standardized hand bone X-ray image data 621b, and standardized hand bone X-ray image data 621c.
  • the image pre-processing step 620 is an image using the image data editing module (not shown) to refer to the reference hand bone X-ray image data 611a, the reference hand bone X-ray image data 611b, and the reference hand bone X-ray image data 611c.
  • the size is adjusted to 256 pixels x 256 pixels, and the black and white contrast is further adjusted to increase the sharpness of the image and reduce the difference in black and white between different reference hand bone X-ray image data.
  • the image data editing module may further perform image chroma expansion processing on each reference hand bone X-ray image data according to requirements, which is to calculate the image gray level of each reference hand bone X-ray image data, and according to the aforementioned calculation
  • the image pixel rows and columns of the reference hand bone X-ray image data 611a, the reference hand bone X-ray image data 611b, and the reference hand bone X-ray image data 611c are automatically filled with colors to convert them into color tones, thereby improving Accuracy of subsequent analysis.
  • the bone age assessment and height prediction model of the present invention can directly extract standardized hand bone X-ray image data 621a, standardized hand bone X-ray image data 621b, and standardized hand bone X-ray image.
  • the physiological age information and gender information of the data 621c is beneficial to eliminate additional analysis procedures and improve the efficiency of analysis.
  • the aforementioned standardized hand bone X-ray image data 621a, standardized hand bone X-ray image data 621b, and standardized hand bone X-ray image data 621c obtained through the image pre-processing step 620 will be further subjected to a feature selection step 630, respectively, to utilize the feature selection module (Not shown) to obtain at least one image feature value after analysis.
  • the feature selection module can cut the palm area and the background area in the standardized hand bone X-ray image data 621a, the standardized hand bone X-ray image data 621b, and the standardized hand bone X-ray image data 621c, respectively, and the palm area
  • the image is used as a positive sample, and the image outside the palm area is a negative sample.
  • the aforementioned positive sample and negative sample are processed by a feature selection module to obtain respective image feature values.
  • FIG. 5 is a schematic diagram illustrating the architecture of a convolutional neural network learning classifier 641 of the bone age assessment and height prediction model of the present invention.
  • the convolutional neural network learning classifier 641 is an Inception-ResNet-v2 convolutional neural network, which includes multiple convolution layers, multiple maximum pooling layers (MaxPool), multiple Average pooling layer (AvgPool) and multiple cascade layers (Concat) to train and analyze image feature values.
  • MaxPool multiple maximum pooling layers
  • AvgPool multiple Average pooling layer
  • Concat cascade layers
  • a two-layer convolution is performed on the image feature values of the standardized hand bone X-ray image data 621a, the standardized hand bone X-ray image data 621b, and the standardized hand bone X-ray image data 621c.
  • Layer and a maximum pooling layer (MaxPool) processing to maximize the output of the extracted image feature values, and repeat the aforementioned two-layer convolutional layer and one layer of maximum pooling layer output, and then use multiple convolutions
  • the layers perform parallel towers training to complete the primary training of image feature values (Inception).
  • the image feature values of the standardized hand bone X-ray image data 621a, the standardized hand bone X-ray image data 621b, and the standardized hand bone X-ray image data 621c will be performed 10 times (10 ⁇ ), 20 Residual module training at different depths, different levels, and different aspects (20 ⁇ ) and 10 times (10 ⁇ ) to train image feature values and achieve convergence.
  • training with the residual module can prevent the convolutional neural network learning classifier 641 from degrading the gradient after the multi-layer training of the aforementioned image feature values, and can effectively improve the convolutional neural network learning classifier. 641 training efficiency.
  • the residual module After completing the deep and repeated training of the residual module, it will sequentially replace the global average pooling layer (GloAvePool2D) with the layer convolution layer, the average pooling layer, and the linear rectifier unit training layer (Rectified LinearLiner). Unit (ReLU) performs final training and processing on the convergent image feature values to determine the hand bone development status of the subject, the subject's bone age, and predict the subject's adult height.
  • the average pooling layer can first calculate the image feature values that have completed the training of the residual module to find the average value of each image feature value.
  • the convolutional neural network can learn the entirety of the classifier 641.
  • the network architecture performs regularization processing to prevent the convolutional neural network learning classifier 641 from overfitting in the training mode that pursues low error, which causes the error value of the judgment result to be too high, resulting in bone age evaluation and
  • the results of the height prediction model are less reliable than expected.
  • the training layer of the linear rectifier unit further activates the image feature values after training and outputs target image feature value weight data 650 for subsequent comparison and analysis.
  • the aforementioned training layer of the linear rectification unit can prevent the target image feature value weight data 650 output from the bone age evaluation and the height prediction model from approaching zero or approaching infinity, which is beneficial to the subsequent comparison steps and further improves the bone age of the present invention. Evaluation and judgment accuracy of height prediction models.
  • the bone age assessment and height prediction model of the present invention is used to judge the hand bone development status, bone age and adult height of a subject
  • the established bone age assessment and height prediction model is further used to judge the hand bone development status of the subject, the subject's bone age, and predict the subject's adult height.
  • the steps are as follows: Provide the previously established bone age assessment and height prediction model.
  • Provide the target hand bone X-ray image data where the target hand bone X-ray image data includes physiological age information and gender information.
  • the target hand bone X-ray image data is pre-processed.
  • the aforementioned image data editing module is used to adjust the image size and black and white contrast of the target hand bone X-ray image data to obtain a standardized target hand bone X-ray image data.
  • the at least one image feature value is obtained after analyzing the standardized target hand bone X-ray image data by using the aforementioned feature selection module.
  • Using the aforementioned bone age assessment and height prediction model to analyze image feature values to determine the hand bone development status of the subject, the subject's bone age, and predict the subject's adult height.
  • the previously established bone age assessment and height prediction model will be applied to the bone age assessment and height prediction system of the present invention, so as to judge the subject's hand bone development status judgment result, the subject's bone age judgment result, and the test subject.
  • the adult height prediction results were further integrated into the reference database to optimize the bone age assessment and height prediction model.
  • the detailed structure of the bone age assessment and height prediction system of the present invention has been described in FIG. 3 and the foregoing, and will not be repeated here.
  • FIG. 6 is a schematic diagram illustrating an application result 700 of the bone age assessment and height prediction system (not shown) of the present invention.
  • Bone age assessment and height prediction system After the analysis of its bone age assessment and height prediction model is completed, it will further output the judgment result of the hand's bone development state and the subject's bone age judgment result, and it can be displayed on the display module ( Figure (Not shown).
  • the application result 700 of the bone age assessment and height prediction system of the present invention may include a result field 701, a result field 702, a result field 703, and a result field 704.
  • the result field 701 can display the basic data of the subject, which includes the subject's physiological age information, gender information, and other personal data such as the medical record number and test number.
  • the result field 702 is the test subject that has not undergone pre-image processing.
  • the target hand bone X-ray image data, the result field 703 is the bone age of the subject judged by the bone age assessment and height prediction model, and the result field 704 is the subject judged by the aforementioned bone age assessment and height prediction system
  • the bone age atlas of 12 months before and after the results of the bone age for comparison and analysis by subsequent analysts.
  • the bone age assessment and height prediction system of the present invention can further determine the hand bone development status judgment result of the subject and the subject's bone age judgment result with the bone age height atlas data of the male bone age height atlas data.
  • Sub-sets or female bone age height atlas data sub-sets are compared to predict the adult height of subjects of different genders, and the predicted results of the adult heights of the subjects are simultaneously output and displayed in the aforementioned display module, and
  • the present invention is not limited to the contents disclosed in the foregoing description or drawings.
  • the bone age assessment and height prediction system of the present invention may further include a warning module (not shown in the figure).
  • a warning module (not shown in the figure).
  • the alert module will issue an active alert notification as soon as possible, It is displayed in red in the result column 703 to facilitate the implementation of subsequent treatment or other relevant countermeasures.
  • the bone age assessment and height prediction model, the bone age assessment and height prediction system, and the bone age assessment and height prediction method of the present invention can automatically perform image characteristics on the subject's target hand bone X-ray image data through the bone age assessment and height prediction model.
  • the value extraction and deep neural network training can not only effectively reduce the time required for bone age assessment and height prediction, but also avoid the result error caused by different analysts' feature selection and comparison methods in the known bone age assessment methods.
  • the bone age assessment and height prediction model including a convolutional neural network learning classifier can not only effectively improve the accuracy and sensitivity of the bone age assessment and height prediction, so that the bone age assessment and height prediction model, bone age assessment, and height of the present invention can be effectively improved.
  • Prediction systems and bone age assessment and height prediction methods are more efficient in determining bone age and height prediction, and can implement appropriate treatment or related application measures based on the bone age judgment results of individual cases to reduce children's developmental delay or precocity. Of disease.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Physics & Mathematics (AREA)
  • Primary Health Care (AREA)
  • Epidemiology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • Pathology (AREA)
  • Molecular Biology (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Databases & Information Systems (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Optics & Photonics (AREA)
  • Veterinary Medicine (AREA)
  • Animal Behavior & Ethology (AREA)
  • Surgery (AREA)
  • Biophysics (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • Artificial Intelligence (AREA)
  • Dentistry (AREA)
  • Physiology (AREA)
  • Orthopedic Medicine & Surgery (AREA)
  • Quality & Reliability (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Apparatus For Radiation Diagnosis (AREA)
  • Image Analysis (AREA)
  • Medical Treatment And Welfare Office Work (AREA)

Abstract

本发明公开了一种骨龄评估与身高预测模型、其系统及其预测方法，骨龄评估与身高预测系统包含影像撷取单元以及非瞬态机器可读介质。影像撷取单元用以取得受试者的目标手骨X光影像数据。非瞬态机器可读介质用以储存程序，当程序由处理单元执行时用以判断受试者的手骨发育状态、骨龄和预测受试者的成年身高。借此，本发明的骨龄评估与身高预测系统可有效提升骨龄评估及身高预测的准确度与敏感度，并可缩短骨龄评估及身高预测的判定时间。

Description

骨龄评估与身高预测模型、其系统及其预测方法 技术领域
本发明是有关于一种医疗信息分析模型、系统以及方法，特别是一种骨龄评估与身高预测模型、骨龄评估与身高预测系统以及骨龄评估与身高预测方法。
背景技术
骨骼年龄为人体生理年龄的重要指标之一，其通过骨骼的生长、发育、成熟、衰老的规律来推断人体的生理年龄。骨龄评估(bone age assessment，BAA)为小儿科医师常用以判读儿童生长发育的常规检查，其通过分析不同生长阶段时骨骼的不同型态表现，并参照人体骨骼的连续性与阶段性的发育状态，进而准确地评估个体的生长发育水平和成熟程度，并可进一步评估与预测受试儿童的生长发育潜力以及性成熟的趋势。
公知的骨龄评估方式系利用低剂量的X光摄影方式取得受试者的左手或右手的指骨、掌骨与腕骨的X光影像，并将前述的X光影像通过Greulich and Pyle(G-P)方法与Tanner-Whitehouse(TW)方法而以图谱比对的方式进行骨龄评估。G-P方法在操作上是以人工方式将受试者手骨的原始X光影像与数据库中的手骨X光片影像依据不同年龄区间进行一对一的比对，而TW方法则是取左手掌与左手腕的手骨X光影像中二十个感兴趣区域(Regions of Interests，ROI)进行逐一比对分析，并将手骨的发育状况分成九个成熟等级，以进行后续的评估。然而，当使用G-P方法进行骨龄评估时，同一受试者的骨龄评估结果容易因为不同分析者的不同比对习惯而有所不同，而利用TW方法进行骨龄评估时所得的骨龄评估结果虽较为客观，但因所需评分的骨头较多，过程也较为繁琐耗时，并无法在短时间内对受试者手骨的原始X光影像进行分析并即时获得相应的骨龄评估结果。
因此，如何发展出一种具有高度准确率及快速检测的骨龄评估及身高预测系统，实为一具有商业价值的技术课题。
发明内容
本发明的目的在于提供一种骨龄评估与身高预测模型、其系统及其预测方法，其可有效提升骨龄评估及身高预测的准确度与敏感度，并可缩短骨龄评估及身高预测的判定时间。
本发明的一方面是在于提供一种骨龄评估与身高预测模型，所述骨龄评估与身高预测模型由以下步骤建立：取得参照数据库、进行影像前处理步骤、进行特征选取步骤以及进 行训练步骤。参照数据库包含骨龄身高图谱数据集合及多个参照手骨X光影像数据，其中各参照手骨X光影像数据包含生理年龄信息以及性别信息。影像前处理步骤系利用影像数据编辑模块调整各参照手骨X光影像数据的影像大小及影像黑白对比度，以取得多个标准化手骨X光影像数据。特征选取步骤是利用特征选取模块分析标准化手骨X光影像数据后以得至少一个影像特征值。训练步骤是将影像特征值通过卷积神经网络学习分类器进行训练而达到收敛，以得骨龄评估与身高预测模型，其中骨龄评估与身高预测模型是用以判断受试者的手骨发育状态、受试者的骨龄以及预测受试者的成年身高。
依据前述的骨龄评估与身高预测模型，其中卷积神经网络学习分类器可为Inception-ResNet-v2卷积神经网络。
依据前述的骨龄评估与身高预测模型，其中参照手骨X光影像数据的影像格式可为数字医疗影像储存标准协定(Digital Imaging and Communications in Medicine，DICOM)的影像格式。
依据前述的骨龄评估与身高预测模型，其中影像前处理步骤可还对各参照手骨X光影像数据进行影像色度扩展处理。
依据前述的骨龄评估与身高预测模型，其中骨龄身高图谱数据集合可包含男性骨龄身高图谱数据子集合及女性骨龄身高图谱数据子集合。
依据前述的骨龄评估与身高预测模型，其中各参照手骨X光影像数据可为非习惯用手的参照手骨X光影像数据。
本发明的另一方面是在于提供一种骨龄评估与身高预测方法，包含下述步骤。提供如前段的骨龄评估与身高预测模型。提供受试者的目标手骨X光影像数据，其中前述的目标手骨X光影像数据包含生理年龄信息以及性别信息。对目标手骨X光影像数据进行前处理，其是利用前述的影像数据编辑模块调整目标手骨X光影像数据的影像大小及影像黑白对比度，以取得标准化目标手骨X光影像数据。利用前述的特征选取模块分析标准化目标手骨X光影像数据后以得至少一个影像特征值。利用前述的骨龄评估与身高预测模型分析影像特征值，以判断受试者的手骨发育状态、受试者的骨龄以及预测受试者的成年身高。
依据前述的骨龄评估与身高预测方法，其中目标手骨X光影像数据的影像格式可为数字医疗影像储存标准协定的影像格式。
依据前述的骨龄评估与身高预测方法，其中影像数据编辑模块可还对目标手骨X光影像数据进行影像色度扩展处理。
依据前述的骨龄评估与身高预测方法，其中前述的目标手骨X光影像数据可为非习惯 用手的目标手骨X光影像数据。
本发明的又一方面是在于提供一种骨龄评估与身高预测系统，包含影像撷取单元以及非瞬态机器可读介质。影像撷取单元用以取得受试者的目标手骨X光影像数据，其中前述的目标手骨X光影像数据包含生理年龄信息以及性别信息。非瞬态机器可读介质信号连接前述的影像撷取单元，其中非瞬态机器可读介质用以储存程序，当前述的程序由处理单元执行时是用以评估受试者的手骨发育状态、受试者的骨龄和预测受试者的成年身高，且前述的程序包含参照数据库取得模块、第一影像数据编辑模块、特征选取模块、训练模块、第二影像数据编辑模块、目标特征选取模块及比对模块。参照数据库取得模块用以取得参照数据库，且前述的参照数据库包含骨龄身高图谱数据集合及多个参照手骨X光影像数据，其中各参照手骨X光影像数据包含生理年龄信息以及性别信息。第一影像数据编辑模块系调整各参照手骨X光影像数据的影像大小及影像黑白对比度，以取得多个标准化手骨X光影像数据。特征选取模块用以分析标准化手骨X光影像数据后以得至少一个参照影像特征值。训练模块用以将前述的参照影像特征值通过卷积神经网络学习分类器进行训练而达到收敛，以得到骨龄评估与身高预测模型。第二影像数据编辑模块系调整目标手骨X光影像数据的影像大小及影像黑白对比度，以取得标准化目标手骨X光影像数据。目标特征选取模块用以分析标准化目标手骨X光影像数据后以得至少一个目标影像特征值。比对模块用以将前述的至少一个目标影像特征值以前述的骨龄评估与身高预测模型进行分析，以得到目标影像特征值权重数据，并将前述的至少一个目标影像特征值权重数据与参照数据库进行比对，以输出受试者的手骨发育状态判断结果、受试者的骨龄判断结果以及受试者的成年身高预测结果。
依据前述的种骨龄评估与身高预测系统，其中卷积神经网络学习分类器可为Inception-ResNet-v2卷积神经网络。
依据前述的种骨龄评估与身高预测系统，其中前述的目标手骨X光影像数据的影像格式可为数字医疗影像储存标准协定的影像格式，前述的参照手骨X光影像数据与的影像格式可为数字医疗影像储存标准协定的影像格式。
依据前述的种骨龄评估与身高预测系统，其中第一影像数据编辑模块可还对各参照手骨X光影像数据进行影像色度扩展处理，第二影像数据编辑模块可还对目标手骨X光影像数据进行影像色度扩展处理。
依据前述的种骨龄评估与身高预测系统，其中骨龄身高图谱数据集合可包含男性骨龄身高图谱数据子集合及女性骨龄身高图谱数据子集合。
依据前述的种骨龄评估与身高预测系统，其中各参照手骨X光影像数据为非习惯用手的参照手骨X光影像数据，目标手骨X光影像数据为非习惯用手的目标手骨X光影像数据。
依据前述的种骨龄评估与身高预测系统，可还包含警示模块，用以在标准化目标手骨X光影像数据通过骨龄评估与身高预测模型进行分析以后，发出主动警示通知。
借此，本发明的骨龄评估与身高预测模型、骨龄评估与身高预测系统及骨龄评估与身高预测方法通过将参照手骨X光影像数据与目标手骨X光影像数据进行影像标准化前处理，并利用特征选取模块分析并得至少一个影像特征值后，再以卷积神经网络对影像特征值进行训练，以对手骨发育状态、骨龄以及成年身高进行分析判断，不仅可有效缩短骨龄评估与身高预测所需的时间，亦可避免公知的骨龄评估方式中因不同分析者的特征选取及比对方式的不同所造成的结果误差。再者，通过包含卷积神经网络学习分类器的骨龄评估与身高预测模型能有效提升骨龄评估及身高预测的准确度与敏感度，不仅使本发明的骨龄评估与身高预测模型、骨龄评估与身高预测系统及骨龄评估与身高预测方法在骨龄判断及身高预测方面更有效率，并可准确地评估不同受试者的生长发育水平和成熟程度，以及预测受试者未来的生长发育潜力。
上述发明内容旨在提供本揭示内容的简化摘要，以使阅读者对本揭示内容具备基本的理解。此发明内容并非本揭示内容的完整概述，且其用意并非在指出本发明实施例的重要/关键元件或界定本发明的范围。
附图说明
为让本发明的上述和其他目的、特征、优点与实施例能更明显易懂，结合附图说明如下:
图1是绘示本发明一实施方式的骨龄评估与身高预测模型的建立步骤流程图；
图2是绘示本发明另一实施方式的骨龄评估与身高预测方法的步骤流程图；
图3是绘示本发明再一实施方式的骨龄评估与身高预测系统的架构示意图；
图4是绘示本发明的骨龄评估与身高预测模型的部分建立步骤流程图；
图5是绘示本发明的骨龄评估与身高预测模型的卷积神经网络学习分类器的架构示意图；以及
图6是绘示本发明的骨龄评估与身高预测系统的应用结果的示意图。
具体实施方式
下述将更详细讨论本发明各实施方式。然而，此实施方式可为各种发明概念的应用，可被具体实行在各种不同的特定范围内。特定的实施方式是仅以说明为目的，且不受限于公开的范围。
请参照图1，其是绘示本发明一实施方式的骨龄评估与身高预测模型100的建立步骤流程图。骨龄评估与身高预测模型100是用以判断受试者的手骨发育状态、受试者的骨龄以及预测受试者的成年身高，且骨龄评估与身高预测模型100的建立步骤包含步骤110、步骤120、步骤130以及步骤140。
步骤110为取得参照数据库，其中参照数据库包含骨龄身高图谱数据集合及多个参照手骨X光影像数据，其中各参照手骨X光影像数据包含生理年龄信息以及性别信息。优选地，前述的参照手骨X光影像数据可为非习惯用手的参照手骨X光影像数据，避免本发明的骨龄评估与身高预测模型100受习惯用手的使用频率或使用习惯所导致的骨骼型态变异而影响其判断准确率。
优选地，前述的参照手骨X光影像数据的影像格式可为数字医疗影像储存标准协定(Digital Imaging and Communications in Medicine，DICOM)的影像格式，以将各参照手骨X光影像数据的生理年龄信息、性别信息等基本数据储存于参照手骨X光影像数据的档头(header)中，以利于后续的分析。再者，由于男性与女性的生理成熟历程并不尽相同，骨骼的发育形态及其对应的生理年龄也不相同，是以本发明的骨龄评估与身高预测模型100可进一步对不同性别的参照手骨X光影像数据分别进行特征选取步骤与训练步骤，以进行性别化的手骨发育状态、骨龄及成年身高的判断与预测。优选地，前述的骨龄身高图谱数据集合可包含男性骨龄身高图谱数据子集合及女性骨龄身高图谱数据子集合，以利于对不同性别的受试者进行分析。
步骤120为进行影像前处理步骤，其是利用影像数据编辑模块调整各参照手骨X光影像数据的影像大小及影像黑白对比度，以取得多个标准化手骨X光影像数据。详细而言，影像数据编辑模块可分别将不同的参照手骨X光影像数据的影像大小调整为256像素(pixel)×256像素后，并调整其黑白对比度，以减少不同参照手骨X光影像数据之间的黑白色度差异以及增加影像的清晰度，以利于后续的分析。
另外，在步骤120中，影像数据编辑模块可进一步对各参照手骨X光影像数据进行影像色度扩展处理。详细而言，影像数据编辑模块可计算各参照手骨X光影像数据的影像灰阶程度，并依据前述的计算结果而依序对各参照手骨X光影像数据的影像像素行、列自动填补色彩，以将呈现灰阶色调的各参照手骨X光影像数据转换为彩色色调，进而提升后续 分析的准确度，但本发明并不以前述说明与附图公开的内容为限。
步骤130为进行特征选取步骤，其是利用特征选取模块分析标准化手骨X光影像数据后以得至少一个影像特征值。详细而言，本发明的骨龄评估与身高预测模型100可利用特征选取模块自动地对标准化手骨X光影像数据的影像信息进行分析，并自动提取对应的影像特征值，借以增进本发明的骨龄评估与身高预测模型100的评估与预测效率。
步骤140为进行训练步骤，其是将前述的影像特征值通过卷积神经网络学习分类器进行训练而达到收敛，以得骨龄评估与身高预测模型100。优选地，前述的卷积神经网络学习分类器可为Inception-ResNet-v2卷积神经网络。Inception-ResNet-v2卷积神经网络为一种基于ImageNet可视化数据数据库的大规模视觉辨识(Large Scale Visual Recognition)卷积神经网络，其通过残差连接(Residual connections)的方式而可有效扩展卷积神经网络的训练深度，进而使Inception-ResNet-v2卷积神经网络于图像分类与辨识方面具有相当高的准确率。
请参照图2，其是绘示本发明另一实施方式的骨龄评估与身高预测方法200的步骤流程图。骨龄评估与身高预测方法200包含步骤210、步骤220、步骤230、步骤240以及步骤250。
步骤210为提供骨龄评估与身高预测模型，而骨龄评估与身高预测模型系经由前述步骤110至步骤140所建立。
步骤220为提供受试者的目标手骨X光影像数据，其中目标手骨X光影像数据包含生理年龄信息以及性别信息。优选地，前述的目标手骨X光影像数据可为非习惯用手的目标手骨X光影像数据，避免受试者的受习惯用手的使用频率或使用习惯所导致的骨骼型态变异影响骨龄评估与身高预测方法200的分析准确率。
优选地，前述的目标手骨X光影像数据的影像格式可为数字医疗影像储存标准协定的影像格式，以将目标手骨X光影像数据的生理年龄信息、性别信息等基本数据储存于目标手骨X光影像数据的档头中，以利于后续的分析。再者，由于男性与女性在生理成熟历程并不尽相同，骨骼的发育形态及其对应的生理年龄也不相同，故本发明的骨龄评估与身高预测方法200将分别对不同性别的目标手骨X光影像数据进行性别化的手骨发育状态、骨龄及其成年身高的评估与分析。
步骤230为对目标手骨X光影像数据进行前处理，其是利用前述的影像数据编辑模块调整目标手骨X光影像数据的影像大小及影像黑白对比度，以取得标准化目标手骨X光影像数据。详细而言，影像数据编辑模块是将目标手骨X光影像数据的影像大小调整为 256像素×256像素后，并调整其黑白对比度以增加影像的清晰度，以利于后续的分析。
另外，在步骤230中，影像数据编辑模块可进一步对目标手骨X光影像数据进行影像色度扩展处理。详细而言，影像数据编辑模块可计算目标手骨X光影像数据的影像灰阶程度，并依据前述的计算结果而依序对目标手骨X光影像数据的影像像素行、列自动填补色彩，以将呈现灰阶色调的目标手骨X光影像数据转换为彩色色调，进而提升后续分析的准确度，但本发明并不以前述说明与附图公开的内容为限。
步骤240为利用特征选取模块分析标准化目标手骨X光影像数据后以得至少一个影像特征值。详细而言，本发明的骨龄评估与身高预测方法200可利用特征选取模块自动地对标准化目标手骨X光影像数据的影像信息进行分析，并自动提取对应的影像特征值，借以增进本发明的骨龄评估与身高预测方法200的评估与预测效率。
步骤250为利用前述的骨龄评估与身高预测模型分析影像特征值，以判断受试者的手骨发育状态、受试者的骨龄以及预测受试者的成年身高。
请参照图3，其是绘示绘示本发明再一实施方式的骨龄评估与身高预测系统300的架构示意图。骨龄评估与身高预测系统300包含影像撷取单元400以及非瞬态机器可读介质500。
影像撷取单元400是用以取得受试者的目标手骨X光影像数据，其中目标手骨X光影像数据包含生理年龄信息以及性别信息。详细而言，影像撷取单元400可为X光检测仪器，其利用低剂量的X光射线照射受试者的手部，以取得解析度适当的目标手骨X光影像数据。优选地，前述的目标手骨X光影像数据可为非习惯用手的目标手骨X光影像数据，避免本发明的骨龄评估与身高预测系统300受习惯用手的使用频率或使用习惯所导致的骨骼型态变异而影响其分析准确率。优选地，前述的目标手骨X光影像数据的影像格式可为数字医疗影像储存标准协定的影像格式，以将目标手骨X光影像数据的生理年龄信息、性别信息等基本数据储存于目标手骨X光影像数据的档头中，以利于后续的分析。
非瞬态机器可读介质500信号连接影像撷取单元400，其中非瞬态机器可读介质500用以储存程序(图未绘示)，当前述的程序由处理单元(图未绘示)执行时是用以评估受试者的手骨发育状态、受试者的骨龄和预测受试者的成年身高，且前述的程序包含参照数据库取得模块510、第一影像数据编辑模块520、特征选取模块530、训练模块540、第二影像数据编辑模块550、目标特征选取模块560及比对模块570。
参照数据库取得模块510用以取得参照数据库，且前述的参照数据库包含骨龄身高图谱数据集合及多个参照手骨X光影像数据，其中各参照手骨X光影像数据包含生理年龄 信息以及性别信息。优选地，各参照手骨X光影像数据可为非习惯用手的参照手骨X光影像数据，而前述的骨龄身高图谱数据集合可包含男性骨龄身高图谱数据子集合及女性骨龄身高图谱数据子集合，以利于对不同性别的受试者进行分析。
优选地，前述的参照手骨X光影像数据的影像格式可为数字医疗影像储存标准协定的影像格式，以将各参照手骨X光影像数据的生理年龄信息、性别信息等基本数据储存于参照手骨X光影像数据的档头中，以利于后续的分析。
第一影像数据编辑模块520是调整各参照手骨X光影像数据的影像大小及影像黑白对比度，以取得多个标准化手骨X光影像数据。详细而言，第一影像数据编辑模块520是将不同的参照手骨X光影像数据的影像大小调整为256像素×256像素后，并调整其黑白对比度，以减少不同参照手骨X光影像数据之间的黑白色度差异以及增加影像的清晰度。
另外，第一影像数据编辑模块520可更进一步对各参照手骨X光影像数据进行影像色度扩展处理。详细而言，第一影像数据编辑模块520可计算各参照手骨X光影像数据的影像灰阶程度，并依据前述的计算结果而依序对各参照手骨X光影像数据的影像像素行、列自动填补色彩，以将呈现灰阶色调的各参照手骨X光影像数据转换为彩色色调，以提升后续分析的准确度，但本发明并不以前述说明与附图公开的内容为限。
特征选取模块530用以分析标准化手骨X光影像数据后以得至少一个参照影像特征值。详细而言，本发明的骨龄评估与身高预测系统300可利用特征选取模块530自动地对标准化手骨X光影像数据的影像信息进行分析，并自动提取对应的影像特征值。
训练模块540用以将前述的参照影像特征值通过卷积神经网络学习分类器进行训练而达到收敛，以得到骨龄评估与身高预测模型。优选地，前述的卷积神经网络学习分类器可为Inception-ResNet-v2卷积神经网络，以有效扩展卷积神经网络的训练深度，进而提升训练模块540的图像分类与辨识能力。
第二影像数据编辑模块550是调整目标手骨X光影像数据的影像大小及影像黑白对比度，以取得标准化目标手骨X光影像数据。详细而言，第二影像数据编辑模块550是将目标手骨X光影像数据的影像大小调整为256像素×256像素后，并调整其黑白对比度以增加影像的清晰度，进而获得前述的标准化目标手骨X光影像数据。优选地，第二影像数据编辑模块550可进一步对目标手骨X光影像数据进行影像色度扩展处理，其是计算目标手骨X光影像数据的影像灰阶程度，并根据前述的计算结果而依序对目标手骨X光影像数据的影像像素行、列自动填补色彩，以将呈现灰阶色调的目标手骨X光影像数据转换为彩色色调，但本发明并不以前述说明与附图公开的内容为限。
目标特征选取模块560用以分析标准化目标手骨X光影像数据后以得至少一个目标影像特征值。详细而言，目标特征选取模块560可自动地对标准化目标手骨X光影像数据的影像信息进行分析，并自动提取对应的影像特征值。具体来说，目标特征选取模块560可自动将目标手骨X光影像数据中的手掌区域和背景区域进行切割，并将手掌区域的影像作为正样本，而手掌区域的外的影像为负样本，接着将前述的正样本与负样本以目标特征选取模块560进行处理后以获得标准化目标手骨X光影像数据的目标影像特征值，以进行后续的分析。
比对模块570用以将前述的目标影像特征值以前述的骨龄评估与身高预测模型进行分析，以得到目标影像特征值权重数据，并将前述的目标影像特征值权重数据与前述的参照数据库进行比对，以输出受试者的手骨发育状态判断结果、受试者的骨龄判断结果以及受试者的成年身高预测结果。
再者，由于男性与女性在生理成熟历程不尽相同，骨骼的发育形态及其对应的生理年龄也不相同，故比对模块570可进一步将不同性别的受试者的标准化目标手骨X光影像数据分别与男性骨龄身高图谱数据子集合或女性骨龄身高图谱数据子集合进行比对，以对不同性别的受试者进行手骨发育状态、骨龄及其成年身高的分析与预测。
再者，虽图未绘示，本发明的骨龄评估与身高预测系统300可还包含警示模块(图未绘示)。当标准化目标手骨X光影像数据与骨龄身高图谱数据集合进行比对后，倘若受试者的骨龄比对结果明显超前或落后其生理年龄时，警示模块可在第一时间发出主动警示通知，以利于后续的治疗或其他相关应对措施的实施。
根据上述实施方式，以下提出具体试验例并配合附图予以详细说明。
<试验例>
一、参照数据库
本发明所使用的参照数据库为中国医学大学附属医院所搜集的回溯性儿科骨龄X光影像数据，为经中国医药大学附属医院研究伦理委员会(China Medical University&Hospital Research Ethics Committee)核准的临床试验计划，其编号为：CMUH 107-REC2-097。前述的参照数据库包含2758位男性受试者以及4462位女性受试者、共计7220位受试者的参照手骨X光影像数据，而受试者的年龄范围则落于2岁至16岁之间，且前述的参照手骨X光影像数据的影像格式皆为数字医疗影像储存标准协定的影像格式，以将各受试者的生理年龄信息、性别信息、病历号码、受试编号等相关数据储存于影像数据的档头中，以利于后续的分析。
前述的参照数据库亦包含骨龄身高图谱数据集合。详细而言，前述的参照手骨X光影像数据为受试者的非习惯用手的参照手骨X光影像数据，避免习惯用手的使用频率或使用习惯所导致的骨骼型态变异而影响参照数据库的可信度，而骨龄身高图谱数据集合则包含骨骼生长图谱、生长曲线图谱等参照数据。再者，本试验例的骨龄身高图谱数据集合可包含男性骨龄身高图谱数据子集合及女性骨龄身高图谱数据子集合，以对不同性别的参照受试者进行分析。
二、本发明的骨龄评估与身高预测模型
请参照图4，其是绘示本发明的骨龄评估与身高预测模型(图未绘示)的部分建立步骤流程图。在图4的试验例中将以参照手骨X光影像数据611a、参照手骨X光影像数据611b与参照手骨X光影像数据611c为例，以说明本发明的骨龄评估与身高预测模型的操作方法以及分析方式。
首先，在取得前述的参照数据库后，参照手骨X光影像数据611a、参照手骨X光影像数据611b与参照手骨X光影像数据611c将分别进行影像前处理步骤620，以对其进行尺寸与色度的标准化，借以取得标准化手骨X光影像数据621a、标准化手骨X光影像数据621b与标准化手骨X光影像数据621c。详细而言，影像前处理步骤620是利用影像数据编辑模块(图未绘示)将参照手骨X光影像数据611a、参照手骨X光影像数据611b与参照手骨X光影像数据611c的影像大小调整为256像素×256像素，并进一步调整其黑白对比度以增加影像的清晰度，以及减少不同参照手骨X光影像数据之间的黑白色度差异。
优选地，影像数据编辑模块可视需求而进一步对各参照手骨X光影像数据进行影像色度扩展处理，其是计算各参照手骨X光影像数据的影像灰阶程度，并依据前述的计算结果而分别对参照手骨X光影像数据611a、参照手骨X光影像数据611b与参照手骨X光影像数据611c的影像像素行、列自动填补色彩，以将其转换为彩色色调，进而提升后续分析的准确度。
再者，由于各受试者的生理年龄信息、性别信息等基本数据是直接储存于呈现数字医疗影像储存标准协定的影像格式的标准化手骨X光影像数据621a、标准化手骨X光影像数据621b与标准化手骨X光影像数据621c的档头中，本发明的骨龄评估与身高预测模型可直接提取标准化手骨X光影像数据621a、标准化手骨X光影像数据621b与标准化手骨X光影像数据621c的生理年龄信息以及性别信息，而无须通过人工方式额外进行标注作业，有利于省去额外的分析程序并提升分析的效率。
前述的经过影像前处理步骤620所得的标准化手骨X光影像数据621a、标准化手骨 X光影像数据621b与标准化手骨X光影像数据621c将进一步分别进行特征选取步骤630，以利用特征选取模块(图未绘示)分析后以得至少一个影像特征值。详细而言，特征选取模块可分别将标准化手骨X光影像数据621a、标准化手骨X光影像数据621b与标准化手骨X光影像数据621c中的手掌区域和背景区域进行切割，并将手掌区域的影像作为正样本，而手掌区域的外的影像为负样本，接着将前述的正样本与负样本以特征选取模块进行处理后以获得各别的影像特征值。
接着，请一并参照图4与图5，图5是绘示本发明的骨龄评估与身高预测模型的卷积神经网络学习分类器641的架构示意图。在图5的试验例中，卷积神经网络学习分类器641为Inception-ResNet-v2卷积神经网络，其包含多个卷积层(Convolution)、多个最大池化层(MaxPool)、多个平均池化层(AvgPool)以及多个级联层(Concat)，以对影像特征值进行训练与分析。
在对影像特征值进行训练的过程中，首先将分别对标准化手骨X光影像数据621a、标准化手骨X光影像数据621b与标准化手骨X光影像数据621c的影像特征值进行二层卷积层及一层最大池化层(MaxPool)处理，以将所提取的影像特征值进行最大输出，并再次重复前述的二层卷积层与一层最大池化层输出后，利用多个卷积层进行并行塔(parallel towers)训练，以完成影像特征值的初级训练(Inception)。
而在完成前述的初级训练后，标准化手骨X光影像数据621a、标准化手骨X光影像数据621b与标准化手骨X光影像数据621c的影像特征值将分别进行10次(10×)、20次(20×)与10次(10×)的不同深度、不同阶层与不同方面的残差(Residual)模块训练，以对影像特征值进行训练并达到收敛。详细而言，利用残差模块进行训练可防止卷积神经网络学习分类器641对前述的影像特征值进行多层的训练后发生梯度消失的退化现象，并可有效提升卷积神经网络学习分类器641的训练效率。
而在完成深层且重复的残差模块训练后，将依序以层卷积层、平均池化层、取代全局平均池化层(Global Average Pooling 2D，GloAvePool2D)以及线性整流单元训练层(Rectified Linear Unit，ReLU)对收敛的影像特征值进行最终训练与处理，借以判断受试者的手骨发育状态、受试者的骨龄以及预测受试者的成年身高。其中，平均池化层可先对完成残差模块训练的影像特征值进行计算，以求各影像特征值的平均值，取代全局平均池化层则可对卷积神经网络学习分类器641的整体网络架构进行正则化(Regularization)处理，防止卷积神经网络学习分类器641在追求低误差的训练模式下发生过拟合现象(Overfitting)，而导致判断结果的误差值过高，致使骨龄评估与身高预测模型的结果可信度不如预期。最后，线 性整流单元训练层则进一步对完成训练后的影像特征值进行激活，并输出目标影像特征值权重数据650，以进行后续的比对与分析。前述的线性整流单元训练层可避免骨龄评估与身高预测模型输出的目标影像特征值权重数据650趋近于零或趋近于无限大，以利于后续比对步骤的进行，进而提升本发明的骨龄评估与身高预测模型的判断准确率。
三、本发明的骨龄评估与身高预测模型用于判断受试者的手骨发育状态、骨龄与成年身高
在本试验例中进一步将所建立的骨龄评估与身高预测模型用于判断受试者的手骨发育状态、受试者的骨龄以及预测受试者的成年身高。其步骤如下：提供前述建立的骨龄评估与身高预测模型。提供受试者的目标手骨X光影像数据，其中目标手骨X光影像数据包含生理年龄信息以及性别信息。对目标手骨X光影像数据进行前处理，其是利用前述的影像数据编辑模块调整目标手骨X光影像数据的影像大小及影像黑白对比度，以取得一标准化目标手骨X光影像数据。利用前述的特征选取模块分析标准化目标手骨X光影像数据后以得至少一个影像特征值。利用前述的骨龄评估与身高预测模型分析影像特征值，以判断受试者的手骨发育状态、受试者的骨龄以及预测受试者的成年身高。
再者，前述所建立的骨龄评估与身高预测模型将应用于本发明的骨龄评估与身高预测系统中，以将受试者的手骨发育状态判断结果、受试者的骨龄判断结果以及受试者的成年身高预测结果进一步整合于参照数据库中，以对骨龄评估与身高预测模型进行优化。此外，本发明的骨龄评估与身高预测系统的细部架构已如图3与前文所述，在此则不再赘述。
请参照图6，其是绘示本发明的骨龄评估与身高预测系统(图未绘示)的应用结果700的示意图。骨龄评估与身高预测系统在其骨龄评估与身高预测模型完成分析后，将进一步输出受试者的手骨发育状态判断结果与受试者的骨龄判断结果，并可将其显示于显示模块(图未绘示)中。而如图6所示，本发明的骨龄评估与身高预测系统的应用结果700可包含结果栏位701、结果栏位702、结果栏位703以及结果栏位704。结果栏位701可显示受试者的基本数据，其包含受试者的生理年龄信息、性别信息及病历号码、受试编号等其他个人数据，结果栏位702为未经过影像前处理的受试者的目标手骨X光影像数据，结果栏位703为骨龄评估与身高预测模型所判断的受试者的骨龄，而结果栏位704则为前述骨龄评估与身高预测系统所判断的受试者的骨龄结果的前后12个月的骨龄图谱，以供后续分析者进行对照与分析。
再者，虽图未揭示，本发明的骨龄评估与身高预测系统可将受试者的手骨发育状态判断结果以及受试者的骨龄判断结果进一步与骨龄身高图谱数据集合的男性骨龄身高图谱 数据子集合或女性骨龄身高图谱数据子集合进行比对，以对不同性别的受试者进行成年身高的预测，并将受试者的成年身高预测结果同步输出并显示于前述的显示模块中，且本发明并不以前述说明或附图公开的内容为限。
另外，虽图未揭示，本发明的骨龄评估与身高预测系统可还包含警示模块(图未绘示)。当骨龄评估与身高预测模型输出受试者的骨龄判断结果后，倘若受试者的骨龄判断结果明显超前或落后于受试者的生理年龄，警示模块将可在第一时间发出主动警示通知，并于结果栏位703中以红字进行显示，以利于后续的治疗或其他相关应对措施的实施。
借此，本发明的骨龄评估与身高预测模型、骨龄评估与身高预测系统及骨龄评估与身高预测方法通过骨龄评估与身高预测模型可自动对受试者的目标手骨X光影像数据进行影像特征值提取以及进行深度神经网络训练，不仅可有效缩短骨龄评估与身高预测所需的时间，亦可避免公知的骨龄评估方式中因不同分析者的特征选取及比对方式的不同所造成的结果误差。再者，通过包含卷积神经网络学习分类器的骨龄评估与身高预测模型不仅能有效提升骨龄评估及身高预测的准确度与敏感度，使本发明的骨龄评估与身高预测模型、骨龄评估与身高预测系统及骨龄评估与身高预测方法在骨龄判断及身高预测方面更有效率，并可通过对个案的骨龄判断结果而实施适当的治疗或相关的应用措施，以降低儿童因发育迟缓或早熟所造成的疾病的发生。
虽然本发明已以实施方式公开如上，然其并非用以限定本发明，任何所属领域的技术人员，在不脱离本发明的精神和范围内，当可作各种的更动与润饰，因此本发明的保护范围当视权利要求所界定的为准。

Claims (17)

  1. 一种骨龄评估与身高预测模型，其特征在于，所述骨龄评估与身高预测模型由以下步骤建立：
    取得参照数据库，其中所述参照数据库包含骨龄身高图谱数据集合及多个参照手骨X光影像数据，其中各所述参照手骨X光影像数据包含生理年龄信息以及性别信息；
    进行影像前处理步骤，其是利用影像数据编辑模块调整各所述参照手骨X光影像数据的影像大小及影像黑白对比度，以取得多个标准化手骨X光影像数据；
    进行特征选取步骤，其是利用特征选取模块分析所述多个标准化手骨X光影像数据后以得至少一个影像特征值；以及
    进行训练步骤，其是将所述至少一个影像特征值通过卷积神经网络学习分类器进行训练而达到收敛，以得所述骨龄评估与身高预测模型，其中所述骨龄评估与身高预测模型是用以判断受试者的手骨发育状态、所述受试者的骨龄以及预测所述受试者的成年身高。
  2. 如权利要求1所述的骨龄评估与身高预测模型，其特征在于，所述卷积神经网络学习分类器为Inception-ResNet-v2卷积神经网络。
  3. 如权利要求1所述的骨龄评估与身高预测模型，其特征在于，所述多个参照手骨X光影像数据的影像格式为数字医疗影像储存标准协定的影像格式。
  4. 如权利要求1所述的骨龄评估与身高预测模型，其特征在于，所述影像前处理步骤还对各所述参照手骨X光影像数据进行影像色度扩展处理。
  5. 如权利要求1所述的骨龄评估与身高预测模型，其特征在于，所述骨龄身高图谱数据集合包含男性骨龄身高图谱数据子集合及女性骨龄身高图谱数据子集合。
  6. 如权利要求1所述的骨龄评估与身高预测模型，其特征在于，各所述参照手骨X光影像数据为非习惯用手的参照手骨X光影像数据。
  7. 一种骨龄评估与身高预测方法，其特征在于，包含：
    提供如权利要求1项的骨龄评估与身高预测模型；
    提供受试者的目标手骨X光影像数据，其中所述目标手骨X光影像数据包含生理年龄信息以及性别信息；
    对所述目标手骨X光影像数据进行前处理，其是利用所述影像数据编辑模块调整所述目标手骨X光影像数据的影像大小及影像黑白对比度，以取得标准化目标手骨X 光影像数据；
    利用所述特征选取模块分析所述标准化目标手骨X光影像数据后以得至少一个影像特征值；以及
    利用所述骨龄评估与身高预测模型分析所述至少一个影像特征值，以判断所述受试者的手骨发育状态、所述受试者的骨龄以及预测所述受试者的成年身高。
  8. 如权利要求7所述的骨龄评估与身高预测方法，其特征在于，所述目标手骨X光影像数据的影像格式为数字医疗影像储存标准协定的影像格式。
  9. 如权利要求7所述的骨龄评估与身高预测方法，其特征在于，所述影像数据编辑模块还对所述目标手骨X光影像数据进行影像色度扩展处理。
  10. 如权利要求7所述的骨龄评估与身高预测方法，其特征在于，所述目标手骨X光影像数据为非习惯用手的目标手骨X光影像数据。
  11. 一种骨龄评估与身高预测系统，其特征在于，包含：
    影像撷取单元，用以取得受试者的目标手骨X光影像数据，其中所述目标手骨X光影像数据包含生理年龄信息以及性别信息；以及
    非瞬态机器可读介质，信号连接所述影像撷取单元，其中所述非瞬态机器可读介质用以储存程序，当所述程序由处理单元执行时是用以评估所述受试者的手骨发育状态、所述受试者的骨龄和预测所述受试者的成年身高，且所述程序包含：
    参照数据库取得模块，用以取得参照数据库，且所述参照数据库包含骨龄身高图谱数据集合及多个参照手骨X光影像数据，其中各所述参照手骨X光影像数据包含生理年龄信息以及性别信息；
    第一影像数据编辑模块，其是调整各所述参照手骨X光影像数据的影像大小及影像黑白对比度，以取得多个标准化手骨X光影像数据；
    特征选取模块，用以分析所述多个标准化手骨X光影像数据后以得至少一个参照影像特征值；
    训练模块，用以将所述至少一个参照影像特征值通过卷积神经网络学习分类器进行训练而达到收敛，以得到骨龄评估与身高预测模型；
    第二影像数据编辑模块，其是调整所述目标手骨X光影像数据的影像大小及影像黑白对比度，以取得标准化目标手骨X光影像数据；
    目标特征选取模块，用以分析所述标准化目标手骨X光影像数据后以得至少一个目标影像特征值；及
    比对模块，用以将所述至少一个目标影像特征值以所述骨龄评估与身高预测模型进行分析，以得到目标影像特征值权重数据，并将所述目标影像特征值权重数据与所述参照数据库进行比对，以输出所述受试者的手骨发育状态判断结果、所述受试者的骨龄判断结果以及所述受试者的成年身高预测结果。
  12. 如权利要求11所述的骨龄评估与身高预测系统，其特征在于，所述卷积神经网络学习分类器为Inception-ResNet-v2卷积神经网络。
  13. 如权利要求11所述的骨龄评估与身高预测系统，其特征在于，所述目标手骨X光影像数据的影像格式为数字医疗影像储存标准协定的影像格式，所述多个参照手骨X光影像数据与的影像格式为数字医疗影像储存标准协定的影像格式。
  14. 如权利要求11所述的骨龄评估与身高预测系统，其特征在于，所述第一影像数据编辑模块还对各所述参照手骨X光影像数据进行一影像色度扩展处理，所述第二影像数据编辑模块还对所述目标手骨X光影像数据进行影像色度扩展处理。
  15. 如权利要求11所述的骨龄评估与身高预测系统，其特征在于，所述骨龄身高图谱数据集合包含男性骨龄身高图谱数据子集合及女性骨龄身高图谱数据子集合。
  16. 如权利要求11所述的骨龄评估与身高预测系统，其特征在于，各所述参照手骨X光影像数据为非习惯用手的参照手骨X光影像数据，所述目标手骨X光影像数据为非习惯用手的目标手骨X光影像数据。
  17. 如权利要求11所述的骨龄评估与身高预测系统，其特征在于，还包含：
    警示模块，用以在所述标准化目标手骨X光影像数据通过所述骨龄评估与身高预测模型进行分析以后，发出主动警示通知。
PCT/CN2018/097915 2018-08-01 2018-08-01 骨龄评估与身高预测模型、其系统及其预测方法 WO2020024127A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
PCT/CN2018/097915 WO2020024127A1 (zh) 2018-08-01 2018-08-01 骨龄评估与身高预测模型、其系统及其预测方法
JP2020528374A JP6999812B2 (ja) 2018-08-01 2018-08-01 骨年齢評価と身長予測モデルの確立方法、そのシステム及びその予測方法
US17/259,685 US20210142477A1 (en) 2018-08-01 2018-08-01 Bone Age Assessment And Height Prediction Model, System Thereof And Prediction Method Thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/097915 WO2020024127A1 (zh) 2018-08-01 2018-08-01 骨龄评估与身高预测模型、其系统及其预测方法

Publications (1)

Publication Number Publication Date
WO2020024127A1 true WO2020024127A1 (zh) 2020-02-06

Family

ID=69230983

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/097915 WO2020024127A1 (zh) 2018-08-01 2018-08-01 骨龄评估与身高预测模型、其系统及其预测方法

Country Status (3)

Country Link
US (1) US20210142477A1 (zh)
JP (1) JP6999812B2 (zh)
WO (1) WO2020024127A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111563874A (zh) * 2020-03-05 2020-08-21 北京深睿博联科技有限责任公司 骨龄评测方法和装置
CN112801994A (zh) * 2021-02-05 2021-05-14 广东顺德工业设计研究院(广东顺德创新设计研究院) 骨龄评估方法和系统
KR20220059244A (ko) * 2020-11-02 2022-05-10 사회복지법인 삼성생명공익재단 심층 학습을 이용한 성인 신장 예측 방법 및 그 시스템
CN117524503A (zh) * 2024-01-08 2024-02-06 深圳市早知道科技有限公司 一种基于生物遗传数据的身高预测方法及系统

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020135812A1 (en) 2018-12-29 2020-07-02 Shanghai United Imaging Intelligence Co., Ltd. Systems and methods for ossification center detection and bone age assessment
TWI772663B (zh) * 2019-08-05 2022-08-01 長佳智能股份有限公司 以個人裝置進行手骨骨齡即時判讀之系統及方法
US20210398280A1 (en) * 2020-06-23 2021-12-23 Vuno Inc. Bone age assessment method for bone image
KR102575890B1 (ko) * 2020-11-16 2023-09-08 본와이즈 주식회사 인공지능 시스템을 훈련시키기 위한 이미지 데이터를 처리하도록 구성되는 전자 장치
CN113298780B (zh) * 2021-05-24 2023-09-12 云南大学 一种基于深度学习的儿童骨龄评估方法及系统
CN113362292B (zh) * 2021-05-27 2023-04-28 重庆邮电大学 一种基于可编程逻辑门阵列的骨龄评估方法及系统
KR102559654B1 (ko) * 2021-08-03 2023-07-24 가톨릭대학교 산학협력단 뇌실-복강 단락술을 위한 압력 산출 방법 및 장치
CN114387680B (zh) * 2022-03-24 2022-05-27 广东红橙云大数据有限公司 评估信息的生成方法、装置、电子设备及介质
CN115661052B (zh) * 2022-10-13 2023-09-12 高峰医疗器械(无锡)有限公司 牙槽骨的骨质检测方法、装置、设备及存储介质
CN116342516B (zh) * 2023-03-17 2024-04-16 四川文理学院 基于模型集成的儿童手骨x光图像骨龄评估方法及系统
CN116523840B (zh) * 2023-03-30 2024-01-16 苏州大学 一种基于深度学习的肺部ct图像检测系统以及方法
CN117094951A (zh) * 2023-07-25 2023-11-21 中国医学科学院北京协和医院 一种新型的自动骨龄预测算法模型

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1804868A (zh) * 2006-01-19 2006-07-19 昆明利普机器视觉工程有限公司 机器自动图像识别方法及装置
WO2017022908A1 (ko) * 2015-08-04 2017-02-09 재단법인 아산사회복지재단 심층신경망을 이용한 골 연령 산출방법 및 프로그램
CN107595248A (zh) * 2017-08-31 2018-01-19 郭淳 一种检测和评价儿童生长发育状况的方法和系统
CN107767376A (zh) * 2017-11-02 2018-03-06 西安邮电大学 基于深度学习的x线片骨龄预测方法及系统
WO2018057714A1 (en) * 2016-09-21 2018-03-29 The General Hospital Corporation Systems, methods and media for automatically generating a bone age assessment from a radiograph
CN107871316A (zh) * 2017-10-19 2018-04-03 浙江工业大学 一种基于深度神经网络的x光片手骨兴趣区域自动提取方法
CN107895367A (zh) * 2017-11-14 2018-04-10 中国科学院深圳先进技术研究院 一种骨龄识别方法、系统及电子设备
CN107944496A (zh) * 2017-12-06 2018-04-20 电子科技大学 基于改进后的残差网络的骨龄自动化识别系统
CN108334899A (zh) * 2018-01-28 2018-07-27 浙江大学 基于手骨x光片骨骼和关节量化信息集成的骨龄评估方法

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003009758A1 (en) * 2001-07-24 2003-02-06 Sunlight Medical, Ltd. Bone age assessment using ultrasound
KR100942699B1 (ko) * 2007-12-10 2010-02-17 한국전자통신연구원 요골원위부 골간단 경계추출 방법 및 장치
CN104780843A (zh) * 2012-09-13 2015-07-15 隆奥克医学技术公司 骨矿物质密度测量设备和方法
US9848818B1 (en) * 2013-08-09 2017-12-26 O.N.Diagnostics, LLC Clinical assessment of fragile bone strength
KR101977174B1 (ko) * 2017-09-13 2019-05-10 이재준 영상 분석 방법, 장치 및 컴퓨터 프로그램
US10825564B1 (en) * 2017-12-11 2020-11-03 State Farm Mutual Automobile Insurance Company Biometric characteristic application using audio/video analysis
DE102019208496B4 (de) * 2019-06-12 2024-01-25 Siemens Healthcare Gmbh Computerimplementierte Verfahren und Vorrichtungen zum Bereitstellen eines Differenzbilddatensatzes eines Untersuchungsvolumens und zum Bereitstellen einer trainierten Generatorfunktion
TWI772663B (zh) * 2019-08-05 2022-08-01 長佳智能股份有限公司 以個人裝置進行手骨骨齡即時判讀之系統及方法
KR102575890B1 (ko) * 2020-11-16 2023-09-08 본와이즈 주식회사 인공지능 시스템을 훈련시키기 위한 이미지 데이터를 처리하도록 구성되는 전자 장치

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1804868A (zh) * 2006-01-19 2006-07-19 昆明利普机器视觉工程有限公司 机器自动图像识别方法及装置
WO2017022908A1 (ko) * 2015-08-04 2017-02-09 재단법인 아산사회복지재단 심층신경망을 이용한 골 연령 산출방법 및 프로그램
WO2018057714A1 (en) * 2016-09-21 2018-03-29 The General Hospital Corporation Systems, methods and media for automatically generating a bone age assessment from a radiograph
CN107595248A (zh) * 2017-08-31 2018-01-19 郭淳 一种检测和评价儿童生长发育状况的方法和系统
CN107871316A (zh) * 2017-10-19 2018-04-03 浙江工业大学 一种基于深度神经网络的x光片手骨兴趣区域自动提取方法
CN107767376A (zh) * 2017-11-02 2018-03-06 西安邮电大学 基于深度学习的x线片骨龄预测方法及系统
CN107895367A (zh) * 2017-11-14 2018-04-10 中国科学院深圳先进技术研究院 一种骨龄识别方法、系统及电子设备
CN107944496A (zh) * 2017-12-06 2018-04-20 电子科技大学 基于改进后的残差网络的骨龄自动化识别系统
CN108334899A (zh) * 2018-01-28 2018-07-27 浙江大学 基于手骨x光片骨骼和关节量化信息集成的骨龄评估方法

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111563874A (zh) * 2020-03-05 2020-08-21 北京深睿博联科技有限责任公司 骨龄评测方法和装置
CN111563874B (zh) * 2020-03-05 2023-04-28 北京深睿博联科技有限责任公司 骨龄评测方法和装置
KR20220059244A (ko) * 2020-11-02 2022-05-10 사회복지법인 삼성생명공익재단 심층 학습을 이용한 성인 신장 예측 방법 및 그 시스템
KR102481112B1 (ko) * 2020-11-02 2022-12-26 사회복지법인 삼성생명공익재단 심층 학습을 이용한 성인 신장 예측 방법 및 그 시스템
KR20230002238A (ko) * 2020-11-02 2023-01-05 사회복지법인 삼성생명공익재단 심층 학습을 이용한 성인 신장 예측 방법 및 그 시스템
KR102633676B1 (ko) * 2020-11-02 2024-02-05 사회복지법인 삼성생명공익재단 심층 학습을 이용한 성인 신장 예측 방법 및 그 시스템
CN112801994A (zh) * 2021-02-05 2021-05-14 广东顺德工业设计研究院(广东顺德创新设计研究院) 骨龄评估方法和系统
CN112801994B (zh) * 2021-02-05 2023-12-26 广东顺德工业设计研究院(广东顺德创新设计研究院) 骨龄评估方法和系统
CN117524503A (zh) * 2024-01-08 2024-02-06 深圳市早知道科技有限公司 一种基于生物遗传数据的身高预测方法及系统
CN117524503B (zh) * 2024-01-08 2024-04-30 深圳市早知道科技有限公司 一种基于生物遗传数据的身高预测方法及系统

Also Published As

Publication number Publication date
JP2021504816A (ja) 2021-02-15
JP6999812B2 (ja) 2022-01-19
US20210142477A1 (en) 2021-05-13

Similar Documents

Publication Publication Date Title
WO2020024127A1 (zh) 骨龄评估与身高预测模型、其系统及其预测方法
TWI684997B (zh) 骨齡評估與身高預測模型之建立方法、骨齡評估與身高預測系統及骨齡評估與身高預測方法
CN107767376B (zh) 基于深度学习的x线片骨龄预测方法及系统
WO2022063199A1 (zh) 一种肺结节自动检测方法、装置及计算机系统
TWI687937B (zh) 染色體異常檢測模型之建立方法、染色體異常檢測系統及染色體異常檢測方法
WO2021139258A1 (zh) 基于图像识别的细胞识别计数方法、装置和计算机设备
Lakshminarayanan et al. Deep Learning-Based Hookworm Detection in Wireless Capsule Endoscopic Image Using AdaBoost Classifier.
CN104484886B (zh) 一种mr图像的分割方法及装置
CN112733950A (zh) 一种基于图像融合与目标检测结合的电力设备故障诊断方法
EP3633682A1 (en) Chromosome abnormality detecting model, detecting system thereof, and method for detecting chromosome abnormality
CN108378869B (zh) 一种从超声图像中自动测量胎儿头围长度的图像处理方法及处理系统
CN109670489B (zh) 基于多实例学习的弱监督式早期老年性黄斑病变分类方法
CN113298780B (zh) 一种基于深度学习的儿童骨龄评估方法及系统
CN108256527A (zh) 一种基于端到端全卷积网络的皮肤病变多类语义分割方法
CN110543912A (zh) 自动获取胎儿关键切面超声视频中心动周期视频的方法
Tang et al. CNN-based qualitative detection of bone mineral density via diagnostic CT slices for osteoporosis screening
CN116579982A (zh) 一种肺炎ct图像分割方法、装置及设备
CN111598868B (zh) 一种肺部超声图像识别方法和系统
CN114240874A (zh) 基于深度卷积神经网络和特征融合的骨龄评定方法、设备及计算机可读存储介质
CN117152433A (zh) 一种基于多尺度跨层注意力融合网络的医学图像分割方法
TWI663960B (zh) 足畸形檢測模型、足畸形檢測系統及足畸形檢測方法
CN116778587A (zh) 一种基于深度学习的活体年龄推断方法、设备、存储介质
CN116229236A (zh) 一种基于改进YOLO v5模型的结核杆菌检测方法
CN115547488A (zh) 基于vgg卷积神经网络和面部识别孤独症早筛系统及方法
CN113469942B (zh) 一种ct图像病变检测方法

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: 18928533

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2020528374

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18928533

Country of ref document: EP

Kind code of ref document: A1