CN116228802B - Cardiac MRI auxiliary imaging control method - Google Patents
Cardiac MRI auxiliary imaging control method Download PDFInfo
- Publication number
- CN116228802B CN116228802B CN202310491678.3A CN202310491678A CN116228802B CN 116228802 B CN116228802 B CN 116228802B CN 202310491678 A CN202310491678 A CN 202310491678A CN 116228802 B CN116228802 B CN 116228802B
- Authority
- CN
- China
- Prior art keywords
- period
- magnetic resonance
- nuclear magnetic
- screened
- images
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration by the use of more than one image, e.g. averaging, subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
- G06T2207/20152—Watershed segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
- G06T2207/20192—Edge enhancement; Edge preservation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- 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
- G06T2207/30048—Heart; Cardiac
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/30—Assessment of water resources
Abstract
The invention relates to the field of intelligent control, in particular to a heart MRI auxiliary imaging control method. Continuously acquiring a plurality of cardiac nuclear magnetic resonance images, dividing the cardiac nuclear magnetic resonance images into a plurality of heartbeat periods, selecting a central period in the plurality of heartbeat periods, and acquiring a period to be screened of each central period; matching the central period with the corresponding cardiac nuclear magnetic resonance image in the period to be screened; obtaining a reference value of a period to be screened; taking the period to be screened with the reference value larger than the threshold value as the reference period of each center period; and acquiring a quality parameter of each central period, acquiring the central period with the maximum quality parameter as an optimal period, and acquiring the heart nuclear magnetic resonance image overlapped in the optimal period as an optimal heart nuclear magnetic resonance image. According to the invention, the center period and the reference period are selected for matching, and the matched images are correspondingly overlapped, so that the edges in the images can be accurately obtained.
Description
Technical Field
The invention relates to the field of intelligent control, in particular to a heart MRI auxiliary imaging control method.
Background
The cardiac nuclear magnetic resonance image plays a very great role in understanding the condition of a patient in medical aspects, but continuous beating of the heart and high-speed movement of blood often form artifacts and uneven gray scales, and the boundaries of different areas are difficult to accurately obtain due to the complex structures of different organs in the heart, so that the obtained image effect is poor, great influence is brought to assisting doctors in consulting the cardiac MRI image of the patient, and therefore, the cardiac nuclear magnetic resonance image needs to be subjected to auxiliary imaging, so that the artifacts in the nuclear magnetic resonance image are removed, and the edge area is obtained.
In the prior art, the method for processing the images, such as threshold segmentation, K-means clustering, graph segmentation, ACM/ASM and fuzzy set theory, is poor in effect when the image segmentation is actually performed due to complex structures in body tissues, such as very similar image gray levels of cardiac muscle and surrounding tissues, and is difficult to find accurate edge lines, and a good edge segmentation effect can be realized through a neural network based on multi-scale edge perception, but the neural network needs a large number of data sets for training, so that the efficiency in actual use is not high.
Disclosure of Invention
In order to solve the problem of poor segmentation effect of conventional image segmentation on cardiac nuclear magnetic resonance images in the prior art, the invention provides a cardiac MRI auxiliary imaging control method, which comprises the following steps: continuously acquiring a plurality of cardiac nuclear magnetic resonance images, dividing the cardiac nuclear magnetic resonance images into a plurality of heartbeat periods, selecting a central period in the plurality of heartbeat periods, and acquiring a period to be screened of each central period; matching the central period with the corresponding cardiac nuclear magnetic resonance image in the period to be screened; obtaining a reference value of a period to be screened; taking the period to be screened with the reference value larger than the threshold value as the reference period of each center period; and acquiring a quality parameter of each central period, acquiring the central period with the maximum quality parameter as an optimal period, and acquiring the heart nuclear magnetic resonance image overlapped in the optimal period as an optimal heart nuclear magnetic resonance image. According to the invention, the center period and the reference period are selected for matching, and the matched images are correspondingly overlapped, so that the edges in the images can be accurately acquired, and the optimal cardiac nuclear magnetic resonance image is obtained.
The invention adopts the following technical scheme that the heart MRI auxiliary imaging control method comprises the following steps:
collecting images; a plurality of cardiac nmr images are acquired sequentially.
Selecting a period to be screened; dividing a plurality of continuous cardiac nuclear magnetic resonance images into a plurality of heart cycles, selecting a central cycle in the heart cycles, and acquiring the heart cycles in the adjacent setting range of each central cycle as the cycles to be screened of the central cycle.
Matching images; the heart nuclear magnetic resonance image of each central period is respectively matched with the heart nuclear magnetic resonance image of the period to be screened corresponding to the central period, and the matching relation between the heart nuclear magnetic resonance image of each central period and the heart nuclear magnetic resonance image of the period to be screened corresponding to the central period is obtained; and acquiring the number of the cardiac nuclear magnetic resonance images in each period to be screened, wherein the cardiac nuclear magnetic resonance images in each period to be screened and the corresponding central period accord with a set matching relation.
Screening a reference period; acquiring a reference value of each period to be screened according to the heart nuclear magnetic resonance quantity of each period to be screened, which accords with a set matching relation with the heart nuclear magnetic resonance image of the corresponding center period; and taking the period to be screened, of which the reference value is larger than the threshold value in the period to be screened in each central period, as the reference period of each central period.
Determining an optimal period; and acquiring the quality parameter of each center period according to the number of the heart nuclear magnetic resonance images, which accord with the set matching relation in the corresponding reference period, of each heart nuclear magnetic resonance image in each center period, and acquiring the center period with the maximum quality parameter as the optimal period.
Obtaining an optimal image; and superposing each cardiac nuclear magnetic resonance image in the optimal period with the cardiac nuclear magnetic resonance image which accords with the set matching relation in the corresponding reference period to obtain the optimal cardiac nuclear magnetic resonance image.
Further, a method for controlling cardiac MRI assisted imaging, which divides a plurality of continuous cardiac nmr images into a plurality of heart cycles, comprises:
the heart cycle is the time taken from the start of one heart beat to the start of the next heart beat;
dividing the acquired continuous multiple cardiac nuclear magnetic resonance images according to the time of each heartbeat cycle to obtain multiple heartbeat cycles; each heart cycle contains a plurality of cardiac nmr images.
Further, a method for matching the cardiac MRI auxiliary imaging control method for each central period with the cardiac MRI image in each period to be screened corresponding to the central period includes:
performing factor analysis on the acquired continuous multiple nuclear magnetic resonance images to acquire independent factor vectors of each nuclear magnetic resonance image;
acquiring cosine similarity mean values between independent factor vectors of each nuclear magnetic resonance image in each central period and independent factor vectors of all nuclear magnetic resonance images in each period to be screened in the central period, and taking the cosine similarity mean values as node values of each nuclear magnetic resonance image in the corresponding central period in matching;
and similarly, obtaining the cosine similarity mean value between the independent factor of each nuclear magnetic resonance image in each period to be screened and the independent factors of all nuclear magnetic resonance images in the corresponding center period, and taking the cosine similarity mean value as the node value of each nuclear magnetic resonance image in matching in each period to be screened.
Further, a method for performing factor analysis on a plurality of continuous nuclear magnetic resonance images by using the heart MRI auxiliary imaging control method comprises the following steps:
converting each nuclear magnetic resonance image into a vector format in a sequence from left to right and from top to bottom;
factor analysis is carried out on vectors corresponding to all nuclear magnetic resonance images, and common factor vectors of the vectors corresponding to all nuclear magnetic resonance images are obtained;
and subtracting the common factor vector from the vector corresponding to each nuclear magnetic resonance image to obtain an independent factor vector of each nuclear magnetic resonance image.
Further, a method for controlling cardiac MRI auxiliary imaging, the method for obtaining the reference value of each period to be screened corresponding to each center period is as follows:
acquiring a time difference value of each center period and each period to be screened corresponding to the center period;
acquiring the number of nuclear magnetic resonance images in each central period and the number of nuclear magnetic resonance images in each period to be screened corresponding to each central period;
acquiring the duty ratio of the number of all nuclear magnetic resonance images in each period to be screened by using the number of the cardiac nuclear magnetic resonance images, in which each period to be screened and the cardiac nuclear magnetic resonance image in the corresponding center period accord with a set matching relation;
acquiring the product of the minimum value of the number of nuclear magnetic resonance images between each period to be screened and the corresponding central period and the time difference between the period to be screened and the corresponding central period;
and obtaining a reference value of each period to be screened according to the ratio of the heart nuclear magnetic resonance image quantity ratio of each period to be screened and the heart nuclear magnetic resonance image in the corresponding center period, which accords with a set matching relation, to the product.
Further, a method for superposing each cardiac nuclear magnetic resonance image in the optimal period and the cardiac nuclear magnetic resonance image in the corresponding reference period according with a set matching relation comprises the following steps:
converting the nuclear magnetic resonance image in the optimal period into a binary image in the reference period corresponding to the optimal period;
acquiring connected domains in each nuclear magnetic resonance binary image by using a watershed segmentation method, and constructing an adjacency graph structure of each nuclear magnetic resonance image according to the connected domains in each nuclear magnetic resonance binary image;
overlapping the adjacent graph structure of each nuclear magnetic resonance image in the optimal period with the adjacent graph structure of the nuclear magnetic resonance image which accords with the set matching relation in the reference period corresponding to the optimal period;
and carrying out morphological operation on the adjacent graph structure of each nuclear magnetic resonance image in the optimal period after superposition.
Further, a cardiac MRI assisted imaging control method, the method for obtaining the quality parameter of each center period is as follows:
acquiring the number of the cardiac nuclear magnetic resonance images, which accord with a set matching relationship in a corresponding reference period, of each cardiac nuclear magnetic resonance image in each central period;
acquiring the number of nuclear magnetic resonance images in each center period;
acquiring the product of the number of the cardiac nuclear magnetic resonance images, which accords with a set matching relation in the corresponding reference period, of each cardiac nuclear magnetic resonance image in each central period and the number of the nuclear magnetic resonance images in the corresponding central period;
and obtaining the quality parameter of each center period according to the ratio of the product to the maximum nuclear magnetic resonance quantity in all the center periods.
The beneficial effects of the invention are as follows: according to the invention, a plurality of heart nuclear magnetic resonance images are continuously acquired, heartbeat periods are divided, and the set central period and the period to be screened are selected, so that the matching of each central period is facilitated, and the difference between the superimposed images is ensured to be as small as possible when the images are superimposed; when the periods to be screened of the center periods are matched, factor analysis is firstly carried out on the nuclear magnetic resonance images in each period, so that the matching relation of the nuclear magnetic resonance images among the periods is established through independent factor vectors, the similarity of the obtained characteristics of the reference period and the center period is ensured, supporting conditions are provided for image superposition of the subsequent center period and the reference period, the images are correspondingly superposed by combining the matching relation, the images are screened again according to the effects of the images after superposition of different center periods, the optimal nuclear magnetic resonance images with clear and accurate edges are finally obtained, and the edge detection effect in the nuclear magnetic resonance images is good.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.
FIG. 1 is a schematic diagram of a method for controlling cardiac MRI assisted imaging according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a heartbeat cycle stacking according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of DTW matching according to an embodiment of the invention;
fig. 4 is a schematic diagram of an adjacency graph according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in fig. 1, a schematic structural diagram of a cardiac MRI auxiliary imaging control method according to an embodiment of the present invention is provided, including:
101. collecting images; a plurality of cardiac nmr images are acquired sequentially.
The object of the invention is to obtain the more accurate demarcation between the different organs of the heart region on the MRI image, so that firstly the MRI image of the heart needs to be acquired, and then the heart region is obtained by semantic segmentation.
102. Selecting a period to be screened; dividing a plurality of continuous cardiac nuclear magnetic resonance images into a plurality of heart cycles, selecting a central cycle in the heart cycles, and acquiring the heart cycles in the adjacent setting range of each central cycle as the cycles to be screened of the central cycle.
The cardiac cycle refers to the process that the cardiovascular system undergoes from the start of one heartbeat to the start of the next, in which process the boundaries of different organs in the heart vary in definition on the one hand and in the location of the same boundaries on the other hand, for example: in the first frame of MRI images, the boundary line of the left atrium and the left ventricle corresponds to the pixel point set a, and in the second frame of MRI images, the boundary line of the left atrium and the left ventricle corresponds to the pixel point set B, which are unequal, because the left atrium contracts in the cardiac cycle.
The heart beat is detected by the electrocardiogram, so that the time range of each cardiac cycle can be determined by the time of each heart beat in the electrocardiogram, and the nuclear magnetic resonance image in the range is taken as an image set in the cycle, and the heart rate is not maintained at a fixed value, and is fluctuant, normally, the heart rate is maintained at 60-100 times/minute, namely, different cycles, even the time of adjacent cardiac cycles is not necessarily the same, and therefore, the number of MRI images obtained in each cycle is different.
If the time of two periods is the same, the boundary conditions of the images in the same order in the periods are similar, and there may be a smaller difference, and at this time, the probability that the edge line of the middle region at the intersection of the two edges is taken as a true edge line is greater, as shown in fig. 2, a represents one edge of the first MRI image in the first period, B represents the corresponding edge in the first MRI image in the second period, the situation that the two are overlapped through the position information is shown in the middle diagram in fig. 2, the intersection region after the two are overlapped is the region C, and the center line of the region C is taken as the edge better, because the property of the edge C is the most stable, and is the least possible false edge caused by heart beat.
The method for dividing the continuous multi-heart nuclear magnetic resonance images into a plurality of heartbeat cycles comprises the following steps:
the heart cycle is the time taken from the start of one heart beat to the start of the next heart beat;
dividing the acquired continuous multiple cardiac nuclear magnetic resonance images according to the time of each heartbeat cycle to obtain multiple heartbeat cycles; each heart cycle contains a plurality of cardiac nmr images.
In the invention, the heartbeat cycle in the range of 7 left and right heartbeat cycles is taken as a calculation cycle, so that when a center cycle is selected, the eighth cycle to the eighth last heartbeat cycle in all the heartbeat cycles are taken as the center cycle, and the cycle which is similar to the edge change condition of the center cycle is calculated in the 7 neighborhood of the center cycle and is taken as a reference cycle, for example: the 7 left periods of the 8 th period are the first period to the 7 th period, the 7 right periods are the 9 th period to the 15 th period, and the reference period is calculated in the 14 periods around each central period, namely, the 14 periods around each central period are taken as periods to be screened.
103. Matching images; the heart nuclear magnetic resonance image of each central period is respectively matched with the heart nuclear magnetic resonance image of the period to be screened corresponding to the central period, and the matching relation between the heart nuclear magnetic resonance image of each central period and the heart nuclear magnetic resonance image of the period to be screened corresponding to the central period is obtained; and acquiring the number of the cardiac nuclear magnetic resonance images in each period to be screened, wherein the cardiac nuclear magnetic resonance images in each period to be screened and the corresponding central period accord with a set matching relation.
Firstly, the probability of the similar period to the time of the central period to the edge change condition of the central period is larger; second, the greater the number of one-to-one matches, the greater the probability of being close to the edge variation of the center period.
The time of each cycle, which refers to the time of each cardiac cycle, can be obtained directly, for example: the time of the first cardiac cycle is 0.01s.
The one-to-one matching in the two cardiac cycles indicates the change consistency of the edge change condition in the two cardiac cycles, and the matching relationship of nuclear magnetic resonance images in the two cardiac cycles is obtained through the DTW (dynamic time warping algorithm) matching between the central cycle and the cycle to be screened, as shown in fig. 3, the matching relationship comprises one-to-one, one-to-many and many-to-one relationships, the one-to-one relationship indicates the change consistency of the two, the more the one-to-one relationship is, the more the change consistency of the two curves is, the more the one-to-many relationship is, the smaller the change consistency of the two curves is, and the DTW matching content is:
for MRI continuous images of the heart, other information in the continuous images can be considered as unchanged except for the boundary changes of different organs generated by heart beating, namely, the different information in the continuous images only comprises boundary information among the different organs.
Because factor analysis is performed on vectors, and images are in matrix format, it is first necessary to convert the images in matrix format into vector format, convert each MRI image into vector format by left-to-right and top-to-bottom order, called MRI vector, and perform factor analysis on successive MRI vectors to obtain a common factor vector, that is, all MRI vectors correspond to a common factor vector, where the common factor vector represents the same information in all MRI vectors, and each MRI vector corresponds to an independent factor vector, where each MRI vector represents the remaining information after subtracting the information represented by the common factor vector from the information contained in each MRI vector.
Taking the cosine similarity mean value of the independent factor vector of each MRI image in the central period and the cosine similarity mean value of the independent factor vectors of all MRI images in the neighborhood period as the corresponding value of the image node, as shown in fig. 3, wherein the upper curve represents the central period curve, the lower curve represents the neighborhood period curve, the first point of the central period curve represents the cosine similarity mean value of the independent factor vector of the first MRI image in the period, and the other points are the same.
The DTW matching method for the heart nuclear magnetic resonance image of each central period and the heart nuclear magnetic resonance image of each period to be screened of the central period comprises the following steps:
performing factor analysis on the acquired continuous multiple nuclear magnetic resonance images to acquire independent factor vectors of each nuclear magnetic resonance image;
acquiring cosine similarity mean values between independent factor vectors of each nuclear magnetic resonance image in each central period and independent factor vectors of all nuclear magnetic resonance images in each period to be screened in the central period, and taking the cosine similarity mean values as node values of each nuclear magnetic resonance image in the corresponding central period in DTW matching;
and similarly, obtaining the cosine similarity mean value between the independent factors of each nuclear magnetic resonance image in each period to be screened and the independent factors of all nuclear magnetic resonance images in the corresponding center period, and taking the cosine similarity mean value as the node value of each nuclear magnetic resonance image in the DTW matching in each period to be screened.
The method for performing factor analysis on the continuous multiple nuclear magnetic resonance images comprises the following steps:
converting each nuclear magnetic resonance image into a vector format in a sequence from left to right and from top to bottom;
factor analysis is carried out on vectors corresponding to all nuclear magnetic resonance images, and common factor vectors of the vectors corresponding to all nuclear magnetic resonance images are obtained;
and subtracting the common factor vector from the vector corresponding to each nuclear magnetic resonance image to obtain an independent factor vector of each nuclear magnetic resonance image.
104. Screening a reference period; acquiring a reference value of each period to be screened according to the heart nuclear magnetic resonance quantity of each period to be screened, which accords with a set matching relation with the heart nuclear magnetic resonance image of the corresponding center period; and taking the period to be screened, of which the reference value is larger than the threshold value in the period to be screened in each central period, as the reference period of each central period.
The method for acquiring the reference value of each period to be screened corresponding to each center period comprises the following steps:
acquiring a time difference value of each center period and each period to be screened corresponding to the center period;
acquiring the number of nuclear magnetic resonance images in each central period and the number of nuclear magnetic resonance images in each period to be screened corresponding to each central period;
the heart nuclear magnetic resonance images in each period to be screened and the heart nuclear magnetic resonance images in the corresponding center period accord with the duty ratio of the number of the heart nuclear magnetic resonance images in the period to be screened in all nuclear magnetic resonance images in the set matching relation, and the set matching relation is one-to-one matching;
acquiring the product of the minimum value of the number of nuclear magnetic resonance images between each period to be screened and the corresponding central period and the time difference between the period to be screened and the corresponding central period;
obtaining a reference value of each period to be screened according to the ratio of the heart nuclear magnetic resonance image quantity ratio of each period to be screened and the heart nuclear magnetic resonance image in the corresponding center period, which accords with a set matching relation, to the product, wherein the reference value is as follows:
wherein c represents the time difference between the period to be screened and the central period, a represents the ratio of the number of nuclear magnetic resonance images which are matched one by one in each central period and each period to be screened corresponding to the central period, b represents the minimum number of nuclear magnetic resonance images between the period to be screened and the corresponding central period, and in the process of calculating the DTW distance of two curves, the relationship between one by one matching between the central period node and the neighborhood period node and one by one matching can be obtained, wherein the one by one matching is the similarity node of the two curves, the one by one matching is the matching which is generated by the similarity of the two curves to the maximum, so that the one by one matching can represent that the two images of the corresponding nodes are in the same edge information change place, namely the edge information of the two images are the same, so that the larger the ratio of the one by one matching in all the matching relationships is, the larger the reference value in the period is, and the period to be screened with the reference value larger than 0.9 is reserved as the reference period of each central period.
105. Determining an optimal period; and acquiring the quality parameter of each center period according to the number of the heart nuclear magnetic resonance images, which accord with the set matching relation in the corresponding reference period, of each heart nuclear magnetic resonance image in each center period, and acquiring the center period with the maximum quality parameter as the optimal period. Since it is impossible for a doctor to analyze all MRI images in practice, it is necessary to select a period with best edge information as an optimal period, and the MRI images in the optimal period are taken as MRI images that the doctor needs to see; first, the greater the number of MRI images in each central cycle, the better the quality of that cycle, since the physician can identify the patient's condition from the information in more images; and secondly, the edges in each MRI image in the central period are obtained by overlapping a plurality of matching images, and the more times of overlapping, the obtained edges are closer to real edge information.
The number of MRI images per cycle can be directly obtained, and the number of overlaps per image refers to how many one-to-one matching intersections the edges in the final image are calculated from, for example: the MRI images in central cycle 1 are: [ a1, a2, a3], the MRI images in reference period 2 are: [ b1, b2, b3, b4], the MRI images in reference period 3 are: [ c1, c2, c3], with cycle 1 as the center cycle, cycle 2 and cycle 3 as the reference cycles of cycle 1, the matching relationship in cycle 1 and cycle 2 is: a1-b1, a2-b2, a3- (b 3, b 4), the matching relationship in period 1 and period 3 is: a1-c1, a2-c2, a3-c3, the number of overlapping a1 in period 1 is 2, i.e. a1 is overlapped with b1, c1 respectively; the number of overlapping a2 is 2, and a3 is not overlapped with the reference period 2 because there is no one-to-one matching in the matching relation with the reference period 2, so the number of overlapping a3 is 1, and it is obvious that the more the number of overlapping is, the closer the edge finally obtained is to the real edge.
Acquiring the number of the cardiac nuclear magnetic resonance images, which accord with a set matching relationship in a corresponding reference period, of each cardiac nuclear magnetic resonance image in each central period;
acquiring the number of nuclear magnetic resonance images in each center period;
acquiring the product of the number of the cardiac nuclear magnetic resonance images, which accords with a set matching relation in the corresponding reference period, of each cardiac nuclear magnetic resonance image in each central period and the number of the nuclear magnetic resonance images in the corresponding central period;
and obtaining the quality parameter of each center period according to the ratio of the product to the maximum nuclear magnetic resonance number in all the center periods, wherein the expression is as follows:
wherein, the liquid crystal display device comprises a liquid crystal display device,
a quality parameter representing the jth center period,
representing the number of nmr images in the jth central cycle,
indicating the number of times of superposition of the ith nuclear magnetic resonance image in the jth center period,
representing the number of maximum nmr images in all central periods,
representing the normalized image quantity value for each cycle, the greater the value, the more the cycle should be selected;
the sum of the number of superimposed times of all MRI images in each period is represented, the larger the value, the more the corresponding period should be selected.
After the quality parameter of each period is obtained by calculation, the period corresponding to the maximum quality parameter is selected as the optimal period, and after the optimal period is obtained, each nuclear magnetic resonance image in the period after the edge enhancement is obtained for assisting a doctor, the whole edge information of the image in the optimal period is the best, and the periodic image can play the largest auxiliary role in diagnosis and the like of the doctor.
106. Obtaining an optimal image; and superposing each cardiac nuclear magnetic resonance image in the optimal period with the cardiac nuclear magnetic resonance image which accords with the set matching relation in the corresponding reference period to obtain the optimal cardiac nuclear magnetic resonance image.
Firstly, carrying out otsu threshold segmentation on each MRI image to obtain a binary image, then obtaining different connected domains through a watershed segmentation method, wherein the boundaries between the different connected domains are of multiple line widths, and then constructing an adjacent graph structure for each MRI image, wherein nodes in the adjacent graph structure represent each connected domain, namely e, f, g and h represent four connected domains respectively, and the edges connecting the connected domains represent the common boundary between every two connected domains as shown in fig. 4; as can be seen from the description of the left image in fig. 4, the connected domain e and the connected domain f and the connected domain g have a common boundary, the connected domain f and the connected domain e, the connected domain g and the connected domain h have a common boundary, the connected domain g and the connected domain e, the connected domain f and the connected domain h have a common boundary, the connected domain h and the connected domain f and the connected domain g are converted into the right adjacent image in fig. 4 according to the content, the main basis of the conversion is that the conversion is performed according to the common boundary between each connected domain and other connected domains, and the right image in fig. 4 is converted by the left image according to the principle;
converting the nuclear magnetic resonance image in the optimal period into a binary image in the reference period corresponding to the optimal period;
acquiring connected domains in each nuclear magnetic resonance binary image by using a watershed segmentation method, and constructing an adjacency graph structure of each nuclear magnetic resonance image according to the connected domains in each nuclear magnetic resonance binary image;
overlapping the adjacent graph structure of each nuclear magnetic resonance image in the optimal period with the adjacent graph structure of the nuclear magnetic resonance image which accords with the set matching relation in the reference period corresponding to the optimal period;
and carrying out morphological operation on the adjacent graph structure of each nuclear magnetic resonance image in the optimal period after superposition.
The structure of the adjacent graph of each MRI image can be obtained through calculation, one center period corresponds to a plurality of reference periods, the overlapping intersection areas after overlapping (multiple pixel widths) are calculated through overlapping the images, the central line of the intersection areas is used as a corresponding real edge, namely, overlapping of each image and a matched image in the center period is calculated, the central line of the intersection area of each boundary line is obtained through a morphological corrosion method, updating of each edge in fig. 4 is completed, and finally an accurate edge area of each heart nuclear magnetic resonance image is obtained, so that the optimal heart nuclear magnetic resonance image is obtained.
According to the invention, a plurality of heart nuclear magnetic resonance images are continuously acquired, heartbeat periods are divided, and the set central period and the period to be screened are selected, so that the matching of each central period is facilitated, and the difference between the superimposed images is ensured to be as small as possible when the images are superimposed; when the periods to be screened of the center periods are matched, factor analysis is firstly carried out on the nuclear magnetic resonance images in each period, so that the matching relation of the nuclear magnetic resonance images among the periods is established through independent factor vectors, the similarity of the obtained characteristics of the reference period and the center period is ensured, supporting conditions are provided for image superposition of the subsequent center period and the reference period, the images are correspondingly superposed by combining the matching relation, the images are screened again according to the effects of the images after superposition of different center periods, the optimal nuclear magnetic resonance images with clear and accurate edges are finally obtained, and the edge detection effect in the nuclear magnetic resonance images is good.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (7)
1. A cardiac MRI assisted imaging control method, comprising:
collecting images; continuously acquiring a plurality of cardiac nuclear magnetic resonance images;
selecting a period to be screened; dividing a plurality of continuous cardiac nuclear magnetic resonance images into a plurality of heart beat periods, selecting a central period in the heart beat periods, and acquiring the heart beat period in a set range adjacent to each central period as a period to be screened of the central period;
matching images; the heart nuclear magnetic resonance image of each central period is respectively matched with the heart nuclear magnetic resonance image of the period to be screened corresponding to the central period, and the matching relation between the heart nuclear magnetic resonance image of each central period and the heart nuclear magnetic resonance image of the period to be screened corresponding to the central period is obtained; acquiring the number of the cardiac nuclear magnetic resonance images in each period to be screened, which accords with a set matching relation with the cardiac nuclear magnetic resonance images in the corresponding center period;
screening a reference period; acquiring a reference value of each period to be screened according to the heart nuclear magnetic resonance quantity of each period to be screened, which accords with a set matching relation with the heart nuclear magnetic resonance image of the corresponding center period; taking the period to be screened, of which the reference value is larger than the threshold value in the period to be screened in each central period, as the reference period of each central period;
determining an optimal period; acquiring a quality parameter of each center period according to the number of the heart nuclear magnetic resonance images, which accord with a set matching relation in the corresponding reference period, of each heart nuclear magnetic resonance image in each center period, and acquiring the center period with the maximum quality parameter as an optimal period;
obtaining an optimal image; and superposing each cardiac nuclear magnetic resonance image in the optimal period with the cardiac nuclear magnetic resonance image which accords with the set matching relation in the corresponding reference period to obtain the optimal cardiac nuclear magnetic resonance image.
2. The method of claim 1, wherein the method of dividing the continuous plurality of cardiac MRI images into a plurality of heart cycles comprises:
the heart cycle is the time taken from the start of one heart beat to the start of the next heart beat;
dividing the acquired continuous multiple cardiac nuclear magnetic resonance images according to the time of each heartbeat cycle to obtain multiple heartbeat cycles; each heart cycle contains a plurality of cardiac nmr images.
3. The method for controlling cardiac MRI-assisted imaging according to claim 1, wherein the method for matching the cardiac MRI image of each central cycle with the cardiac MRI image of each period to be screened corresponding to the central cycle comprises:
performing factor analysis on the acquired continuous multiple nuclear magnetic resonance images to acquire independent factor vectors of each nuclear magnetic resonance image;
acquiring cosine similarity mean values between independent factor vectors of each nuclear magnetic resonance image in each central period and independent factor vectors of all nuclear magnetic resonance images in each period to be screened in the central period, and taking the cosine similarity mean values as node values of each nuclear magnetic resonance image in the corresponding central period in matching;
and similarly, obtaining the cosine similarity mean value between the independent factor of each nuclear magnetic resonance image in each period to be screened and the independent factors of all nuclear magnetic resonance images in the corresponding center period, and taking the cosine similarity mean value as the node value of each nuclear magnetic resonance image in matching in each period to be screened.
4. A method of cardiac MRI-assisted imaging control according to claim 3, characterized in that the method of factor analysis of a plurality of successive nuclear magnetic resonance images comprises:
converting each nuclear magnetic resonance image into a vector format in a sequence from left to right and from top to bottom;
factor analysis is carried out on vectors corresponding to all nuclear magnetic resonance images, and common factor vectors of the vectors corresponding to all nuclear magnetic resonance images are obtained;
and subtracting the common factor vector from the vector corresponding to each nuclear magnetic resonance image to obtain an independent factor vector of each nuclear magnetic resonance image.
5. The method for controlling cardiac MRI-assisted imaging according to claim 1, wherein the method for acquiring the reference value of each period to be screened corresponding to each center period comprises:
acquiring a time difference value of each center period and each period to be screened corresponding to the center period;
acquiring the number of nuclear magnetic resonance images in each central period and the number of nuclear magnetic resonance images in each period to be screened corresponding to each central period;
acquiring the duty ratio of the number of all nuclear magnetic resonance images in each period to be screened by using the number of the cardiac nuclear magnetic resonance images, in which each period to be screened and the cardiac nuclear magnetic resonance image in the corresponding center period accord with a set matching relation;
acquiring the product of the minimum value of the number of nuclear magnetic resonance images between each period to be screened and the corresponding central period and the time difference between the period to be screened and the corresponding central period;
and obtaining a reference value of each period to be screened according to the ratio of the heart nuclear magnetic resonance image quantity ratio of each period to be screened and the heart nuclear magnetic resonance image in the corresponding center period, which accords with a set matching relation, to the product.
6. The method for controlling cardiac MRI-assisted imaging according to claim 1, wherein the method for superimposing each cardiac MRI image in the optimal period and the cardiac MRI image in the corresponding reference period according to the set matching relationship comprises:
converting the nuclear magnetic resonance image in the optimal period into a binary image in the reference period corresponding to the optimal period;
acquiring connected domains in each nuclear magnetic resonance binary image by using a watershed segmentation method, and constructing an adjacency graph structure of each nuclear magnetic resonance image according to the connected domains in each nuclear magnetic resonance binary image;
overlapping the adjacent graph structure of each nuclear magnetic resonance image in the optimal period with the adjacent graph structure of the nuclear magnetic resonance image which accords with the set matching relation in the reference period corresponding to the optimal period;
and carrying out morphological operation on the adjacent graph structure of each nuclear magnetic resonance image in the optimal period after superposition.
7. The method for cardiac MRI-assisted imaging control of claim 1, wherein the method for obtaining the quality parameter for each center period comprises:
acquiring the number of the cardiac nuclear magnetic resonance images, which accord with a set matching relationship in a corresponding reference period, of each cardiac nuclear magnetic resonance image in each central period;
acquiring the number of nuclear magnetic resonance images in each center period;
acquiring the product of the number of the cardiac nuclear magnetic resonance images, which accords with a set matching relation in the corresponding reference period, of each cardiac nuclear magnetic resonance image in each central period and the number of the nuclear magnetic resonance images in the corresponding central period;
and obtaining the quality parameter of each center period according to the ratio of the product to the maximum nuclear magnetic resonance quantity in all the center periods.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310491678.3A CN116228802B (en) | 2023-05-05 | 2023-05-05 | Cardiac MRI auxiliary imaging control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310491678.3A CN116228802B (en) | 2023-05-05 | 2023-05-05 | Cardiac MRI auxiliary imaging control method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116228802A CN116228802A (en) | 2023-06-06 |
| CN116228802B true CN116228802B (en) | 2023-07-04 |
Family
ID=86579049
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310491678.3A Active CN116228802B (en) | 2023-05-05 | 2023-05-05 | Cardiac MRI auxiliary imaging control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116228802B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102525468A (en) * | 2012-01-13 | 2012-07-04 | 大连灵动科技发展有限公司 | Cerebral function imaging-based method for analyzing intracerebral activation trend |
| CN103020945A (en) * | 2011-09-21 | 2013-04-03 | 中国科学院电子学研究所 | Remote sensing image registration method of multi-source sensor |
| CN106203255A (en) * | 2016-06-24 | 2016-12-07 | 华中科技大学 | A kind of pedestrian based on time unifying heavily recognition methods and system |
| CN106709919A (en) * | 2015-11-17 | 2017-05-24 | 上海联影医疗科技有限公司 | Method for processing cardiac perfusion magnetic resonance image |
| CN109001660A (en) * | 2018-06-12 | 2018-12-14 | 上海联影医疗科技有限公司 | Film imaging method and magnetic resonance imaging system |
| CN109754448A (en) * | 2018-12-29 | 2019-05-14 | 深圳安科高技术股份有限公司 | A kind of CT heart scanning artifact correction method and its system |
| CN110599499A (en) * | 2019-08-22 | 2019-12-20 | 四川大学 | MRI image heart structure segmentation method based on multipath convolutional neural network |
| EP3951705A1 (en) * | 2020-08-04 | 2022-02-09 | Siemens Healthcare GmbH | Method, device and system for determining abnormality in myocardium region |
| CN114746008A (en) * | 2019-12-02 | 2022-07-12 | 美敦力公司 | Generating representative cardiac information |
| CN115880511A (en) * | 2022-12-27 | 2023-03-31 | 凌云光技术股份有限公司 | Image screening method, device and system and storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11393136B2 (en) * | 2019-03-25 | 2022-07-19 | Canon Medical Systems Corporation | Image reconstruction method and reconstruction apparatus |
-
2023
- 2023-05-05 CN CN202310491678.3A patent/CN116228802B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020945A (en) * | 2011-09-21 | 2013-04-03 | 中国科学院电子学研究所 | Remote sensing image registration method of multi-source sensor |
| CN102525468A (en) * | 2012-01-13 | 2012-07-04 | 大连灵动科技发展有限公司 | Cerebral function imaging-based method for analyzing intracerebral activation trend |
| CN106709919A (en) * | 2015-11-17 | 2017-05-24 | 上海联影医疗科技有限公司 | Method for processing cardiac perfusion magnetic resonance image |
| CN106203255A (en) * | 2016-06-24 | 2016-12-07 | 华中科技大学 | A kind of pedestrian based on time unifying heavily recognition methods and system |
| CN109001660A (en) * | 2018-06-12 | 2018-12-14 | 上海联影医疗科技有限公司 | Film imaging method and magnetic resonance imaging system |
| CN109754448A (en) * | 2018-12-29 | 2019-05-14 | 深圳安科高技术股份有限公司 | A kind of CT heart scanning artifact correction method and its system |
| CN110599499A (en) * | 2019-08-22 | 2019-12-20 | 四川大学 | MRI image heart structure segmentation method based on multipath convolutional neural network |
| CN114746008A (en) * | 2019-12-02 | 2022-07-12 | 美敦力公司 | Generating representative cardiac information |
| EP3951705A1 (en) * | 2020-08-04 | 2022-02-09 | Siemens Healthcare GmbH | Method, device and system for determining abnormality in myocardium region |
| CN115880511A (en) * | 2022-12-27 | 2023-03-31 | 凌云光技术股份有限公司 | Image screening method, device and system and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| MRI在乳腺癌新辅助化疗中的研究进展;周世凯 等;《现代医药卫生》;第2294-2297页 * |
| Patch-based 3D U-Net and transfer learning for longitudinal piglet brain segmentation on MRI;P. Coupeau 等;《 Computer Methods and Programs in Biomedicine》;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116228802A (en) | 2023-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Suzuki et al. | Extraction of left ventricular contours from left ventriculograms by means of a neural edge detector | |
| Liu et al. | Automatic whole heart segmentation using a two-stage u-net framework and an adaptive threshold window | |
| CN110120051A (en) | A kind of right ventricle automatic division method based on deep learning | |
| US11475278B2 (en) | Artificial intelligence based cardiac motion classification | |
| Liu et al. | Cardiac magnetic resonance image segmentation based on convolutional neural network | |
| CN111047607A (en) | Method for automatically segmenting coronary artery | |
| Liu et al. | A simple and effective method for detecting myocardial infarction based on deep convolutional neural network | |
| CN112950737A (en) | Fundus fluorescence radiography image generation method based on deep learning | |
| CN112819818A (en) | Image recognition module training method and device | |
| Zou et al. | A novel approach for left ventricle segmentation in tagged MRI | |
| Stalidis et al. | Model-based processing scheme for quantitative 4-D cardiac MRI analysis | |
| WO2022086910A1 (en) | Anatomically-informed deep learning on contrast-enhanced cardiac mri | |
| CN113689441A (en) | DeepLabV3 network-based left ventricle ultrasonic dynamic segmentation method | |
| CN116228802B (en) | Cardiac MRI auxiliary imaging control method | |
| Wen et al. | Analysis on SPECT myocardial perfusion imaging with a tool derived from dynamic programming to deep learning | |
| CN115841472A (en) | Method, device, equipment and storage medium for identifying high-density characteristics of middle cerebral artery | |
| CN111329467A (en) | Heart disease auxiliary detection method based on artificial intelligence | |
| Ting et al. | Multiview 3-d echocardiography image fusion with mutual information neural estimation | |
| CN114612484A (en) | Retina OCT image segmentation method based on unsupervised learning | |
| Zhu | Estimating left ventricular volume with ROI-based convolutional neural network | |
| Kermani et al. | Quantitative analysis of left ventricular performance from sequences of cardiac magnetic resonance imaging using active mesh model | |
| Bernard et al. | Measurement and quantification | |
| Suzuki et al. | Contour extraction of left ventricular cavity from digital subtraction angiograms using a neural edge detector | |
| Kanakatte et al. | 3D Cardiac Substructures Segmentation from CMRI using Generative Adversarial Network (GAN) | |
| CN113223104B (en) | Cardiac MR image interpolation method and system based on causal relationship |
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 |