[go: up one dir, main page]

WO2021035712A1 - Méthode de détection rapide de morphologie anormale de structure de tissu osseux et dispositif électronique - Google Patents

Méthode de détection rapide de morphologie anormale de structure de tissu osseux et dispositif électronique Download PDF

Info

Publication number
WO2021035712A1
WO2021035712A1 PCT/CN2019/103787 CN2019103787W WO2021035712A1 WO 2021035712 A1 WO2021035712 A1 WO 2021035712A1 CN 2019103787 W CN2019103787 W CN 2019103787W WO 2021035712 A1 WO2021035712 A1 WO 2021035712A1
Authority
WO
WIPO (PCT)
Prior art keywords
bone tissue
control sample
tested
abnormal
bone
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2019/103787
Other languages
English (en)
Chinese (zh)
Inventor
李晨天
李朝阳
张秋旭
蒋康慧
马驰
吕维加
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Bone's Technology Shenzhen Ltd
Original Assignee
Bone's Technology Shenzhen Ltd
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 Bone's Technology Shenzhen Ltd filed Critical Bone's Technology Shenzhen Ltd
Priority to PCT/CN2019/103787 priority Critical patent/WO2021035712A1/fr
Publication of WO2021035712A1 publication Critical patent/WO2021035712A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis

Definitions

  • the invention relates to a bone tissue structure assessment auxiliary technology, in particular to a rapid detection method and an electronic device for abnormal bone tissue structure morphology.
  • the imaging observation and evaluation of bone tissue structure is an important basis for evaluating the health status of bone tissue in clinical and scientific research.
  • the assessment of bone tissue structure relies on the visual observation and subjective judgment of pathologists and radiologists.
  • the assessment results are prone to errors.
  • the main purpose of the present invention is to provide a rapid detection method and electronic device for abnormal bone tissue structure morphology, so as to solve the problem that the existing evaluation methods of bone tissue structure rely on the visual observation and subjective judgment of pathologists and radiologists. , The evaluation results are prone to errors.
  • a rapid detection method for abnormal bone tissue structure and morphology including:
  • Step A Obtain and preprocess the CT scan images of the bone tissue to be tested and each bone tissue control sample in the control sample group, and segment the region of interest from the obtained preprocessed image to obtain the bone tissue to be tested and The region of interest of each bone tissue control sample in the control sample group;
  • Step B Calculate the feature vector of the bone tissue to be tested and each bone tissue control sample in the control sample group based on the obtained region of interest of each bone tissue control sample in the bone tissue to be tested and the control sample group;
  • Step C Calculate the similarity between the feature vector of the bone tissue to be tested and the standard feature vector of the control sample established based on the feature vector of each bone tissue control sample, and determine whether the structural shape of the bone tissue to be tested is based on the similarity abnormal.
  • the method of preprocessing the CT scan image includes:
  • the spatial resolution of the CT scan image is down-sampled to 1mm ⁇ 1mm ⁇ 1mm isotropic spatial resolution.
  • the region of interest of the bone tissue to be tested and the region of interest of each bone tissue control sample are bone tissues with a complete anatomical structure.
  • the method of calculating the feature vector of the bone tissue to be tested or the bone tissue control sample includes:
  • Step B1 Calculate the average bone density of each region of interest of the bone tissue to be tested or the bone tissue control sample
  • Step B2 Calculate the statistical mean and standard deviation of the average bone density of the region of interest of the bone tissue to be tested or the bone tissue control sample, the calculation method is:
  • the area is divided into three areas: area M1, area M2, and area M3 according to their voxel signal values. Among them:
  • the voxel signal value range of area M1 is: (- ⁇ ,S1], the voxel signal value range of area M2 is (- ⁇ ,S2], and the voxel signal value range of area M3 is (- ⁇ ,+ ⁇ );
  • Step B5 According to the set ⁇ Li1 ⁇ , Li2 ⁇ , Li3 ⁇ , calculate the intra-volume average of each feature in the area M1: Obtain the feature average set ⁇ L1 ⁇ , L2 ⁇ , L3 ⁇ of the bone tissue to be tested or bone tissue control samples, where VM1 is the volume of region M1, ⁇ [(0°,0°),(360°,360° )].
  • step C the method of establishing the standard feature vector of the control sample according to the feature vector of each bone tissue control sample includes:
  • the statistical average and standard deviation of each feature in the feature value set of the bone tissue control sample in the control sample group are calculated by statistics, Obtain the average feature vector of the region of interest of each bone tissue control sample in the control sample group Where ⁇ [(0°,0°),(360°,360°)].
  • step C includes:
  • n is a custom parameter, n ⁇ R+.
  • the method of cosine similarity, k-means, Euclidean distance, svm, linear discriminant analysis, or neural network is used to calculate the similarity between the bone tissue control sample in the control sample group and the standard average feature vector.
  • step C includes:
  • Score D(V(T1 ⁇ , T2 ⁇ , T3 ⁇ )), where V(T1 ⁇ , T2 ⁇ , T3 ⁇ ) is the characteristic vector of the bone tissue to be tested, and Score is the similarity between the characteristics of the bone tissue to be tested and the characteristics of different groups;
  • the first setting value is -2.5
  • the second setting value is -1.
  • a computer-readable storage medium has a computer program stored thereon, and when the computer program is executed by a processor, the rapid detection method for abnormal bone tissue structure morphology as described above is realized.
  • An electronic device comprising a memory, a processor, and a computer program stored in the memory and capable of running in the processor, and when the processor executes the computer program, the bone tissue structure shape as described above is realized Quick detection method of abnormality.
  • the technical solution of the present invention combines the standardization step of bone density calibration and the standardization step of image down-sampling. While standardizing the image, it indirectly improves the signal-to-noise ratio and improves the measurement robustness from multiple levels; Third, the structural morphological characteristics measured by the technical solution of the present invention are correlated with the mechanical stress concentration coefficient, which can indirectly reflect the mechanical influence caused by the structural characteristics.
  • the technical solution of the present invention effectively solves the defect that the existing evaluation methods of bone tissue structure rely on the visual observation and subjective judgment of pathologists and radiologists, and ensures the accuracy of the evaluation results.
  • FIG. 1 is a schematic diagram of the main flow of a rapid detection method for abnormal bone tissue structure and morphology according to an embodiment of the present invention
  • Figure 2 is a schematic diagram of the feature measurement process.
  • Step A Obtain and preprocess the CT scan images of the bone tissue to be tested and each bone tissue control sample in the control sample group, and segment the region of interest from the obtained preprocessed image to obtain the bone tissue to be tested and The region of interest of each bone tissue control sample in the control sample group.
  • Step B Calculate the feature vector of the bone tissue to be tested and each bone tissue control sample in the control sample group based on the obtained region of interest of each bone tissue control sample in the bone tissue to be tested and the control sample group.
  • Step C Calculate the similarity between the feature vector of the bone tissue to be tested and the standard feature vector of the control sample established based on the feature vector of each bone tissue control sample, and determine whether the structure and shape of the bone tissue to be tested is abnormal according to the similarity.
  • the CT scan image of the bone tissue to be tested and each bone tissue control sample in the control sample group can be obtained by the CT scanning device.
  • the bone tissue to be tested and each bone tissue control sample in the control sample group can be respectively placed in a CT scanner for scanning, and the scanning parameters are selected as: 120kV, 300mAs.
  • reconstruction nucleus selection for example, the reconstruction nucleus of Siemens Somaton series scanners is B80s or B30f, or the equivalent reconstruction nucleus of other models of brand scanners.
  • the xy-axis reconstruction resolution is: 0.5mm ⁇ 0.5mm, and the reconstruction layer thickness is 0.625. mm.
  • step A the method for preprocessing the CT scan image includes:
  • the spatial resolution of the CT scan image is down-sampled to 1mm ⁇ 1mm ⁇ 1mm isotropic spatial resolution.
  • step A when segmenting the region of interest, the image of the part of interest can be marked and stored, and the background image therein can be removed.
  • the region of interest of the bone tissue to be tested and the region of interest of each bone tissue control sample are bone tissues with a complete anatomical structure, such as a complete vertebral body part, a complete femoral head part, and so on.
  • Step B mainly includes T-score/Z-score conversion, T-score/Z-score gray gradient threshold segmentation, anisotropic feature calculation, and comparison sample feature coefficient calculation.
  • the method for calculating the feature vector of the bone tissue to be tested or the bone tissue control sample includes:
  • Step B1 Calculate the average bone density of each region of interest of the bone tissue to be tested or the bone tissue control sample, the calculation method is: From the above formula, the average bone density of each control sample area of interest can be obtained.
  • Step B2 Calculate the statistical mean and standard deviation of the average bone density of the region of interest of the bone tissue to be tested or the bone tissue control sample, the calculation method is:
  • the area is divided into three areas: area M1, area M2, and area M3 according to their voxel signal values, as shown in FIG. 2 for details. among them:
  • the voxel signal value range of area M1 is: (- ⁇ ,S1], the voxel signal value range of area M2 is (- ⁇ ,S2], and the voxel signal value range of area M3 is (- ⁇ ,+ ⁇ ).
  • ⁇ [(0°,0°),(360°,360°)];
  • Step B5 According to the set ⁇ Li1 ⁇ , Li2 ⁇ , Li3 ⁇ , calculate the intra-volume average of each feature in the area M1: Obtain the feature average set ⁇ L1 ⁇ , L2 ⁇ , L3 ⁇ of the bone tissue to be tested or bone tissue control samples, where VM1 is the volume of region M1, ⁇ [(0°,0°),(360°,360° )].
  • the feature vector of the bone tissue to be tested and the feature vector of each bone tissue control sample in the control sample group are calculated in the same way. After the feature vector of each bone tissue control sample in the control sample group is calculated, the control sample needs to be established accordingly Standard feature vector, so as to compare the standard feature vector of the reference sample with the feature vector of the bone tissue to be tested for similarity.
  • the method of establishing the standard feature vector of the control sample according to the feature vector of each bone tissue control sample includes:
  • the statistical average and standard deviation of each feature in the feature value set of the bone tissue control sample in the control sample group are calculated by statistics, Obtain the average feature vector of the region of interest of each bone tissue control sample in the control sample group Where ⁇ [(0°,0°),(360°,360°)].
  • step C can be performed to compare the feature vector of the bone tissue to be tested with the standard feature vector of the control sample for similarity.
  • step C has two implementation manners. In the first implementation manner, step C includes:
  • n is a custom parameter, n ⁇ R+.
  • the cosine similarity method can be used to calculate the similarity between the bone tissue control sample in the control sample group and the standard average feature vector:
  • V(T1 ⁇ , T2 ⁇ , T3 ⁇ ) is the feature vector of the bone tissue to be tested
  • Sim is the degree of similarity between the characteristics of the bone tissue to be tested and the characteristics of the standard product.
  • the cosine similarity method can also be calculated by k-means, Euclidean distance, svm, linear discriminant analysis, or neural network.
  • step C specifically includes:
  • Score D(V(T1 ⁇ , T2 ⁇ , T3 ⁇ )), where V(T1 ⁇ , T2 ⁇ , T3 ⁇ ) is the characteristic vector of the bone tissue to be tested, and Score is the similarity between the characteristics of the bone tissue to be tested and the characteristics of different groups;
  • the first setting value is -2.5
  • the second setting value is -1.
  • the features participating in the cosine similarity calculation can be any combination of the three groups, that is, any single one of the three groups, or a combination of two or two, or a complete set of three features.
  • the function conversion processing here can be performed separately for multiple sets of features or for a single set of features.
  • the purpose of the above-mentioned function conversion processing is to convert the obtained original features into a specific statistical distribution type (such as normal distribution, Poisson distribution) to obtain a better statistical processing effect.
  • the present invention also provides a computer-readable storage medium and electronic device.
  • a computer program is stored on the computer readable storage medium, and when the computer program is executed by the processor, the above rapid detection method for abnormal bone tissue structure morphology is realized.
  • the computer storage medium may also be a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a RAM, a magnetic disk, or an optical disk, etc., which can store program codes.
  • the electronic device includes a memory, a processor, and a computer program that is stored in the memory and can be run in the processor. When the processor executes the computer program, the above rapid detection method for abnormal bone tissue structure morphology is realized.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Apparatus For Radiation Diagnosis (AREA)

Abstract

L'invention concerne une méthode de détection rapide d'une morphologie anormale d'une structure de tissu osseux. Une méthode d'extraction de caractéristiques mise en œuvre pendant le processus de détection de la morphologie de la structure de tissu osseux et des caractéristiques de mesure utilisées présentent une invariance de rotation et une invariance de translation, et l'analyse peut être effectuée rapidement sans enregistrement rigide strict ni autres opérations lors de la collecte d'images de détection. Simultanément, la méthode de détection combine une étape de normalisation d'étalonnage de densité osseuse et une étape de normalisation de sous-échantillonnage d'image. Pendant la normalisation des images, le rapport signal/bruit est indirectement amélioré et la robustesse de mesure est améliorée à plusieurs niveaux. En outre, les caractéristiques morphologiques de la structure mesurée par la méthode sont corrélées avec un coefficient de concentration de contrainte mécanique, ce qui peut refléter indirectement l'influence mécanique provoquée par les caractéristiques de la structure. Grâce à la méthode, le défaut que représente le fait que le moyen d'évaluation existant d'une structure de tissu osseux repose sur l'observation visuelle et la détermination subjective de pathologistes et de radiologues est résolu efficacement, ce qui permet d'assurer la précision des résultats d'évaluation. L'invention concerne en outre un dispositif électronique correspondant.
PCT/CN2019/103787 2019-08-30 2019-08-30 Méthode de détection rapide de morphologie anormale de structure de tissu osseux et dispositif électronique Ceased WO2021035712A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/103787 WO2021035712A1 (fr) 2019-08-30 2019-08-30 Méthode de détection rapide de morphologie anormale de structure de tissu osseux et dispositif électronique

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/103787 WO2021035712A1 (fr) 2019-08-30 2019-08-30 Méthode de détection rapide de morphologie anormale de structure de tissu osseux et dispositif électronique

Publications (1)

Publication Number Publication Date
WO2021035712A1 true WO2021035712A1 (fr) 2021-03-04

Family

ID=74684468

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/103787 Ceased WO2021035712A1 (fr) 2019-08-30 2019-08-30 Méthode de détection rapide de morphologie anormale de structure de tissu osseux et dispositif électronique

Country Status (1)

Country Link
WO (1) WO2021035712A1 (fr)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2002038045A2 (fr) * 2000-11-08 2002-05-16 The Johns Hopkins University Techniques de derivation de structure tissulaire a partir d'absorptiometrie a rayons x bienergique de projection multiple
CN1849094A (zh) * 2003-03-25 2006-10-18 成像治疗仪股份有限公司 在本发明的射线照相影像领域的处理中用于补偿成像技术的方法
US20110213242A1 (en) * 2010-02-18 2011-09-01 Budoff Matthew J Method for thoracic vertebral bone density measurement by thoracic quantitative computed tomography
CN102648482A (zh) * 2009-09-11 2012-08-22 斯特拉克斯联合控股有限公司 用于所选组织结构的图像分析的方法及系统
CN108292432A (zh) * 2015-11-13 2018-07-17 奥卢大学 用于确定宏拓扑、毫拓扑、微拓扑和纳米拓扑中的至少一个的评估系统和方法
CN108498075A (zh) * 2017-02-24 2018-09-07 西门子保健有限责任公司 骨骼健康的个性化评估
CN108830835A (zh) * 2018-05-25 2018-11-16 北京长木谷医疗科技有限公司 识别脊柱矢状位图像异常的方法及计算设备

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2002038045A2 (fr) * 2000-11-08 2002-05-16 The Johns Hopkins University Techniques de derivation de structure tissulaire a partir d'absorptiometrie a rayons x bienergique de projection multiple
CN1849094A (zh) * 2003-03-25 2006-10-18 成像治疗仪股份有限公司 在本发明的射线照相影像领域的处理中用于补偿成像技术的方法
CN102648482A (zh) * 2009-09-11 2012-08-22 斯特拉克斯联合控股有限公司 用于所选组织结构的图像分析的方法及系统
US20110213242A1 (en) * 2010-02-18 2011-09-01 Budoff Matthew J Method for thoracic vertebral bone density measurement by thoracic quantitative computed tomography
CN108292432A (zh) * 2015-11-13 2018-07-17 奥卢大学 用于确定宏拓扑、毫拓扑、微拓扑和纳米拓扑中的至少一个的评估系统和方法
CN108498075A (zh) * 2017-02-24 2018-09-07 西门子保健有限责任公司 骨骼健康的个性化评估
CN108830835A (zh) * 2018-05-25 2018-11-16 北京长木谷医疗科技有限公司 识别脊柱矢状位图像异常的方法及计算设备

Similar Documents

Publication Publication Date Title
CN110415234A (zh) 基于多参数磁共振成像的脑部肿瘤分割方法
CN111340827A (zh) 一种肺部ct影像数据处理分析方法及系统
CN111932492B (zh) 一种医学图像处理方法、装置及计算机可读存储介质
CN108062749B (zh) 肛提肌裂孔的识别方法、装置和电子设备
CN111260671A (zh) 一种ct影像肺部分叶方法及系统
CN110660044B (zh) 骨组织结构形态异常的快速检测方法及电子装置
US20230115927A1 (en) Systems and methods for plaque identification, plaque composition analysis, and plaque stability detection
CN115760858B (zh) 基于深度学习的肾脏病理切片细胞识别方法及系统
CN111652300B (zh) 脊柱曲度分类方法、计算机设备和存储介质
CN111951215A (zh) 一种图像检测方法、装置及计算机可读存储介质
CN114881976B (zh) 股骨肱骨标准切面图像确定方法、装置和计算机设备
CN117745704A (zh) 一种用于骨质疏松识别的脊椎区域分割系统
CN106778793B (zh) 一种影像特征的可重复性测量方法及装置
CN113610923B (zh) 扫描定位方法、装置、计算机设备及计算机可读存储介质
WO2018098697A1 (fr) Procédé et dispositif de mesure de répétabilité de caractéristique d'image
CN118039142A (zh) 一种基于舌象的阿尔兹海默症预警方法及装置
CN112102327A (zh) 一种图像处理方法、装置及计算机可读存储介质
CN115760934B (zh) 一种眼鼻三角区的自动标定方法
CN113012127A (zh) 基于胸部医学影像的心胸比测量方法
CN112734740A (zh) 训练目标检测模型的方法、目标检测的方法及其装置
WO2021035712A1 (fr) Méthode de détection rapide de morphologie anormale de structure de tissu osseux et dispositif électronique
CN118864622A (zh) 冠状动脉钙化积分评估模型训练方法、评估方法及系统
CN115761373A (zh) 一种基于核磁共振图像的乳腺检测分类方法
CN109993754B (zh) 从图像上进行颅骨分割的方法和系统
WO2023133929A1 (fr) Procédé de détection et d'analyse de symétrie de tissu humain basé sur des ultrasons

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: 19943542

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: 19943542

Country of ref document: EP

Kind code of ref document: A1