[go: up one dir, main page]

CN111127414B - Perfusion image judgment system and method based on OPENCV and intelligent terminal - Google Patents

Perfusion image judgment system and method based on OPENCV and intelligent terminal Download PDF

Info

Publication number
CN111127414B
CN111127414B CN201911314612.7A CN201911314612A CN111127414B CN 111127414 B CN111127414 B CN 111127414B CN 201911314612 A CN201911314612 A CN 201911314612A CN 111127414 B CN111127414 B CN 111127414B
Authority
CN
China
Prior art keywords
image
pixel
value
perfusion
color
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.)
Active
Application number
CN201911314612.7A
Other languages
Chinese (zh)
Other versions
CN111127414A (en
Inventor
陈儇
刘鹏程
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.)
First Hospital Jinlin University
Original Assignee
First Hospital Jinlin University
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 First Hospital Jinlin University filed Critical First Hospital Jinlin University
Priority to CN201911314612.7A priority Critical patent/CN111127414B/en
Publication of CN111127414A publication Critical patent/CN111127414A/en
Application granted granted Critical
Publication of CN111127414B publication Critical patent/CN111127414B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • G06T5/92—Dynamic range modification of images or parts thereof based on global image properties
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/90—Determination of colour characteristics
    • 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
    • G06T2207/10072—Tomographic images
    • G06T2207/10088—Magnetic resonance imaging [MRI]
    • 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
    • G06T2207/30016—Brain

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Quality & Reliability (AREA)
  • Radiology & Medical Imaging (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Apparatus For Radiation Diagnosis (AREA)
  • Magnetic Resonance Imaging Apparatus (AREA)
  • Image Analysis (AREA)

Abstract

本发明属于图像处理技术领域，公开了一种基于OPENCV的灌注图像判定系统及方法、智能终端，将图片进行分割，将患者信息，时间序列信息等对结果判断没有意义的噪声去除掉；图像的脑区位置分割出来，把单张图片中的左右脑区分割；灌注图片的颜色系统用像素点去转化RGB‑>HSV；遍历像素点将像素点设置成图像在RGB颜色空间的值中的黑色或者白色；把目标色点用黑色像素点显示出来，非目标色点本发明用白色色素点显示出来；将数组中的每个元素给予一个权值；得到比较合适阈值参数。本发明可以实现对头核磁灌注图像的智能分析，智能的判别核磁灌注的临床分期，对于缺血性疾病的治疗有指导意义；可以判断出病变脑组织的严重程度。

Figure 201911314612

The invention belongs to the technical field of image processing, and discloses an OPENCV-based perfusion image determination system and method, and an intelligent terminal, which can segment the image, remove patient information, time-series information and other noises that are not meaningful for result determination; The location of the brain region is segmented, and the left and right brain regions in a single image are segmented; the color system of the perfusion image uses pixels to convert RGB->HSV; traverse the pixels to set the pixels to black in the value of the image in the RGB color space Or white; the target color point is displayed with black pixels, and the non-target color point is displayed with white pigment points; each element in the array is given a weight; a more appropriate threshold parameter is obtained. The invention can realize intelligent analysis of head nuclear magnetic perfusion images, and intelligently judge clinical stages of nuclear magnetic perfusion, which has guiding significance for the treatment of ischemic diseases; and can judge the severity of diseased brain tissue.

Figure 201911314612

Description

Perfusion image judgment system and method based on OPENCV and intelligent terminal
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a perfusion image judgment system and method based on OPENCV and an intelligent terminal.
Background
Currently, the closest prior art: due to aging population, the number of people suffering from ischemic brain diseases in China increases year by year, head nuclear magnetic perfusion examination is performed after patients suffering from cerebral ischemic diseases are admitted, and clinical doctors can grade the ischemic degree of the patients according to the head nuclear magnetic perfusion examination of the patients provided by the imaging department so as to formulate a treatment scheme. At present, the low perfusion state before infarction is divided into four stages by the commonly used brain perfusion stage in China according to the researches of Hieropi and the like: stage I1: TTP prolongation, MTT, rCBF and rCBV normal; and I2 stage: TTP and MTT prolongation, rCBF normal, rCBV normal or mild elevation; stage II 1: TTP, prolonged MTT and decreased rCBF, with essentially normal or mild decrease in rCBV; and II, stage 2: TTP, MTT prolonged, rCBF and rCBV decreased.
At present, the main applications are: the early cerebral infarction, transient ischemic attack, carotid artery stenosis, intracranial vascular stenosis, smog and other ischemic cerebrovascular disease intervention and before and after surgical evaluation. At present, the identification, interpretation and staging of perfusion images only depend on the artificial identification of each clinical doctor on the images, the possible abnormality on the images is identified by visual perception, and the subjective decision is made to a certain extent by depending on knowledge, experience and understanding of normal nuclear magnetic perfusion performance. Each clinician may have slight errors in the discrimination of the images, resulting in errors in the ranking of the images.
In summary, the problems of the prior art are as follows: at present, the identification, interpretation and division stages of perfusion images have errors on the judgment of the images, so that the errors on the image classification are caused.
The difficulty of solving the technical problems is as follows:
the reason why the above problems are present is that the judgment is highly subjective, and therefore, different diagnosis results are generated among different doctors. The problem is solved in the field of computer images, so that the artificial subjectivity can be removed, and a scientific diagnosis result can be obtained to assist a doctor in diagnosis and treatment. The technical difficulties in solving the current problems are the following aspects. 1. The color distribution of the perfusion images does not have a completely uniform standard, and different imaging doctors have large color matching difference when the perfusion images are output, which is also the main reason that clinical doctors have differences in judgment. 2. There is no gold standard when determining the degree of ischemia of a patient by perfusion images, that is, doctors can only estimate the degree of ischemia by so-called "feeling", and different doctors have their own measures in mind, which is also an important reason for the difference in determination. At this stage, the perfusion image can be further processed by image methods of a computer, such as image preprocessing, image segmentation and the like, so that more accurate image information can be obtained. The position and the severity of a focus of a patient are obtained more intuitively through pixel identification, and then a threshold value is set to judge a pixel point to obtain a result.
The significance of solving the technical problems is as follows:
by solving the problems, the method is a one-time organic collision between computer science and medicine, and the trend of future development is to solve the medical problem which is difficult to break through in the medical field through the computer science. In this regard, the results obtained in the present invention can be used by clinicians for clinical diagnosis and treatment. The method can help clinical doctors to improve the interpretation accuracy of the head nuclear magnetic perfusion images, help clinical doctors who lack clinical experience interpret and learn the head nuclear magnetic perfusion images in stages, and accelerate the maturity of the clinical doctors. Meanwhile, the accuracy and efficiency of interpretation and staging of perfusion images are improved, and help is provided for the work of clinical doctors; the combination of computer and clinical medicine is realized, and thus a way for fusing more computers and more clinical medicine is opened up.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a perfusion image judgment system and method based on OPENCV and an intelligent terminal.
The present invention is achieved in such a way that an OPENCV-based perfusion image determination method includes the steps of:
firstly, taking out corresponding perfusion images from an image library of a hospital by each perfusion image;
second, what color the target color point is in the perfusion image;
thirdly, preprocessing the perfusion image, segmenting the image, and removing noise which does not make sense in result judgment such as patient information, time sequence information and the like;
fourthly, the brain area position of the image is divided, and the left and right brain areas in a single picture are divided;
fifthly, converting RGB- > HSV by a color system of the perfusion picture by using pixel points;
sixthly, setting the pixel points to be black or white in the value of the RGB color space by traversing the pixel points through the threshold value of the pixel points to the HSV color space;
seventhly, obtaining pixel point differences of a single picture, obtaining all target color points of the single picture in a certain mode, displaying the target color points by using black pixel points, and displaying non-target color points by using white pixel points;
eighthly, obtaining pixel point difference of a single mode, giving a weight to each element in the array, wherein the sum of the weights corresponding to all the elements in the array is 1;
and ninthly, judging the disease condition classification and stage of the current mode by a threshold method, and adjusting a threshold parameter through a large amount of data after obtaining the value of the pixel color point of each mode to obtain a more appropriate threshold parameter.
Further, the function prototype of the first step reading picture is as follows:
CV_EXPORTS_W Mat imread(const String&filename，int flags＝IMREAD_COLOR)；
the returned value of Imread is of Mat type, the read image is returned, and an empty matrix object is returned when the image reading fails (Mat: data is NULL); parameter 1filename, picture filename read, using a relative path or an absolute path; parameter 2flags, a reading flag, for selecting the mode of reading the picture, the default value is IMREAD _ COLOR, and the setting of the flag value is related to what COLOR format the picture is read;
a function that recursively traverses the file directory, using the function findAllFile () function, whose function prototype is:
voidfindAllFile(const char*path，const char*format，intsum_pixel[4])；
the return value of the function is null type, the parameter 1path represents the address of the path, the parameter 2 represents the file type to be read, the format value is bmp, and the parameter 3 represents the sum of the number of traversal pixels.
Further, the third step divides the original segment 1168 807 into 680 x 700 segments.
Further, the algorithm for equally dividing the brain region in the fourth step includes: setting a pixelistanceminSuch a variable is initialized to 10000, which represents the minimum value of the number of pixel points from the image edge to the brain region; defining PixelDistanceleftAnd pixelistancerightThe two variables are initialized to 0, and represent the number of pixels from the edge of the image to the pixels which are not black pixels;
for a perfusion image, passing pixel pointsTraversing the images, taking one image as a matrix, and obtaining the value of the pixel point at the fixed position through coordinates; when traversing the image matrix, traversing from left to right, and for each row, when touching the first pixel which is not the background color, obtaining the column number of the pixel and the pixel distanceminBy comparison, if the row column number is less than pixelDistanceminIf so, the pixelDistance is updatedminAnd record the pixelDistance of the lineleftAnd pixelistancerightThese two values; in the process of traversing the image, pixelistanceminThis value may be updated at any time; PixelDistanceleftAnd pixelistancerightAnd pixelistanceminAre corresponding; when the image traversal is completed, the pixelistanceminThe minimum pixel point of the brain area, which is far away from the edge of the image, is stored as pixel distanceleftAnd pixelistancerightRespectively storing the number of pixel points from the brain area to the left edge of the image and the number of pixel points from the brain area to the right edge of the image;
pixelistanceminThe line with the minimum distance is used as a segmentation reference, the central axis of the brain region is found through the line, and the calculation formula is as follows:
axis＝pixelDis tan cemin+((picturewide-(pixelDis tan celeft+pixelDis tan ceright) 2) according to
The central axis of the brain region, i.e. the first column of the matrix, is obtained by the above formula, and the image is divided left and right on the basis of the central axis to obtain left and right brain regions.
Further, the fifth step includes that the color system of the perfusion picture is composed of RGB- > HSV:
(1)max＝max(R，G，B)；
(2)min＝min(R，G，B)；
(3)V＝max(R，G，B)；
(4)S＝(max-min)/max；
(5)if R＝max，H＝(G-B)/(max-min)*60；
(6)ifG＝max，H＝120+(B-R)/(max-min)*60；
(7)ifB＝max，H＝240+(R-G)/(max-min)*60；
(8)ifH<0，H＝H+360。
further, in the sixth step, through the set threshold of the pixel point to the HSV color space, the traversal pixel point sets the pixel point to be pixelRGBBlack or white, 0 for black, 255 for white; the threshold algorithm is as follows:
Figure GDA0003377136390000051
in which pixelRGB，pixelHSV，pixelTagHSVRespectively represent the values of three dimensions; pixelRGBRepresenting the value of an image in RGB color space if pixelRGBWhen the value is 255, the values of the three dimensions are 255; PixelTagHSVRepresenting the range of three dimensional values of the current target color point H, S, V, pixelTagHSVThe value of (b) is obtained according to the physician selected color used.
Further, the seventh step records the number of target pixel points to obtain pixelleftAnd pixelrightSo that the difference value pixeldifferenceComprises the following steps:
pixeldifference＝pixelleft-pixelright；
by pixelsleftSubtract pixelrightIf the obtained result is a positive value, the color points of the target pixels of the left brain are more; if the color point of the target pixel of the right brain is negative, more color points are formed; if 0, it indicates that the pixel points of left and right brains are consistent in number, pixeldifferenceThe final result pixel point result of a single picture is obtained;
the eighth step obtains pixelArrdifferenceAfter array, now pixelArrdifferenceAll elements in the array are sorted in descending order, pixelArrdifference[0]The value stored is maximum, pixelArrdifference[17]The value deposited is minimal; each element in the array is given a weight value, and the sum of the corresponding weight values of all the elements in the array is 1(ii) a The specific formula is as follows:
Figure GDA0003377136390000052
wherein, weight [ i ] represents the weight value of the color point of the pixel of the ith picture in the mode;
then the value pixelSum of the total pixel color point in a certain modalityxComprises the following steps:
Figure GDA0003377136390000061
wherein pixelSumxRepresenting the total number of pixel points in X mode, wherein X is TTP, MTT, CBV and CBF;
after the ninth step of obtaining pixelSum of each modality, adjusting the threshold parameter according to data of a large number of patients to finally obtain a relatively proper threshold parameter, wherein the threshold is defined as follows:
Figure GDA0003377136390000062
Figure GDA0003377136390000063
Figure GDA0003377136390000064
Figure GDA0003377136390000065
and judging the final disease staging result according to the obtained results of all the modes, wherein the judgment of the disease staging result is based on the following steps:
Figure GDA0003377136390000066
Figure GDA0003377136390000067
Figure GDA0003377136390000071
Figure GDA0003377136390000072
SumR＝RTTP+RMTT+RCBV+RCBF；
Figure GDA0003377136390000073
from the above, the final diagnosis result is obtained.
Another object of the present invention is to provide an OPENCV-based perfusion image determination system that implements the OPENCV-based perfusion image determination method, the OPENCV-based perfusion image determination system including:
the perfusion image extracting module is used for extracting corresponding perfusion images of each perfusion image in an image library of the hospital;
a target color point labeling module for what color the target color point is in the perfusion image;
the perfusion image preprocessing module is used for preprocessing a perfusion image, segmenting the image and removing noise which does not make sense in result judgment such as patient information, time sequence information and the like;
the image segmentation module is used for segmenting the brain region position of the image and segmenting the left and right brain regions in a single picture;
the color conversion module is used for realizing that pixel points for a color system of the perfusion picture are used for converting RGB- > HSV;
the threshold setting module is used for setting the pixel points to be black or white in the value of the RGB color space through the traversal pixel points by setting the threshold of the pixel points to the HSV color space;
the pixel point display module is used for obtaining the pixel point difference of a single picture, obtaining all target color points of the single picture in a certain mode, displaying the target color points by using black pixel points, and displaying non-target color points by using white pixel points;
the weight value giving module is used for giving a weight value to each element in the array according to the obtained pixel point difference of the single mode, and the sum of the weight values corresponding to all the elements in the array is 1;
and the threshold parameter acquisition module is used for judging the disease condition classification and stage of the current mode by using a threshold method, and adjusting the threshold parameter through a large amount of data after obtaining the value of the pixel color point of each mode to obtain a more appropriate threshold parameter.
Another object of the present invention is to provide an information data processing terminal for implementing the OPENCV-based perfusion image determination method.
It is another object of the present invention to provide a computer-readable storage medium comprising instructions which, when executed on a computer, cause the computer to perform the OPENCV-based perfusion image determination method.
In summary, the advantages and positive effects of the invention are: the present invention provides clinicians with an OPENCV-based perfusion image disease grading algorithm that is implemented as executable software running in a fixed facility in a hospital. According to the invention, through learning and analyzing a plurality of head nuclear magnetic perfusion images, system software capable of identifying the perfusion images is manufactured, intelligent analysis of the head nuclear magnetic perfusion images can be realized more quickly and accurately, and accurate and quick clinical staging is realized.
The invention can realize intelligent analysis of the head nuclear magnetic perfusion image, quickly judge the change of the TTP, MTT, CBF and CBV indexes, intelligently judge the clinical stage of nuclear magnetic perfusion, and has guiding significance for the treatment of ischemic diseases. Even the brain tissue structure with corresponding change can be automatically identified and divided, and the severity of the pathological brain tissue can be judged by analyzing the color pixels of the image.
Drawings
FIG. 1 is a schematic structural diagram of a perfusion image determination system based on OPENCV according to an embodiment of the present invention;
in the figure: 1. a perfusion image extraction module; 2. a target color point marking module; 3. a perfusion image preprocessing module; 4. an image segmentation module; 5. a color conversion module; 6. a threshold setting module; 7. a pixel point display module; 8. a weight value giving module; 9. and a threshold parameter acquisition module.
Fig. 2 is a flowchart of a perfusion image determination method based on OPENCV according to an embodiment of the present invention.
Fig. 3 is a schematic diagram of a target color palette provided in an embodiment of the invention.
Fig. 4 is a flowchart of obtaining piexlTag according to an embodiment of the present invention.
Fig. 5 is a diagram of image segmentation (taking one image of a TTP modality of a patient as an example) provided by an embodiment of the present invention.
Fig. 6 is a schematic diagram of a brain region segmentation result according to an embodiment of the present invention.
Fig. 7 is a schematic diagram of RGB- > HSV color space conversion according to an embodiment of the present invention.
Fig. 8 is a schematic diagram of finding a target color point according to an embodiment of the present invention.
Fig. 9 is a schematic diagram of a program screenshot 1 provided in an embodiment of the present invention.
Fig. 10 is a schematic diagram of a program screenshot 2 provided in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
In view of the problems in the prior art, the present invention provides a perfusion image determination system and method based on OPENCV, and an intelligent terminal, and the present invention is described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the OPENCV-based perfusion image determination system provided by the embodiment of the present invention includes:
a perfusion image extracting module 1, configured to extract a corresponding perfusion image of each perfusion image in an image library of a hospital.
A target color point labeling module 2 for what color the target color point is in the perfusion image.
And the perfusion image preprocessing module 3 is used for preprocessing the perfusion image, segmenting the picture and removing noise which does not have significance on result judgment, such as patient information, time sequence information and the like.
And the image segmentation module 4 is used for segmenting the brain region position of the image and segmenting the left and right brain regions in a single picture.
And the color conversion module 5 is used for realizing that pixel points for a color system of the perfusion picture are used for converting RGB- > HSV.
And the threshold setting module 6 is used for setting the pixel points to be black or white in the value of the RGB color space through the traversal pixel points according to the set threshold of the pixel points to the HSV color space.
And the pixel point display module 7 is used for obtaining the pixel point difference of the single picture, obtaining all target color points of the single picture in a certain mode, displaying the target color points by using black pixel points, and displaying non-target color points by using white pixel points.
And the weight value giving module 8 is used for giving a weight value to each element in the array according to the obtained pixel point difference of the single mode, and the sum of the corresponding weight values of all the elements in the array is 1.
The threshold parameter obtaining module 9 is configured to implement a threshold method to determine the disease condition classification and stage of the current modality, and adjust the threshold parameter through a large amount of data after obtaining the value of the pixel color point of each modality, so as to obtain a relatively appropriate threshold parameter.
As shown in fig. 2, the method for determining a perfusion image based on OPENCV according to an embodiment of the present invention includes the following steps:
s201: taking out the corresponding perfusion image from the image library of the hospital for each perfusion image;
s202: what color the target color point is in the perfusion image;
s203: preprocessing the perfusion image, segmenting the image, and removing noise which does not make sense in result judgment such as patient information, time sequence information and the like;
s204: the brain area position of the image is divided, and the left and right brain areas in a single picture are divided;
s205: the color system of the perfusion picture uses pixel points to convert RGB- > HSV;
s206: traversing pixel points to set the pixel points to be black or white in the value of the image in the RGB color space through the set threshold of the pixel points to the HSV color space;
s207: obtaining pixel point differences of a single picture, obtaining all target color points of the single picture in a certain mode, displaying the target color points by using black pixel points, and displaying non-target color points by using white pixel points;
s208: obtaining pixel point difference of a single mode, giving a weight to each element in the array, wherein the sum of the weights corresponding to all the elements in the array is 1;
s209: the threshold method judges the grading and stage of the disease condition of the current mode, and after the value of the pixel color point of each mode is obtained, the threshold parameter is adjusted through a large amount of data to obtain a more appropriate threshold parameter.
The technical solution of the present invention is further described below with reference to the accompanying drawings.
The perfusion image judging method based on OPENCV provided by the embodiment of the invention comprises the following steps:
deriving perfusion images of a patient from an imaging system
The perfusion image of each patient is taken out of the image library of the hospital before using the device, and the perfusion image of the corresponding patient is used in the subsequent steps. The C + + language used by the invention reads the function prototype of the picture as follows:
CV_EXPORTS_W Mat imread(const String&filename，int flags＝IMREAD_COLOR)；
the returned value of Imread is of Mat type, that is, the read image is returned, and an empty matrix object is returned when the image reading fails (Mat:: data ═ NULL). Parameter 1filename, picture filename read, may use a relative path or an absolute path, but must be accompanied by a full file extension (picture format suffix). Parameter 2flags, a read flag, for selecting the mode of reading the picture, the default value is read _ COLOR, and the flag value is set according to what COLOR format the picture is read.
Since i want to recursively traverse the file directory, a function for recursively traversing the file directory is designed, and a findAllFile () function is used, which is prototype of the function:
voidfindAllFile(const char*path，const char*format，intsum_pixel[4])；
the return value of the function is null type, the parameter 1path represents the address of the path, the parameter 2 represents the file type to be read, the value of format in the invention is 'bmp', and the parameter 3 represents the sum of the number of traversal pixel points.
After the first step, perfusion images of the patient may be obtained from the imaging system for use in subsequent steps.
2 setting a target color point
As shown in fig. 3 and 4, the physician tells the software what color point of the target color in the perfusion image is before use. If the color of the TTP-perfused area is red, red is selected. The software sets the red to the target color point, thus obtaining the value of the HSV color space of the red pixel points, namely pixelTagHSV。
3 preprocessing the perfusion image
After the first step, the medical image of the patient has been taken from the system library, and the perfusion image is then preprocessed. The preprocessing refers to dividing the picture, and aims to remove noise which does not have significance on result judgment, such as patient information, time sequence information and the like, so that the noise information does not influence subsequent judgment of pixel points. The original image segment (1168 807) is segmented into 680 700 sizes to achieve data enhancement. The splitting operation is shown in fig. 5 below.
4 segmentation of left and right semi-brain regions
And segmenting the brain region position of the image. In order to divide the left and right brain areas in a single picture, the resolution of the image is 680 x 700 in the previous step, so that the width of the image has 680 pixel points, but the invention cannot divide 680/2 equally according to the length of the width of the picture, and the division method is inaccurate because the brain area of a patient is not in the middle of the picture. The invention provides an algorithm for equal division of brain regions, which comprises the following steps:
in the algorithm, the invention sets a pixelistanceminSuch a variable is initialized to 10000, which represents the minimum value of the number (distance) of pixels from the edge of the image to the brain region. PixelDistance is also definedleftAnd pixelistancerightThese two variables are initialized to 0, and represent the number (distance) of pixels from the edge of the image to pixels other than black pixels. For example, pixelistanceleftThe number (distance) of pixels at the left edge of the image from the first pixel which is not black is represented. The algorithm idea of the invention is as follows:
for a perfusion image, the invention traverses the image through the pixel points. The invention can obtain the value of the fixed position pixel point through the coordinate by taking an image as a matrix (700 rows and 680 columns). Because the value of the pixel point of the background color is different from the value of the pixel point of the brain area, the invention can easily distinguish whether the pixel point belongs to the brain area part or the background part. When the invention traverses the image (matrix), the image is traversed from left to right, and for each row, when the first pixel which is not the background color is touched, the invention obtains the column number of the pixel and the pixel distanceminThis value is compared if the row column count is less than pixelistanceminIf so, the pixelDistance is updatedminAnd record the pixelDistance of the lineleftAnd pixelistancerightThese two values. In the process of traversing the image, pixelistanceminThis isThe value may be updated at any time. PixelDistanceleftAnd pixelistancerightAnd pixelistanceminIs the corresponding. When the image traversal is completed, the pixelistanceminThe pixel distance which is the minimum distance between the brain area and the image edge pixel point is storedleftAnd pixelistancerightThe number (distance) of pixel points from the brain area to the left edge of the image and the number (distance) of pixel points from the brain area to the right edge of the image are respectively stored.
After these three variables are obtained separately, the invention will use pixelDistanceminThe line with the smallest distance is used as the segmentation reference, i.e. the central axis of the brain region is found through the line. The calculation formula of the invention is as follows:
axis＝pixelDis tan cemin+((picturewide-(pixelDis tan celeft+pixelDis tan ceright) 2) according to the above formula, the present invention can obtain the central axis of the brain region, i.e. the several columns of the matrix, and then divide the image left and right on this basis to obtain the left and right brain regions. According to experiments, the algorithm provided by the invention is very accurate. The operation in this step is as in fig. 6.
5 the color system of the perfusion picture is composed of RGB- > HSV
The color system of the original picture is RGB, which is now converted into HSV. The RGB- > HSV color system conversion is realized by using pixel points, and the algorithm is executed for each pixel point in an image and is divided into eight steps:
(1)max＝max(R，G，B)；
(2)min＝min(R，G，B)；
(3)V＝max(R，G，B)；
(4)S＝(max-min)/max；
(5)if R＝max，H＝(G-B)/(max-min)*60；
(6)ifG＝max，H＝120+(B-R)/(max-min)*60；
(7)ifB＝max，H＝240+(R-G)/(max-min)*60；
(8)ifH<0，H＝H+360。
through the above steps, the perfusion image has become an image of the HSV color system, as shown in fig. 7.
6 finding the target color point
According to the invention, through the threshold value of the pixel point for the HSV color space, the pixel point is set to pixel by traversing the pixel pointRGBBlack or white (0 for black and 255 for white). The threshold algorithm is as follows:
Figure GDA0003377136390000141
in which pixelRGB，pixelHSV，pixelTagHSVRepresented are the values of the three dimensions, respectively. PixelRGBRepresenting the value of an image in RGB color space if pixelRGB255 indicates that the values of all three dimensions are 255. PixelTagHSVRepresenting the range of three dimensional values of H, S, V for the current target color point. PixelTagHSVThe value of (b) is obtained by selecting a color according to the doctor used in the device of the invention. The result of this step is shown in fig. 8.
And 7, obtaining pixel point differences of a single picture, obtaining all target color points of the single picture in a certain mode, displaying the target color points by using black pixel points, and displaying non-target color points by using white pixel points. In the step, the invention records the number of the target pixel points to obtain pixelleftAnd pixelrightTherefore, the differential pixel of the present inventiondifferenceComprises the following steps:
pixeldifference＝pixelleft-pixelright；
here, pixels are used in the inventionleftSubtract pixelrightIf the obtained result is a positive value, the color point of the target pixel of the left brain is more. If negative, the target pixel color point of the right brain is more. If the number is 0, the number of the pixel points of the left brain and the right brain is consistent. PixeldifferenceThe final result pixel point result of a single picture is obtained.
8 obtaining pixel point difference of single mode
The invention obtains the difference pixel of a single target pixel point in a certain modedifference. The invention calculates the pixel point difference of all images in the whole mode to obtain the pixel point difference of the whole mode. The invention obtains all pixel point differences of the current mode by traversing the image in the whole mode and stores the pixel point differences in an array. This array is called pixelArr in the inventiondifference. The present invention does not directly sum all the elements in this array, and such values are not necessarily accurate and the results obtained are certainly not good. Here, the present invention designs a weighted sum algorithm. The algorithm is described as follows:
the invention obtains pixelArrdifferenceAfter array, now pixelArrdifferenceAll elements in the array are sorted in descending order, that is to say pixelArrdifference[0]The value stored is the largest, and there are typically 18 pictures in one modality, pixelArrdifference[17]The value stored is minimal. Then, the invention gives a weight to each element in the array, and the sum of the weights corresponding to all the elements in the array is 1. The specific formula is as follows:
Figure GDA0003377136390000151
wherein, weight [ i ] represents the weight value of the color point of the pixel of the ith picture in the mode.
Then the value pixelSum of the total pixel color point in a certain modalityxComprises the following steps:
Figure GDA0003377136390000152
wherein pixelSumxRepresenting the total number of pixels in the x mode. X can be TTP, MTT, CBV and CBF.
9 threshold value method for judging grading and staging of current modal state of illness
After the pixelSum of each modality is obtained, the method adjusts the threshold parameter according to the data of a large number of patients to finally obtain a more appropriate threshold parameter, and the effect of the algorithm is more accurate through a large number of tests. The threshold of the invention is defined as follows:
Figure GDA0003377136390000153
Figure GDA0003377136390000154
Figure GDA0003377136390000155
Figure GDA0003377136390000156
the final disease staging result is judged according to the results of all the modes, and the judgment basis of the disease staging result is as follows:
Figure GDA0003377136390000161
Figure GDA0003377136390000162
Figure GDA0003377136390000163
Figure GDA0003377136390000164
SumR＝RTTP+RMTT+RCBV+RCBF；
Figure GDA0003377136390000165
from the above, the final diagnosis result can be obtained. According to the judgment of the doctor, the result is more accurate.
The technical effects of the present invention will be described in detail below with reference to program screenshots.
As shown in fig. 9 and 10.
(1) Taking patient A as an example:
the image of the perfusion exam in Guizhou, is reported as follows: head perfusion, left side temporoparietal lobe, frontal occipital lobe partial area and basal ganglia, radiation coronal region MTT, TTP prolongation, CBV, CBF did not see obvious abnormal changes.
The results of the present invention are shown in FIG. 9.
(2) Taking patient B as an example:
the Lixingjie perfusion exam imaging department reports as follows: PWI: the right hemioval center MTT, TTP is slightly longer, and CBV, CBF of bilateral cerebral hemispheres have no clear abnormality.
The results of the present invention are shown in FIG. 10.
Compared with the cases of many patients in hospitals, the diagnosis result of the invention is more accurate with the diagnosis results of image departments and clinicians, and the pixel-level diagnosis can even reflect the real conditions of the patients.
It should be noted that the embodiments of the present invention can be realized by hardware, software, or a combination of software and hardware. The hardware portion may be implemented using dedicated logic; the software portions may be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or specially designed hardware. Those skilled in the art will appreciate that the apparatus and methods described above may be implemented using computer executable instructions and/or embodied in processor control code, such code being provided on a carrier medium such as a disk, CD-or DVD-ROM, programmable memory such as read only memory (firmware), or a data carrier such as an optical or electronic signal carrier, for example. The apparatus and its modules of the present invention may be implemented by hardware circuits such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of hardware circuits and software, e.g., firmware.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.

Claims (10)

1.一种基于OPENCV的灌注图像判定方法，其特征在于，所述基于OPENCV的灌注图像判定方法包括以下步骤：1. a perfusion image determination method based on OPENCV, is characterized in that, described perfusion image determination method based on OPENCV comprises the following steps: 第一步，每个灌注图像在医院的图像库中，将对应的灌注图像取出；In the first step, each perfusion image is in the hospital's image library, and the corresponding perfusion image is taken out; 第二步，在灌注图像中的目标色点是什么颜色；The second step, what color is the target color point in the perfusion image; 第三步，灌注图像进行预处理，将图片进行分割，将患者信息，时间序列信息等对结果判断没有意义的噪声去除掉；In the third step, the perfusion image is preprocessed, the image is segmented, and the noise that is not meaningful to the judgment of the result, such as patient information and time series information, is removed; 第四步，图像的脑区位置分割出来，把单张图片中的左右脑区分割；The fourth step is to segment the brain area of the image, and segment the left and right brain areas in a single image; 第五步，灌注图片的颜色系统用像素点去转化RGB->HSV；The fifth step, the color system of the perfused image uses pixels to convert RGB->HSV; 第六步，通过像素点的对HSV色彩空间的设置阈值，遍历像素点将像素点设置成图像在RGB颜色空间的值中的黑色或者白色；The sixth step is to traverse the pixel points to set the pixel points to be black or white in the value of the image in the RGB color space by setting the threshold value of the HSV color space for the pixel point; 第七步，得到单张图片的像素点差异，得到某个模态的单张照片的所有目标色点，并把目标色点用黑色像素点显示出来，非目标色点本发明用白色素点显示出来；The seventh step is to obtain the pixel point difference of a single picture, obtain all the target color points of a single photo of a certain mode, and display the target color points with black pixels, and the non-target color points use white pixels in the present invention. display; 第八步，得到单个模态的像素点差异，将数组中的每个元素给予一个权值，数组中的所有元素对应权值的和为1；The eighth step, obtain the pixel point difference of a single mode, give each element in the array a weight, and the sum of the corresponding weights of all elements in the array is 1; 第九步，阈值法判断当前模态的病情分级与分期，得到每个模态的像素色点的值之后，通过大量的数据调整阈值参数，得到比较合适阈值参数。In the ninth step, the threshold method is used to judge the disease grading and staging of the current modality, and after obtaining the value of the pixel color point of each modality, the threshold parameter is adjusted through a large amount of data to obtain a more appropriate threshold parameter. 2.如权利要求1所述的基于OPENCV的灌注图像判定方法，其特征在于，其中第一步的具体处理为：2. the perfusion image determination method based on OPENCV as claimed in claim 1, is characterized in that, wherein the concrete processing of the first step is: CV_EXPORTS_W Mat imread(const String&filename，int flags＝IMREAD_COLOR)；CV_EXPORTS_W Mat imread(const String&filename, int flags=IMREAD_COLOR); Imread的返回值为Mat类型，返回读取的图像，读取图像失败时返回一个空的矩阵对象(Mat::data＝＝NULL)；参数1filename，读取的图片文件名，使用相对路径或者绝对路径；参数2flags，一个读取标记，用于选择读取图片的方式，默认值为IMREAD_COLOR，flag值的设定与用什么颜色格式读取图片有关；The return value of Imread is of type Mat, which returns the read image, and returns an empty matrix object (Mat::data==NULL) when the image fails to be read; parameter 1filename, the name of the image file to be read, uses a relative path or an absolute path Path; parameter 2flags, a read flag, used to select the way to read the picture, the default value is IMREAD_COLOR, the setting of the flag value is related to what color format is used to read the picture; 递归遍历文件目录的函数，使用函数findAllFile()函数，其函数原型为：The function to recursively traverse the file directory, use the function findAllFile() function, and its function prototype is: voidfindAllFile(const char*path，const char*format，intsum_pixel[4])；void findAllFile(const char*path, const char*format, intsum_pixel[4]); 此函数的返回值为空类型，参数1代表的是路径的地址，参数2代表的是要读取的文件类型，format的值为bmp，参数3代表的是遍历像素点的个数和。The return value of this function is an empty type, parameter 1 represents the address of the path, parameter 2 represents the file type to be read, the value of format is bmp, and parameter 3 represents the sum of the number of traversed pixels. 3.如权利要求1所述的基于OPENCV的灌注图像判定方法，其特征在于，其中第三步的具体处理为：将原图分1168*807分割成大小为680*700。3 . The method for determining perfusion images based on OPENCV according to claim 1 , wherein the specific processing of the third step is: dividing the original image into 1168*807 with a size of 680*700. 4 . 4.如权利要求1所述的基于OPENCV的灌注图像判定方法，其特征在于，其中第四步的具体处理为：脑区均等分割的算法包括：设置一个pixelDistancemin这样一个变量，并初始化为10000，代表的是图像边缘到脑区的像素点个数的最小值；定义pixelDistanceleft和pixelDistanceright这两个变量，初始化为0，代表的是图像边缘分别到不是黑色像素点的像素点个数；4. the perfusion image determination method based on OPENCV as claimed in claim 1, is characterized in that, wherein the concrete processing of the 4th step is: the algorithm of equal segmentation of brain area comprises: set such a variable as pixelDistance min , and be initialized to 10000 , which represents the minimum number of pixels from the edge of the image to the brain area; define two variables, pixelDistance left and pixelDistance right , initialized to 0, which represent the number of pixels from the edge of the image that are not black pixels; 对于一张灌注图像，通过像素点将图像遍历，把一张图像当做一个矩阵，通过坐标去得到固定位置像素点的值；当遍历图像矩阵时，从左向右遍历，对于每一行，当碰到第一个不是背景色像素点的时候，将该像素点的列数得到并与pixelDistancemin比较，如果该行的列数小于pixelDistancemin，则更新pixelDistancemin，并记录下该行的pixelDistanceleft和pixelDistanceright这两个值；在遍历图像的过程中，pixelDistancemin这个值可能会随时更新的；pixelDistanceleft和pixelDistanceright与pixelDistancemin是对应的；当图像遍历完成时，pixelDistancemin存放的是脑区距离图像边缘像素点最少那个，pixelDistanceleft和pixelDistanceright分别存放的是脑区到图像左边缘的像素点个数和脑区到图像右边缘的像素点个数；For a perfusion image, the image is traversed through pixels, an image is regarded as a matrix, and the value of the pixel at a fixed position is obtained through the coordinates; when traversing the image matrix, it is traversed from left to right, for each row, when touching When the first pixel is not the background color, get the number of columns of the pixel and compare it with pixelDistance min . If the number of columns in the row is less than pixelDistance min , update pixelDistance min and record the pixelDistance left and pixelDistance left of the row. The two values of pixelDistance right ; in the process of traversing the image, the value of pixelDistance min may be updated at any time; pixelDistance left and pixelDistance right correspond to pixelDistance min ; when the image traversal is completed, pixelDistance min stores the distance of the brain area The image edge has the least number of pixels, pixelDistance left and pixelDistance right respectively store the number of pixels from the brain area to the left edge of the image and the number of pixels from the brain area to the right edge of the image; 将pixelDistancemin距离最小的这一行作为分割基准，通过这一行找到脑区的中轴线，计算公式如下：The line with the smallest pixelDistance min distance is used as the segmentation benchmark, and the central axis of the brain area is found through this line. The calculation formula is as follows: axis＝pixelDis tan cemin+((picturewide-(pixelDis tan celeft+pixelDis tanceright)÷2)根据以上公式，得到脑区的中轴线，也就是矩阵的第几列，将图像以这个基准左右分割，得到左右脑区。axis=pixelDis tance min +((picture wide -(pixelDis tance left +pixelDis tance right )÷2)According to the above formula, the central axis of the brain area is obtained, which is the column of the matrix, and the image is left and right based on this benchmark Segmentation to obtain left and right brain regions. 5.如权利要求1所述的基于OPENCV的灌注图像判定方法，其特征在于，其中第五步的具体处理为：将灌注图片的颜色系统由RGB-＞HSV包括：5. The method for judging perfusion images based on OPENCV as claimed in claim 1, wherein the specific processing of the fifth step is: changing the color system of the perfusion images from RGB->HSV to include: (1)max＝max(R，G，B)；(1) max=max(R, G, B); (2)min＝min(R，G，B)；(2) min=min(R, G, B); (3)V＝max(R，G，B)；(3) V=max(R, G, B); (4)S＝(max-min)/max；(4) S=(max-min)/max; (5)ifR＝max，H＝(G-B)/(max-min)*60；(5) ifR=max, H=(G-B)/(max-min)*60; (6)ifG＝max，H＝120+(B-R)/(max-min)*60；(6) ifG=max, H=120+(B-R)/(max-min)*60; (7)ifB＝max，H＝240+(R-G)/(max-min)*60；(7) ifB=max, H=240+(R-G)/(max-min)*60; (8)ifH＜0，H＝H+360。(8) if H<0, H=H+360. 6.如权利要求1所述的基于OPENCV的灌注图像判定方法，其特征在于，所述第六步通过像素点的对HSV色彩空间的设置阈值，遍历像素点将像素点设置成pixelRGB中的黑色或者白色，0代表黑色，255代表白色；阈值算法如下：6. the perfusion image judging method based on OPENCV as claimed in claim 1, is characterized in that, described 6th step is through the setting threshold value of pixel point to HSV color space, traverse pixel point and pixel point is set to pixel RGB . Black or white, 0 represents black, 255 represents white; the threshold algorithm is as follows:
Figure FDA0003377136380000031
Figure FDA0003377136380000031
 其中pixelRGB，pixelHSV，pixelTagHSV分别代表的是三个维度的值；pixelRGB代表图像在RGB颜色空间的值，如果pixelRGB＝255则说明其三个维度的值都是255；pixelTagHSV代表当前目标色点H，S，V三个维度值的范围，pixelTagHSV的值是根据使用的医生选择颜色而得到的。Among them, pixel RGB , pixel HSV , pixelTag HSV respectively represent the value of three dimensions; pixel RGB represents the value of the image in the RGB color space, if pixel RGB = 255, it means that the value of its three dimensions is 255; pixelTag HSV represents The range of the three dimension values of the current target color point H, S, and V, the value of pixelTag HSV is obtained according to the color selected by the doctor. 7.如权利要求1所述的基于OPENCV的灌注图像判定方法，其特征在于，其中第七步的具体处理为：将目标像素点的个数记录，得到pixelleft和pixelright，故差值pixeldifference为：7. the perfusion image determination method based on OPENCV as claimed in claim 1, is characterized in that, wherein the concrete processing of the 7th step is: record the number of target pixel points, obtain pixel left and pixel right , so difference pixel The difference is: pixeldifference＝pixelleft-pixelright；pixel difference = pixel left - pixel right ; 用pixelleft减去pixelright，得到的结果如果为正值，则左脑的目标像素色点较多；如果为负值，则右脑的目标像素色点较多；如果为0，则说明左右脑像素点数量一致，pixeldifference为单张图片最后的结果像素点结果；Subtract pixel right from pixel left . If the result is a positive value, the left-brain target pixel has more color points; if it is a negative value, the right-brain target pixel has more color points; if it is 0, it means the left and right The number of brain pixels is the same, and the pixel difference is the final pixel result of a single image; 所述第八步得到pixelArrdifference数组之后，现将pixelArrdiffeerence数组中的所有元素进行排序，按从小到大的顺序排序，pixelArrdifference[0]存放的值是最大的，pixelArrdifference[17]存放的值是最小的；将数组中的每个元素给予一个权值，数组中的所有元素对应权值的和为1；具体公式如下：After obtaining the pixelArr difference array in the eighth step, now sort all the elements in the pixelArr difference array in ascending order, the value stored in pixelArr difference [0] is the largest, and the value stored in pixelArr difference [17] The value is the smallest; each element in the array is given a weight, and the sum of the corresponding weights of all elements in the array is 1; the specific formula is as follows:
Figure FDA0003377136380000041
Figure FDA0003377136380000041
 其中，weight[i]代表的是此模态第i个图片的像素色点的权重值；Among them, weight[i] represents the weight value of the pixel color point of the i-th picture of this modality; 那么，某个模态中总共的像素色点的值pixelSumx为：Then, the value pixelSum x of the total pixel color points in a certain mode is:
Figure FDA0003377136380000042
Figure FDA0003377136380000042
 其中pixelSumx代表的是x模态的像素点总数，x是TTP，MTT，CBV，CBF四个模态；Among them, pixelSum x represents the total number of pixels in the x mode, and x is the four modes of TTP, MTT, CBV, and CBF; 所述第九步得到每个模态的pixelSum之后，通过大量患者的数据调整阈值参数，最终得到比较合适阈值参数，阈值的定义如下：After obtaining the pixelSum of each modality in the ninth step, adjust the threshold parameters through the data of a large number of patients, and finally obtain more suitable threshold parameters. The definition of the threshold is as follows:
Figure FDA0003377136380000043
Figure FDA0003377136380000043
Figure FDA0003377136380000044
Figure FDA0003377136380000044
Figure FDA0003377136380000045
Figure FDA0003377136380000045
Figure FDA0003377136380000046
Figure FDA0003377136380000046
 根据得到的各个模态的结果判断最后的病情分期结果，病情分期结果的判断依据如下：According to the obtained results of each modal, the final disease staging result is judged, and the judgment basis of the disease staging result is as follows:
Figure FDA0003377136380000051
Figure FDA0003377136380000051
Figure FDA0003377136380000052
Figure FDA0003377136380000052
Figure FDA0003377136380000053
Figure FDA0003377136380000053
Figure FDA0003377136380000054
Figure FDA0003377136380000054
 SumR＝RTTP+RMTT+RCBV+RCBF；Sum R = R TTP + R MTT + R CBV + R CBF ;
Figure FDA0003377136380000055
Figure FDA0003377136380000055
 根据以上，得到最后的诊断结果。According to the above, the final diagnosis result is obtained. 8.一种实施权利要求1～7任意一项所述基于OPENCV的灌注图像判定方法的基于OPENCV的灌注图像判定系统，其特征在于，所述基于OPENCV的灌注图像判定系统包括：8. An OPENCV-based perfusion image determination system implementing the OPENCV-based perfusion image determination method according to any one of claims 1 to 7, wherein the OPENCV-based perfusion image determination system comprises: 灌注图像取出模块，用于将每个灌注图像在医院的图像库中对应的灌注图像取出；a perfusion image extraction module, used for extracting the perfusion image corresponding to each perfusion image in the hospital's image library; 目标色点标注模块，用于在灌注图像中的目标色点是什么颜色；Target color point labeling module, what color is the target color point in the perfusion image; 灌注图像预处理模块，用于灌注图像进行预处理，将图片进行分割，将患者信息，时间序列信息等对结果判断没有意义的噪声去除掉；The perfusion image preprocessing module is used to preprocess the perfusion image, segment the image, and remove the noise that is not meaningful to the judgment of the result, such as patient information and time series information; 图像分割模块，用于将图像的脑区位置分割出来，把单张图片中的左右脑区分割；The image segmentation module is used to segment the brain area of the image, and segment the left and right brain areas in a single image; 颜色转化模块，用于实现灌注图片的颜色系统用像素点去转化RGB->HSV；The color conversion module is used to convert the color system of the perfused image to RGB->HSV with pixels; 阈值设置模块，用于通过像素点的对HSV色彩空间的设置阈值，遍历像素点将像素点设置成图像在RGB颜色空间的值中的黑色或者白色；The threshold value setting module is used to set the threshold value of the HSV color space through the pixel point, and traverse the pixel point to set the pixel point to be black or white in the value of the image in the RGB color space; 像素点显示模块，用于将得到单张图片的像素点差异，得到某个模态的单张照片的所有目标色点，并把目标色点用黑色像素点显示出来，非目标色点本发明用白色色素点显示出来；The pixel point display module is used to obtain the pixel point difference of a single picture, obtain all the target color points of a single photo of a certain mode, and display the target color point with black pixel points, and the non-target color point of the present invention Shown with white pigmented dots; 权值赋予模块，用于将得到单个模态的像素点差异，将数组中的每个元素给予一个权值，数组中的所有元素对应权值的和为1；The weight assignment module is used to obtain the pixel point difference of a single mode, and assign a weight to each element in the array, and the sum of the corresponding weights of all elements in the array is 1; 阈值参数获取模块，用于实现阈值法判断当前模态的病情分级与分期，得到每个模态的像素色点的值之后，通过大量的数据调整阈值参数，得到比较合适阈值参数。The threshold parameter acquisition module is used to implement the threshold method to judge the disease grading and staging of the current modality. After obtaining the value of the pixel color point of each modality, adjust the threshold parameters through a large amount of data to obtain more appropriate threshold parameters. 9.一种实现权利要求1～7任意一项所述基于OPENCV的灌注图像判定方法的信息数据处理终端。9 . An information data processing terminal for implementing the OPENCV-based perfusion image determination method according to any one of claims 1 to 7 . 10.一种计算机可读存储介质，包括指令，当其在计算机上运行时，使得计算机执行如权利要求1～7任意一项所述的基于OPENCV的灌注图像判定方法。10 . A computer-readable storage medium, comprising instructions that, when executed on a computer, cause the computer to execute the OPENCV-based perfusion image determination method according to any one of claims 1 to 7 .
CN201911314612.7A 2019-12-19 2019-12-19 Perfusion image judgment system and method based on OPENCV and intelligent terminal Active CN111127414B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911314612.7A CN111127414B (en) 2019-12-19 2019-12-19 Perfusion image judgment system and method based on OPENCV and intelligent terminal

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911314612.7A CN111127414B (en) 2019-12-19 2019-12-19 Perfusion image judgment system and method based on OPENCV and intelligent terminal

Publications (2)

Publication Number Publication Date
CN111127414A CN111127414A (en) 2020-05-08
CN111127414B true CN111127414B (en) 2022-02-22

Family

ID=70500177

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911314612.7A Active CN111127414B (en) 2019-12-19 2019-12-19 Perfusion image judgment system and method based on OPENCV and intelligent terminal

Country Status (1)

Country Link
CN (1) CN111127414B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113827263B (en) * 2021-11-08 2025-02-07 上海联影智能医疗科技有限公司 Perfusion image processing method, system, electronic device and storage medium
CN115931811B (en) * 2023-03-09 2023-06-09 良渚实验室 High-flux neural loop analysis method and system

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105997128A (en) * 2016-08-03 2016-10-12 上海联影医疗科技有限公司 Method and system for recognizing focus of infection by perfusion imaging
CN109509186A (en) * 2018-11-09 2019-03-22 北京邮电大学 Cerebral arterial thrombosis lesion detection method and device based on brain CT image

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8532353B2 (en) * 2011-11-23 2013-09-10 Vital Images, Inc. Synthetic visualization and quantification of perfusion-related tissue viability

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105997128A (en) * 2016-08-03 2016-10-12 上海联影医疗科技有限公司 Method and system for recognizing focus of infection by perfusion imaging
CN109509186A (en) * 2018-11-09 2019-03-22 北京邮电大学 Cerebral arterial thrombosis lesion detection method and device based on brain CT image

Also Published As

Publication number Publication date
CN111127414A (en) 2020-05-08

Similar Documents

Publication Publication Date Title
Xu et al. A deep convolutional neural network for classification of red blood cells in sickle cell anemia
Harrison et al. Progressive and multi-path holistically nested neural networks for pathological lung segmentation from CT images
Zhang et al. Automated semantic segmentation of red blood cells for sickle cell disease
Mehanian et al. Computer-automated malaria diagnosis and quantitation using convolutional neural networks
US10853449B1 (en) Report formatting for automated or assisted analysis of medical imaging data and medical diagnosis
WO2020260936A1 (en) Medical image segmentation using an integrated edge guidance module and object segmentation network
Faria et al. Content-based image retrieval for brain MRI: An image-searching engine and population-based analysis to utilize past clinical data for future diagnosis
WO2017106645A1 (en) Interpretation and quantification of emergency features on head computed tomography
CN111047610A (en) Focal region presenting method and device
Waheed et al. Hybrid features and mediods classification based robust segmentation of blood vessels
Chen et al. Automated midline shift and intracranial pressure estimation based on brain CT images
CN111127414B (en) Perfusion image judgment system and method based on OPENCV and intelligent terminal
Wang et al. Measurement system of gear parameters based on machine vision
George et al. Automatic psoriasis lesion segmentation in two-dimensional skin images using multiscale superpixel clustering
Shyla et al. RETRACTED ARTICLE: Glaucoma detection and classification using modified level set segmentation and pattern classification neural network
Chang et al. Modified YOLO network model for metaphase cell detection in antinuclear antibody images
Kumar et al. Automatic detection of red lesions in digital color retinal images
Chen et al. Application of artificial intelligence in tongue diagnosis of traditional Chinese medicine: a review
Pal et al. A fully connected reproducible SE-UResNet for multiorgan chest radiographs segmentation
Kim et al. Automatic localization of anatomical landmarks in cardiac MR perfusion using random forests
Rasel et al. Bluish veil detection and lesion classification using custom deep learnable layers with explainable artificial intelligence (XAI)
Bidaki et al. Lesion area assessment in psoriasis patients
Jiji et al. An intelligent technique for detecting Alzheimer's disease based on brain structural changes and hippocampal shape
Chaichana et al. Automatic location of blood vessel bifurcations in digital eye fundus images
Kulyabin et al. Generalist Segmentation Algorithm for Photoreceptors Analysis in Adaptive Optics Imaging

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant