WO2020237525A1 - Système et procédé de détection pour région de lésion d'imagerie de perfusion d'accident vasculaire cérébral sur la base de coefficients de corrélation - Google Patents

Système et procédé de détection pour région de lésion d'imagerie de perfusion d'accident vasculaire cérébral sur la base de coefficients de corrélation Download PDF

Info

Publication number
WO2020237525A1
WO2020237525A1 PCT/CN2019/088994 CN2019088994W WO2020237525A1 WO 2020237525 A1 WO2020237525 A1 WO 2020237525A1 CN 2019088994 W CN2019088994 W CN 2019088994W WO 2020237525 A1 WO2020237525 A1 WO 2020237525A1
Authority
WO
WIPO (PCT)
Prior art keywords
correlation coefficient
image
information
perfusion
voxel point
Prior art date
Application number
PCT/CN2019/088994
Other languages
English (en)
Chinese (zh)
Inventor
朱帆
尚明生
陈琳
Original Assignee
中国科学院重庆绿色智能技术研究院
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国科学院重庆绿色智能技术研究院 filed Critical 中国科学院重庆绿色智能技术研究院
Priority to PCT/CN2019/088994 priority Critical patent/WO2020237525A1/fr
Publication of WO2020237525A1 publication Critical patent/WO2020237525A1/fr

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/05—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves 
    • A61B5/055—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves  involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
    • 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
    • A61B6/02—Arrangements for diagnosis sequentially in different planes; Stereoscopic radiation diagnosis
    • A61B6/03—Computed tomography [CT]

Definitions

  • the invention relates to a medical detection device and a detection and imaging method of the device, in particular to a stroke perfusion imaging lesion area detection system and method based on correlation coefficient.
  • Brain Perfusion Imaging plays an important role in disease diagnosis, disease stage classification and treatment guidance. It is an important method for the diagnosis of cerebral thrombosis, epilepsy, dementia, stroke (stroke) and other diseases. Cerebral perfusion imaging can usually provide brain blood flow, blood volume, average transit time and peak time and other parameter maps. In the case of cerebral blood flow rate (CBF), cerebral blood volume (CBV), mean transit time (MTT) and other blood When the kinetic parameters are quantified, the concentration change curve of the tracer in the artery—arterial input function (AIF) is usually used.
  • CBF cerebral blood flow rate
  • CBV cerebral blood volume
  • MTT mean transit time
  • AIF arterial input function
  • a Chinese invention patent application discloses a method and device for segmentation of an ischemic stroke image.
  • the method includes: preprocessing the first intracranial computed tomography CTP timing diagram to obtain the preprocessed intracranial CTP timing diagram; inputting the preprocessed intracranial CTP timing diagram into the first network for processing to obtain The first MRI image corresponding to the preprocessed intracranial CTP time sequence diagram; the first MRI image is input into the second network for processing to obtain the regional segmentation image of ischemic stroke, wherein Perform a convolution operation on an intracranial computed tomography perfusion imaging sequence diagram to determine the arterial input function (AIF) in the first intracranial computed tomography perfusion imaging sequence diagram; according to the arterial input function and the first intracranial computer A tomographic perfusion imaging sequence diagram, determining a starting time, where the starting time is the time when the arterial input function appears in the first intracranial computed tomography perfusion imaging sequence diagram;
  • AIF arterial input function
  • This method can automatically find the AIF in the CTP sequence diagram through the neural network, and remove the invalid data in the CTP sequence diagram according to the position of the AIF. Then deconvolve the CTP timing chart after removing the invalid data to obtain the preprocessed CTP timing chart, and perform convolution operation on the preprocessed CTP timing chart to generate an NMR image, and then compare the NMR image Perform convolution operation to automatically and efficiently divide the CTP timing diagram into cerebral infarction area, penumbra area and background area.
  • the Chinese Invention Patent discloses a perfusion magnetic resonance imaging method of a vascularized animal subject's region of interest.
  • the method includes: administering an imaging to the vasculature of the subject Agent medicament; determining the region of interest over a series of time values (t) from before the contrast agent reaches the region of interest to at least the end of the first pass of the contrast agent through the region of interest
  • the magnetic resonance signal intensity s i (t) of voxel (i) from the determined value s i (t) of the signal intensity and an arterial input function v(t), the tissue residue is determined for each voxel function value r i (t) of;
  • the image region of interest generated by the value r i (t) is determined; the improvement comprising generating a s i (t) voxel specific arterial function v i ( t) and use the voxel specific arterial function to determine the value of the tissue residual function
  • This method determines the voxel specific value of v i (t) from the magnetic resonance image intensity signal s i (t), the value of r i (t) with greater clinical information and therefore the value of the above-mentioned regional parameters can be It is determined. Further, by v i (t) expressed as a function of time, you can evaluate the blood supply to the organ mode.
  • the purpose of the present invention is to overcome the above-mentioned shortcomings in the prior art, and to provide a system and method for detecting the lesion area of stroke perfusion imaging based on correlation coefficient.
  • a method for detecting lesion area of stroke perfusion imaging based on correlation coefficient including the following steps:
  • each 3D voxel point in the 3D image is translated backward, and then a translation time t with the highest Pearson correlation coefficient is found.
  • a standardized AIF function is used.
  • the AIF function in the brainstem of healthy adults is usually used.
  • a personalized AIF function is used. That is, the AIF in the brainstem of the person to be diagnosed (patient) is used.
  • different data sets are automatically identified. That is, different patients (data sets) can be automatically diagnosed without additional adaptation.
  • step A Gaussian process regression and noise reduction processing based on the voxel point intensity curve are performed on the cerebral perfusion image. Because in the subsequent calculation of the correlation coefficient, it is also based on the voxel point intensity curve, not based on the image.
  • the Pearson correlation coefficient r is as follows:
  • r is the Pearson correlation coefficient
  • cov(X,Y) is the covariance of X, Y, X, Y is the voxel group
  • E[X] is the expected value of X
  • E[Y] is the expected value of Y.
  • the covariance can be simplified as:
  • n is the number of time points
  • x i is the value of the i-th element (voxel) in X
  • y i is the value of the i-th element (voxel) in group Y
  • i is a natural number.
  • the Pearson correlation coefficient r is
  • the Spearsman correlation coefficient ⁇ in the method of the present invention is as follows,
  • [rho] is the correlation coefficient Steven Mann
  • the value of ⁇ or r is generally from -1 to 1.
  • the measured tissue will not be negatively correlated with the reference value (the value of healthy tissue), so the value of ⁇ or r should be positive.
  • Spearsman's correlation coefficient or Pearson's correlation coefficient is equal to 1, it means that the two are perfectly correlated.
  • the Spearsman correlation coefficient or Pearson correlation coefficient is between 0 and 1, it means the health of the tissue. The closer the distance is to 0, the greater the degree of tissue damage. This characteristic can be expressed by the thermal map of the correlation coefficient.
  • step D a one-tailed test is performed on the Pearson correlation coefficient or the Spearsman correlation coefficient.
  • the present invention also includes an image processing system, characterized in that the system includes one or more processors and a storage device, and the storage device is used to store one or more programs, and when the one or more programs are One or more processors execute, so that the one or more processors implement the method as described above.
  • the present invention also provides a stroke perfusion imaging lesion area detection device based on correlation coefficient, which includes:
  • Magnetic resonance perfusion or CT perfusion imaging instrument which is used to obtain image information of contrast agent perfusion in the forebrain, the image information includes three-dimensional information and timing information;
  • the image processing device is used for backward calculation of the time sequence information of each voxel point in the three-dimensional image to obtain the highest flow rate (CBF), cumulative flow rate (CBV), and highest flow delay time ( MTT) information and obtain the AIF function; for each 3D voxel point, calculate the Pearson correlation coefficient or Spearman's correlation coefficient (Person's correlation coefficient or Spearman's correlation coefficient) of the time series signal curve and the arterial input function, and add The result is drawn as a heat map.
  • CBF flow rate
  • CBV cumulative flow rate
  • MTT flow delay time
  • the present invention has beneficial effects: the method of the present invention uses correlation coefficients to correlate the time-concentration curve shape of the tissue, rather than the intensity value. This feature can reduce noise and intensity values in different scanned images. The difference caused by misjudgment and possible damage to the tissue. Since the time-concentration curves of arteries, gray matter and white matter are expected to have similar or identical shapes (with different amplitudes), it is not necessary to divide the types of tissues during processing, but directly use the curve shapes for comparison.
  • Figure 1 is a flow chart of the method of the present invention.
  • Figure 2 is a heat map of the correlation coefficient between Pearson and Spearsman without denoising the source data.
  • Figure 3 is a heat diagram of the correlation coefficients of Pearson and Spearsman after denoising the source data.
  • Figure 4 is a heat map of Pearson's correlation coefficient and a heat map of Spearsman's correlation coefficient.
  • Figure 5 is a schematic diagram after single-tail detection.
  • a specific embodiment of the present invention includes the following steps:
  • the object to be detected can be put into an MRI imager, and a contrast agent is injected into the object to be detected. After the contrast agent is injected, the brain of the object to be detected is imaged to observe the process of the contrast agent passing through the brain and generate 4D (3D + time series) image information.
  • the target to be detected may be a human or other living beings.
  • the image can be derived from the detection equipment, or can be an image stored in a computer storage medium.
  • Brain perfusion images usually contain more information than hemodynamic parameter maps, which can be more conducive to analysis and understanding.
  • Identify and analyze the cerebral perfusion image There are a variety of image recognition technologies in the prior art, which will not be described in detail here.
  • image recognition technologies in the prior art, which will not be described in detail here.
  • different data sets In brain perfusion images, different data sets have different waiting time lengths and different signal strengths before the contrast agent is injected. Therefore, different data sets need to be trained separately, because their signal curve characteristics, whether it is healthy or abnormal tissue types, will be different.
  • unsupervised learning can be used, that is, no manual indexing is required.
  • the purpose of preprocessing is mainly to reduce noise.
  • a preferred method in this embodiment is to use Gaussian process regression and noise reduction processing based on the voxel point intensity curve. Those skilled in the art can use other noise reduction methods known to them. After noise reduction or before noise reduction, the first and last images can be removed, making the operation noise lower.
  • the information of each voxel point with time sequence in the three-dimensional image is calculated backwards to obtain the highest flow rate (CBF), cumulative flow rate (CBV), and maximum flow delay (MTT) information of each voxel point
  • each 3D voxel point in the 3D image it can be translated backward, and then a translation time t with the highest Pearson correlation coefficient can be found.
  • Select AIF In the process of selecting AIF, you can either select a. Use the standardized AIF function, that is, use the AIF in the brainstem of any healthy adult, or use the personalized AIF function, that is, use the brainstem of the patient to be diagnosed AIF.
  • the obtained AIF function is a signal with timing.
  • the method of obtaining AIF is simple. You can choose any location on the brainstem and directly evaluate the volume of the contrast agent in the brainstem.
  • n is the number of time points
  • x i is the value of the i-th element (voxel) in X
  • y i is the value of the i-th element (voxel) in group Y
  • i is a natural number.
  • the Spearsman correlation coefficient ⁇ is as follows,
  • [rho] is the correlation coefficient Steven Mann
  • the value of ⁇ or r is generally from -1 to 1.
  • the measured tissue is positively correlated with the reference value (the value of healthy tissue), so the value of ⁇ or r should be from 0 to 1.
  • the reference value the value of healthy tissue
  • the Spearsman's correlation coefficient or Pearson's correlation coefficient is equal to 1, it means that the two are perfectly correlated.
  • the Spearsman correlation coefficient or Pearson correlation coefficient is between 0 and 1, it means the health of the tissue. The closer the distance is to 0, the greater the degree of tissue damage. This characteristic can be expressed by the thermal map of the correlation coefficient.
  • an existing brain perfusion image is selected for analysis, and when the image is collected, the time from start to imaging is 1 hour and 54 minutes.
  • AIF arterial input function
  • the existing AIF selection method is used, and a healthy AIF curve is selected. Since the correlation coefficient between healthy arteries and gray (white) tissue is relatively large (>0.9), Therefore, only one reference curve is used for all tissue types.
  • the heat maps of the correlation coefficients of Pearson and Spearsman are obtained by processing according to the method of the foregoing embodiment.
  • Figure 2 is a heat map of the correlation coefficient between Pearson and Spearsman without denoising the source data (the left side is the heat map of the Pearson correlation coefficient, and the right side is the heat map of the Spearsman correlation coefficient)
  • Figure 3 is a heat map of the correlation coefficient between Pearson and Spearsman after denoising the source data (the left side is the heat map of the Pearson correlation coefficient, and the right side is the heat map of the Spearsman correlation coefficient). It can be clearly seen that after Gaussian noise reduction, the related heat map lesion area is more obvious.
  • the left side is the heat map of Pearson's correlation coefficient
  • the right side is the heat map of Spearsman's correlation coefficient.
  • Area 1 is the penumbra area, which is displayed in green or blue in the heat map. It means that the correlation coefficient is small and has a higher risk of disease; when the tissue actually dies, it has a false positive error, and its time concentration curve coincidentally has a shape similar to the health curve, that is, in the area where the correlation coefficient is almost 0, that is Area 2 in the figure; when the tissue is actually healthy, but the relevant tests cannot detect this, it is a false negative error because the tissue is affected by noise and low CNR, that is, area 3 in the figure appears as Some dark spots.
  • one-tailed detection can also be performed on each correlation coefficient, that is, the output is only divided into healthy tissues and non-healthy tissues according to a set threshold.
  • the statistical methods and threshold selection of one-tailed detection are common technical solutions in statistics. For example, if the detection value is lower than the selected statistical significance threshold (in our case, 0.05 and 0.01), then the null hypothesis is rejected and we Consider the target organization as an abnormal organization.
  • Fig. 5 is a schematic diagram after adopting one-tailed detection (wherein, the left side is the graph of Pearson's correlation coefficient, and the right side is the graph of Spearsman's correlation coefficient). After single-tail detection is used, the diseased area and the normal area will be more clearly distinguished, but the risk level cannot be seen from it, which can be effectively applied in certain specific embodiments.
  • the aforementioned calculations can be calculated on Intel's Xeon series processor servers, which are clocked at 3G. When processing a data set of 512 ⁇ 512 ⁇ 2 with 36 time intervals, it takes about 10 seconds to use Gaussian process for preprocessing to reduce noise. Pearson correlation coefficient calculation only needs to run for one second. Spearman's correlation coefficient calculation takes 11 seconds, because its sorting requires an additional 10 seconds.
  • the perfusion source image can be accurately analyzed in a reasonable time without expert intervention in the image processing process.
  • the system includes one or more processors and a storage device.
  • the storage device is used to store one or more programs.
  • the one or more programs are One or more processors execute, so that the one or more processors implement the method as described above.
  • a device for detecting lesion area of stroke perfusion imaging based on correlation coefficient includes: magnetic resonance perfusion or CT perfusion imaging instrument, which is used to obtain image information of contrast agent perfusion in the forebral hilum.
  • the information includes three-dimensional information and time series information; the image processing device is used for the backward calculation of the time series information of each voxel point in the three-dimensional image to obtain the highest flow rate (CBF) of each voxel point, and the cumulative flow rate ( CBV), maximum flow delay (MTT) information, and obtain the AIF function; for each 3D voxel point, calculate the curve of its time series signal and the Pearson correlation coefficient or Spearsman correlation coefficient of the arterial input function, and compare the results Draw into a heat map.
  • CBF flow rate
  • CBV cumulative flow rate
  • MTT maximum flow delay

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Medical Informatics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Physics & Mathematics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Surgery (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Radiology & Medical Imaging (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Molecular Biology (AREA)
  • Pathology (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Optics & Photonics (AREA)
  • Magnetic Resonance Imaging Apparatus (AREA)

Abstract

L'invention concerne un procédé et un système de détection pour une région de lésion d'imagerie de perfusion d'accident vasculaire cérébral (AVC) sur la base de coefficients de corrélation. Le procédé comprend les étapes suivantes : A) lire une image de perfusion cérébrale, cette dernière comportant des informations d'image tridimensionnelle et des informations de synchronisation ; B) exécuter une résolution régressive sur des informations de chaque synchronisation de bande de points voxel de l'image tridimensionnelle pour obtenir le flux le plus élevé, le flux accumulé et les informations de retard de flux les plus élevés de chaque point voxel ; C) obtenir une fonction d'entrée artérielle ; D) pour chaque point voxel tridimensionnel de l'image tridimensionnelle, calculer sa courbe de signaux de synchronisation et un coefficient de corrélation de Pearson ou un coefficient de corrélation de Spearman de la fonction d'entrée artérielle, et permettre aux résultats de générer un graphique thermodynamique. Le procédé peut permettre de réduire les fausses détections provoquées par le bruit et la différence de valeurs d'intensité des différentes images balayées, et réduire les lésions possibles d'un tissu.
PCT/CN2019/088994 2019-05-29 2019-05-29 Système et procédé de détection pour région de lésion d'imagerie de perfusion d'accident vasculaire cérébral sur la base de coefficients de corrélation WO2020237525A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/088994 WO2020237525A1 (fr) 2019-05-29 2019-05-29 Système et procédé de détection pour région de lésion d'imagerie de perfusion d'accident vasculaire cérébral sur la base de coefficients de corrélation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/088994 WO2020237525A1 (fr) 2019-05-29 2019-05-29 Système et procédé de détection pour région de lésion d'imagerie de perfusion d'accident vasculaire cérébral sur la base de coefficients de corrélation

Publications (1)

Publication Number Publication Date
WO2020237525A1 true WO2020237525A1 (fr) 2020-12-03

Family

ID=73552450

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/088994 WO2020237525A1 (fr) 2019-05-29 2019-05-29 Système et procédé de détection pour région de lésion d'imagerie de perfusion d'accident vasculaire cérébral sur la base de coefficients de corrélation

Country Status (1)

Country Link
WO (1) WO2020237525A1 (fr)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000057777A1 (fr) * 1999-03-26 2000-10-05 Oestergaard Leif Calcul de coefficients d'hemodynamisme a partir de donnees tomographiques
US7069068B1 (en) * 1999-03-26 2006-06-27 Oestergaard Leif Method for determining haemodynamic indices by use of tomographic data
WO2008024082A2 (fr) * 2006-08-24 2008-02-28 Agency For Science, Technology And Research Localisation de points de repère du cerveau tels que les commissures antérieure et postérieure en fonction d'une adaptation géométrique
CN103339653A (zh) * 2010-12-17 2013-10-02 奥胡斯大学 勾画组织损伤的方法
US20130322718A1 (en) * 2012-06-01 2013-12-05 Yi-Hsuan Kao Method and apparatus for measurements of the brain perfusion in dynamic contrast-enhanced computed tomography images
CN105997128A (zh) * 2016-08-03 2016-10-12 上海联影医疗科技有限公司 利用灌注成像识别病灶的方法及系统
WO2017112554A1 (fr) * 2015-12-21 2017-06-29 The Regents Of The University Of California Angiographie de perfusion par soustraction numérique
CN109662735A (zh) * 2019-02-18 2019-04-23 亿慈(上海)智能科技有限公司 皮肤血流灌注量的测量方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000057777A1 (fr) * 1999-03-26 2000-10-05 Oestergaard Leif Calcul de coefficients d'hemodynamisme a partir de donnees tomographiques
US7069068B1 (en) * 1999-03-26 2006-06-27 Oestergaard Leif Method for determining haemodynamic indices by use of tomographic data
WO2008024082A2 (fr) * 2006-08-24 2008-02-28 Agency For Science, Technology And Research Localisation de points de repère du cerveau tels que les commissures antérieure et postérieure en fonction d'une adaptation géométrique
CN103339653A (zh) * 2010-12-17 2013-10-02 奥胡斯大学 勾画组织损伤的方法
US20130322718A1 (en) * 2012-06-01 2013-12-05 Yi-Hsuan Kao Method and apparatus for measurements of the brain perfusion in dynamic contrast-enhanced computed tomography images
WO2017112554A1 (fr) * 2015-12-21 2017-06-29 The Regents Of The University Of California Angiographie de perfusion par soustraction numérique
CN105997128A (zh) * 2016-08-03 2016-10-12 上海联影医疗科技有限公司 利用灌注成像识别病灶的方法及系统
CN109662735A (zh) * 2019-02-18 2019-04-23 亿慈(上海)智能科技有限公司 皮肤血流灌注量的测量方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
FAN, ZHU ET AL.: "Lesion Area Detection Using Source Image Correlation Coefficient for CT Perfusion Imaging", IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, vol. 17, no. 5, 30 September 2013 (2013-09-30), XP011525477, ISSN: 2168-2194, DOI: 20200218155040X *

Similar Documents

Publication Publication Date Title
AU2013400936B2 (en) Image analysis techniques for diagnosing diseases
Delsanto et al. Characterization of a completely user-independent algorithm for carotid artery segmentation in 2-D ultrasound images
US9036878B2 (en) Method for delineation of tissue lesions
WO2012074039A1 (fr) Dispositif de traitement d'image médicale
CN110236544B (zh) 基于相关系数的中风灌注成像病变区域检测系统及方法
CN111753833A (zh) 基于fMRI与DTI构建脑网络建模的帕金森辅助识别方法
US9767552B2 (en) Systems and methods for detecting the presence of anomalous material within tissue
CN113288186A (zh) 基于深度学习算法的乳腺肿瘤组织检测方法及装置
CN114332910A (zh) 一种面向远红外图像的相似特征计算的人体部位分割方法
US20230143229A1 (en) Method for diagnostic ultrasound of carotid artery
Charoensuk et al. Acute stroke brain infarct segmentation in DWI images
WO2020237525A1 (fr) Système et procédé de détection pour région de lésion d'imagerie de perfusion d'accident vasculaire cérébral sur la base de coefficients de corrélation
Lauersen et al. Kidney segmentation for quantitative analysis applying MaskRCNN architecture
KR20190068254A (ko) 병변 발생 시점 추정 방법, 장치 및 프로그램
US20220280121A1 (en) Non-invasive non-contact system and method for evaluating primary and secondary hypertension conditions using thermal imaging
CN114847922A (zh) 一种基于纤维束自动识别的脑年龄预测方法
KR20210024853A (ko) 영상 진단기기를 이용한 특발성 정상압 수두증의 진단 방법 및 시스템
Lee et al. Automatic Detection of Dense Calcium and Acoustic Shadow in Intravascular Ultrasound Images by Dual-threshold-based Segmentation Approach.
Wang et al. Spatio-temporal analysis of brain MRI images using hidden Markov models
Wang et al. Voxel-based discriminant map classification on brain ventricles for Alzheimer's disease
CN117649403B (zh) 用于偏头痛的视频衍生面部血流脉动定量评测方法、装置及设备
KR102487827B1 (ko) 전두엽 피부 영상을 이용한 치매 중증도 분석 방법 및 장치
US20240013379A1 (en) Computer-implemented method and system for determining the fetus ventricular volume from diffusion-weighted magnetic resonance imaging, and related nmr ventricle volume assessment method
Soman et al. A novel algorithm for earlier down syndrome detection in fetal ultrasound images
CN113628174A (zh) 一种主观及客观超声医学图像质量评价方法及系统

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: 19930954

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: 19930954

Country of ref document: EP

Kind code of ref document: A1