CN117351526A - Intravascular ultrasound image automatic identification method for intima - Google Patents
Intravascular ultrasound image automatic identification method for intima Download PDFInfo
- Publication number
- CN117351526A CN117351526A CN202311654438.7A CN202311654438A CN117351526A CN 117351526 A CN117351526 A CN 117351526A CN 202311654438 A CN202311654438 A CN 202311654438A CN 117351526 A CN117351526 A CN 117351526A
- Authority
- CN
- China
- Prior art keywords
- point
- ellipse
- points
- pixel
- intima
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000002608 intravascular ultrasound Methods 0.000 title claims abstract description 37
- 210000004204 blood vessel Anatomy 0.000 claims abstract description 40
- 238000005070 sampling Methods 0.000 claims description 22
- 210000004026 tunica intima Anatomy 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 239000012528 membrane Substances 0.000 claims description 3
- 239000008280 blood Substances 0.000 abstract description 5
- 210000004369 blood Anatomy 0.000 abstract description 5
- 230000006870 function Effects 0.000 description 7
- 238000013136 deep learning model Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 210000001519 tissue Anatomy 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000002308 calcification Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000002792 vascular Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/14—Vascular patterns
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/457—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by analysing connectivity, e.g. edge linking, connected component analysis or slices
-
- 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
- G06T2207/10132—Ultrasound image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
Landscapes
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Vascular Medicine (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The invention provides an automatic identification method of an intima of an intravascular ultrasound image, which comprises the following steps: acquiring an intravascular ultrasound image; setting a scanning starting direction, a scanning starting point and a pixel moving step length during scanning on the image; and searching the maximum value and the minimum value of the pixel gray difference value in the scanning direction along the scanning direction from the scanning starting point according to the set moving step length, and respectively recording the corresponding pixel point coordinates into a gray difference value maximum value pixel point set and a gray difference value minimum value pixel point set. According to the invention, through 360-degree pixel-by-pixel scanning of the whole image, the pixel points at the edge of the blood vessel intima, which are close to the gray value of blood, can be effectively extracted; the interference points are filtered, curve fitting is carried out on the residual points, and the intima edge obtained by the method can be close to the real edge on the premise that the real intima edge is not easy to identify from the image, so that the identification accuracy of the intima edge is effectively improved.
Description
Technical Field
The invention belongs to the field of intravascular ultrasound image processing and analysis, and particularly relates to an intravascular ultrasound image automatic identification method for an intima.
Background
Three common methods for identifying the intima in intravascular ultrasound images are:
(1) The method for identifying the blood vessel intima based on statistics is mainly characterized in that the identification of the edge of the blood vessel intima is realized by carrying out statistical modeling on the gray level of an ultrasonic image in the blood vessel. But the accuracy of modeling is susceptible to complex features such as guide wire artifacts, tissue calcification, etc. in the image.
(2) The vascular intima identification method based on the deep learning model needs to establish a complex deep learning model, needs to carry out a large number of data marks and is limited in practical application.
(3) The method for identifying the blood vessel intima based on the active contour model algorithm comprises the steps of firstly, establishing and solving a complex energy functional equation, and carrying out a complex solving process; second, the intima of the vessel is identified using an active contour model, requiring a given initial contour. Among the existing initial contour extraction methods, there is a method that cluster-segments an ultrasonic image, extracts a closed region meeting a condition, expands polar coordinates of the region, and converts an edge which is possibly disconnected into a closed curve by combining an interpolation algorithm as an initial contour.
In the method for identifying the intima of the blood vessel by using the active contour model, the method for extracting the initial contour by using the clustering-based method has the following defects: (1) When the gray value corresponding to the intima tissue of the blood vessel in the image is close to the gray value corresponding to the blood region and the intima edge of the blood vessel exists nearby, the clustering method easily recognizes the intima edge of the blood vessel as the intima edge of the blood vessel; (2) The clustering segmentation needs to repeatedly debug a plurality of clustering parameters to obtain an ideal segmentation result; (3) In an intravascular ultrasound image, when a non-intima edge region with a gray value close to that of a pixel region corresponding to the real intima edge exists near the real intima edge, the clustering method often has difficulty in correctly segmenting two partial regions, so that a larger error exists in an initial contour.
The invention is made in order to overcome the defect that the clustering method cannot perform correct clustering when the pixel gray value of the image area corresponding to the edge of the real blood vessel intima is close to the pixel gray value of the image area corresponding to the surrounding blood in the intravascular ultrasound image.
Disclosure of Invention
The invention provides an automatic identification method of an intima of an intravascular ultrasound image, which aims to solve at least one technical problem.
To solve the above problems, as one aspect of the present invention, there is provided an automatic identification method of an intima of an intravascular ultrasound image, comprising the steps of:
step 1, acquiring an intravascular ultrasound image;
step 2, setting a scanning starting direction, a scanning starting point and a pixel moving step length during scanning on the image;
step 3, searching the maximum value and the minimum value of the pixel gray difference value in the scanning direction along the scanning direction from the scanning starting point according to the set moving step length, and respectively recording the corresponding pixel point coordinates into a gray difference value maximum value pixel point set maxGradPoints and a gray difference value minimum value pixel point set minGradPoints;
step 4, increasing the scanning direction by 1 degree, and repeating the steps 3-4 until the scanning within the range of 360 degrees is completed;
step 5, obtaining a once fitting ellipse by combining a random sampling consistency algorithm with a least square method according to the point set minGradPoints;
step 6, extracting the points which are most matched with the primary fitting ellipse in the point sets minGradPoints and maxGradPoints to form a secondary fitting point set;
step 7, obtaining a quadratic fit ellipse by combining a random sampling consistency algorithm with a least square method according to the quadratic fit point set;
step 8, comparing the long and short axes and the center points of the first fitting ellipse and the second fitting ellipse, and selecting a fitting ellipse with larger area when the distance difference between the center points is smaller than the specified distance and the distance difference between the long and short axes is smaller than the specified length, otherwise, taking the second fitting ellipse as the final fitting ellipse;
and 9, dispersing the final fitting ellipse into 360 points, extracting points with the distances between the gradPoints point set and the 360 points being smaller than a preset value so as to obtain a discrete point set similar to the ellipse, and obtaining the edge profile of the blood vessel intima according to the discrete point set.
Preferably, in step 5, obtaining a fitted ellipse by combining a random sampling consistency algorithm with a least square method includes:
step 51, randomly selecting a predetermined number of points from the point set;
step 52, calculating an elliptic equation according to the least square principle, and calculating two focus coordinates corresponding to the ellipse;
step 53, calculating the sum of the distances from the points in the point set minGradPoints to the two focuses as the distance from the points to the ellipse;
and 54, when the distance from the point to the ellipse meets the threshold value, classifying the point as an interior point, and calculating the number of the interior points.
Preferably, in step 5, obtaining a fitted ellipse by combining a random sampling consistency algorithm with a least square method further includes:
step 55, randomly selecting a predetermined number of points from the minGradPoints point set again, repeating the steps 51-54, and repeating the iteration until the iteration times reach the set times;
and 56, selecting an ellipse corresponding to the iteration with the largest number of inner points as a final once-fit ellipse.
Preferably, the method for obtaining the quadratic fit ellipse by combining the random sampling consistency algorithm with the least square method in the step 7 is the same as the method for obtaining the once fit ellipse by combining the random sampling consistency algorithm with the least square method in the step 5.
Preferably, in step 6, extracting a point in the point set minGradPoints that most matches the first fit ellipse includes: and calculating the distance from each point in the point set minGradPoints to the primary fitting ellipse, reserving the points with the distance meeting a set threshold value, and adding the points to the secondary fitting point set.
Preferably, in step 6, extracting a point in the point set maxGradPoints that matches most with the first fitted ellipse includes:
step 61, taking the point in the maxGradPoints point set, calculating the difference between the gray value of the corresponding pixel in the intravascular ultrasound image of the point and the gray value corresponding to a certain neighborhood pixel around the point, calculating the absolute value mean value of the gray difference, and calculating the nearest distance from the point to a fitting ellipse;
and step 62, setting a mean value threshold and a distance threshold, extracting points meeting the mean value threshold and the distance threshold at the same time, and adding the points to the secondary fitting elliptical point set.
Preferably, in step 9, obtaining an edge profile of the intima of the vessel from the set of discrete points includes: and (3) designating the number of control points, uniformly sampling the discrete point set, and then performing cubic uniform spline interpolation, so as to output a smooth vascular intima edge profile.
Preferably, in step 52, calculating the elliptic equation according to the least squares principle comprises:
an ellipse with arbitrary center coordinates and rotation angles is set as follows:a plurality of measuring points are acquired as +.>；
Taking the least time of the loss function according to the least square principleAs a final fit result, wherein the loss function is:
the minimum loss function is obtained by solving the following equation setTo obtain the relative parameters of the ellipse:
。
preferably, step 5 is preceded by:
step a, calculating the minimum value of the distance from each point in the minGradPoints set to other points in the set, if the minimum value of the distance exceeds a set value, considering the point as an isolated point, and deleting the point from the minGradPoints set;
and b, counting 8 neighborhood pixel gray value distribution of the point at the corresponding position on the ultrasonic image for each pixel point in the minGradPoints set after the isolated point is deleted, and if the number of gray values in the gray distribution corresponding to the point is smaller than the specified gray threshold value and smaller than the specified number, reserving the pixel point so as to reserve the pixel point of the edge area of the blood vessel inner membrane in the blood vessel ultrasonic image as far as possible.
Preferably, in step 3, searching for the pixel gray scale difference value in the scanning direction includes:
sequentially calculating the difference of corresponding pixel gray values of the current point and the next point in the intravascular ultrasound image; when the difference exceeds a set threshold, the pixel coordinates of the current point are recorded, and the pixel coordinates of the next point are updated to the pixel coordinates of the current point, and the process is repeated until the next point reaches the image edge.
The invention can effectively extract the pixel points corresponding to the edge of the blood vessel intima from the blood vessel ultrasonic image by scanning the whole image 360 degrees pixel by pixel, and the gray value corresponding to the part of the pixel points is close to the gray value corresponding to the blood region in the blood vessel ultrasonic image; according to the invention, the interference points are filtered, and curve fitting is carried out on the residual pixel points, so that the edge of the blood vessel intima can be identified from the image in a state that the edge of the blood vessel intima is not obvious in the blood vessel ultrasonic image, and the identification precision of the edge of the blood vessel intima is effectively improved.
Drawings
FIG. 1 schematically illustrates a schematic diagram of a vascular intima profile recognition process;
FIG. 2 schematically illustrates a flow chart of steps of a method for identifying an intima profile of a blood vessel;
fig. 3 schematically shows a result one of the recognition of the intima of a blood vessel;
fig. 4 schematically shows a second result of the recognition of the intima of a blood vessel;
fig. 5 schematically shows a result three of the recognition of the intima of a blood vessel.
Detailed Description
The following describes embodiments of the invention in detail, but the invention may be practiced in a variety of different ways, as defined and covered by the claims.
As one aspect of the present invention, there is provided an intravascular ultrasound image intima automatic identification method comprising the steps of:
step 1, acquiring an intravascular ultrasound image;
step 2, setting a scanning starting direction, a scanning starting point and a pixel moving step length during scanning on the image;
step 3, searching the maximum value and the minimum value of the pixel gray difference value in the scanning direction along the scanning direction from the scanning starting point according to the set moving step length, and respectively recording the corresponding pixel point coordinates into a gray difference value maximum value pixel point set maxGradPoints and a gray difference value minimum value pixel point set minGradPoints; in consideration of the gray value difference corresponding to the vascular lumen area and the vascular intima tissue area in the intravascular ultrasound image, when the scanning mode is utilized to scan along the appointed direction, the corresponding pixel gray value can undergo the change process from dark to light to dark, and the pixel points corresponding to the maximum value and the minimum value of the gray difference generally appear at the edge position of the vascular intima on the image. Therefore, the minimum value and the maximum value of the gray value difference are extracted, and the corresponding pixel points can be ensured to appear at the edge position of the blood vessel intima as much as possible.
Step 4, increasing the scanning direction by 1 degree, and repeating the steps 3-4 until the scanning within the range of 360 degrees is completed;
step 5, obtaining a once fitting ellipse by combining a random sampling consistency algorithm with a least square method according to the point set minGradPoints;
step 6, extracting the points which are most matched with the primary fitting ellipse in the point sets minGradPoints and maxGradPoints to form a secondary fitting point set; thus, points near the once-fit ellipse can be extracted from the minGradPoints and maxGradPoints sets, so that the point-to-ellipse distance is less than the distance threshold. Wherein, the most matched point in the invention can refer to a point near the ellipse of the primary or secondary fitting, namely, a point with a distance from the ellipse smaller than a set value.
Step 7, obtaining a quadratic fit ellipse by combining a random sampling consistency algorithm with a least square method according to the quadratic fit point set;
step 8, comparing the long and short axes and the center points of the first fitting ellipse and the second fitting ellipse, and selecting a fitting ellipse with larger area when the distance difference between the center points is smaller than the specified distance and the distance difference between the long and short axes is smaller than the specified length, otherwise, taking the second fitting ellipse as the final fitting ellipse;
and 9, dispersing the final fitting ellipse into 360 points, extracting points with the distances between the gradPoints point set and the 360 points being smaller than a preset value so as to obtain a discrete point set similar to the ellipse, and obtaining the edge profile of the blood vessel intima according to the discrete point set. The gradPoints are pixel coordinates corresponding to the maximum value and the minimum value of the stored pixel gray difference when scanning is performed in the step 3, namely minGradPoints+maxGradPoints.
Preferably, in step 5, obtaining a fitted ellipse by combining a random sampling consistency algorithm with a least square method includes:
step 51, randomly selecting a predetermined number of points from the point set;
step 52, calculating an elliptic equation according to the least square principle, and calculating two focus coordinates corresponding to the ellipse;
step 53, calculating the sum of the distances from the points in the point set minGradPoints to the two focuses as the distance from the points to the ellipse;
and 54, when the distance from the point to the ellipse meets the threshold value, classifying the point as an interior point, and calculating the number of the interior points.
Preferably, in step 5, obtaining a fitted ellipse by combining a random sampling consistency algorithm with a least square method further includes:
step 55, randomly selecting a predetermined number of points from the minGradPoints point set again, repeating the steps 51-54, and repeating the iteration until the iteration times reach the set times;
and 56, selecting an ellipse corresponding to the iteration with the largest number of inner points as a final once-fit ellipse.
Preferably, the method for obtaining the quadratic fit ellipse by combining the random sampling consistency algorithm with the least square method in the step 7 is the same as the method for obtaining the once fit ellipse by combining the random sampling consistency algorithm with the least square method in the step 5.
Preferably, in step 6, extracting a point in the point set minGradPoints that most matches the first fit ellipse includes: and calculating the distance from each point in the point set minGradPoints to the primary fitting ellipse, reserving the points with the distance meeting a set threshold value, and adding the points to the secondary fitting point set.
Preferably, in step 6, extracting a point in the point set maxGradPoints that matches most with the first fitted ellipse includes:
step 61, taking the point in the maxGradPoints point set, calculating the difference between the gray value of the corresponding pixel in the intravascular ultrasound image of the point and the gray value corresponding to a certain neighborhood pixel around the point, calculating the absolute value mean value of the gray difference, and calculating the nearest distance from the point to a fitting ellipse;
and step 62, setting a mean value threshold and a distance threshold, extracting points meeting the mean value threshold and the distance threshold at the same time, and adding the points to the secondary fitting elliptical point set.
Preferably, in step 9, obtaining an edge profile of the intima of the vessel from the set of discrete points includes: and (3) designating the number of control points, uniformly sampling the discrete point set, and then performing cubic uniform spline interpolation, so as to output a smooth vascular intima edge profile.
Preferably, in step 52, calculating the elliptic equation according to the least squares principle comprises:
an ellipse with arbitrary center coordinates and rotation angles is set as follows:a plurality of measuring points are acquired as +.>；
Taking the least time of the loss function according to the least square principleAs a final fit result, wherein the loss function is:
the minimum loss function is obtained by solving the following equation setTo obtain the relative parameters of the ellipse:
。
preferably, step 5 is preceded by:
and a, calculating the minimum value of the distance from each point in the minGradPoints set to other points in the set, and if the minimum value of the distance exceeds a set value, considering the point as an isolated point and deleting the isolated point from the minGradPoints set. For example, counting the number of other points included in the set within a circular area of 3 pixels per point in the set minGradPoints, and deleting the point if the number of other points included is less than 1, so as to delete an isolated point;
and b, counting 8 neighborhood pixel gray value distribution of the point corresponding to the position on the ultrasonic image for each pixel point in the minGradPoints set after the isolated point is deleted, and if the number of gray values in the gray value distribution corresponding to the point is smaller than the specified gray threshold value is smaller than the specified number (for example, 3), reserving the pixel point so as to reserve the pixel point of the edge area of the blood vessel intima in the blood vessel ultrasonic image as much as possible.
Preferably, in step 3, searching for the pixel gray scale difference value in the scanning direction includes: sequentially calculating the difference of corresponding pixel gray values of the current point and the next point in the intravascular ultrasound image; when the difference exceeds a set threshold, the pixel coordinates of the current point are recorded, and the pixel coordinates of the next point are updated to the pixel coordinates of the current point, and the process is repeated until the next point reaches the image edge.
The invention will be further illustrated by a specific example.
Firstly, acquiring an intravascular ultrasound image, and setting a scanning starting direction, a scanning starting point and a pixel movement step length during scanning on the image;
then, starting from a scanning starting point, calculating the difference of corresponding pixel gray values of the current point and the next point in the intravascular ultrasound image according to a set moving step length along the scanning direction, recording the pixel coordinates of the current point when the difference exceeds a set threshold value, updating the pixel coordinates of the next point to the pixel coordinates of the current point, and repeating the process until the next point reaches the image edge. Then, the scanning direction is increased by 1 °, and the above-described manner is repeated until scanning within a 360 ° range is completed.
And searching the maximum value and the minimum value of the pixel gray difference value in each scanning direction, respectively recording the pixel point coordinates to a gray difference value maximum value pixel point set maxGradPoints and a gray difference value minimum value pixel point set minGradPoints, and storing the scanned pixel point coordinates into an array gradPoints set.
Then, counting the number of other points contained in the set in a circular area with 3 pixels of each point in the set minGradPoints, and deleting the point if the number of the other contained points is less than 1; and counting 8 neighborhood pixel gray value distribution of the point at the corresponding position on the ultrasonic image for each pixel point reserved by the minGradPoints in the set, and reserving the pixel point if the number of gray values in the distribution corresponding to the point is smaller than 3 and smaller than the designated gray threshold value, so as to reserve the pixel point of the edge area of the blood vessel intima in the intravascular ultrasonic image as far as possible.
Considering that the edge of the intima of a human blood vessel is similar to an ellipse under normal conditions, the invention firstly obtains an initial contour by performing ellipse fitting on pixel points extracted from an intravascular ultrasound image. However, the pixel points extracted in the above steps contain many noise points (non-edge points), and it is difficult to obtain an ideal result by direct fitting, so the following processing method is adopted:
(1) For ellipses of arbitrary center coordinates and rotation angles, it can be expressed in the following general form
A plurality of measuring points are acquired at presentLoss function according to least squares principle
And at the minimum, the fitting result is the best. In order to makeMinimum, need->The partial derivative for each term is 0, namely:
solving the equation set to obtainAnd obtaining the related parameters of the ellipse. Since the general form of an ellipse contains 6 unknown parameters, 6 observation points are selected for calculation in calculating an ellipse equation.
(2) In order to acquire a stable target ellipse from a minGradPoints point set containing noise, 6 non-collinear points are randomly selected from the point set, an ellipse equation is calculated by the method in (1), two focal coordinates corresponding to the ellipse are calculated, then the sum of the distances from the points in the minGradPoints point set to the two focal points is calculated as the distance from the points to the ellipse, and when the distance from the points to the ellipse meets a set threshold value, the points are classified as inner points.
(3) Randomly selecting 6 non-collinear points from the minGradPoints point set again, and repeatedly calculating the ellipse parameters and the number of internal points according to (1) - (2). Repeating iteration, and when the iteration times reach the set times, selecting the ellipse corresponding to the largest number of internal points as a final one-time fitting ellipse result.
After the equation of the primary fitting ellipse is obtained, the distance between each point in the point set minGradPoints and the ellipse is calculated, and the points with the distance meeting the set threshold value are reserved and added into the secondary fitting point set.
Then, taking the point in the maxGradPoints point set, calculating the difference between the gray value of the corresponding pixel of the point in the intravascular ultrasound image and the gray value corresponding to the 8 neighborhood pixel points around the point, calculating the absolute value average value of the gray difference value, simultaneously calculating the nearest distance from the point to the primary fitting ellipse, setting the average value threshold value and the distance threshold value, extracting the point meeting the two threshold values simultaneously, and adding the point to the secondary fitting point set, thereby forming a complete secondary fitting point set.
And (3) performing ellipse fitting on the point set for secondary fitting by adopting the same method as in the steps (1) - (3), so as to obtain a secondary fitting ellipse.
And then comparing the primary fitted ellipse obtained by the twice fitting with the long and short axes and the central point of the secondary fitted ellipse, and selecting the fitted ellipse with larger area when the distance difference between the central points is smaller than the specified distance and the difference between the long and short axes is smaller than the specified length, otherwise, taking the secondary fitted ellipse as the final fitting result.
Finally, dispersing the ellipse of the final fitting result into 360 points, calculating the distance from each point in the gradPoints point set to the 360 points, reserving the points with the distance smaller than 2 pixels to obtain a discrete point set similar to the ellipse, designating the number of control points, uniformly sampling the point set, then carrying out three times of uniform spline interpolation, and outputting a smooth vascular intima edge profile.
In the technical scheme, through 360-degree pixel-by-pixel scanning of the whole image, the pixel points corresponding to the edges of the blood vessel inner membranes can be effectively extracted from the blood vessel inner ultrasonic image, and the gray value corresponding to the part of the pixel points is close to the gray value corresponding to the blood region in the blood vessel inner ultrasonic image; according to the invention, the interference points are filtered, and the residual point pixel points are subjected to curve fitting, so that the intima edge of the blood vessel can be identified from the image in a state that the intima edge of the blood vessel is not obvious in the intravascular ultrasound image, and the identification accuracy of the intima edge of the blood vessel is effectively improved.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. An automatic identification method of an intravascular membrane of an intravascular ultrasound image is characterized by comprising the following steps:
step 1, acquiring an intravascular ultrasound image;
step 2, setting a scanning starting direction, a scanning starting point and a pixel moving step length during scanning on the image;
step 3, searching the maximum value and the minimum value of the pixel gray difference value in the scanning direction along the scanning direction from the scanning starting point according to the set moving step length, and respectively recording the corresponding pixel point coordinates into a gray difference value maximum value pixel point set maxGradPoints and a gray difference value minimum value pixel point set minGradPoints;
step 4, increasing the scanning direction by 1 degree, and repeating the steps 3-4 until the scanning within the range of 360 degrees is completed;
step 5, obtaining a once fitting ellipse by combining a random sampling consistency algorithm with a least square method according to the point set minGradPoints;
step 6, extracting the points which are most matched with the primary fitting ellipse in the point sets minGradPoints and maxGradPoints to form a secondary fitting point set;
step 7, obtaining a quadratic fit ellipse by combining a random sampling consistency algorithm with a least square method according to the quadratic fit point set;
step 8, comparing the long and short axes and the center points of the first fitting ellipse and the second fitting ellipse, and selecting a fitting ellipse with larger area when the distance difference between the center points is smaller than the specified distance and the distance difference between the long and short axes is smaller than the specified length, otherwise, taking the second fitting ellipse as the final fitting ellipse;
and 9, dispersing the final fitting ellipse into 360 points, extracting points with the distances between the gradPoints point set and the 360 points being smaller than a preset value so as to obtain a discrete point set similar to the ellipse, and obtaining the edge profile of the blood vessel intima according to the discrete point set.
2. The method for automatically identifying an intima of an intravascular ultrasound image according to claim 1, wherein in step 5, obtaining a fitted ellipse by a random sampling consistency algorithm in combination with a least square method comprises:
step 51, randomly selecting a predetermined number of points from the point set;
step 52, calculating an elliptic equation according to the least square principle, and calculating two focus coordinates corresponding to the ellipse;
step 53, calculating the sum of the distances from the points in the point set minGradPoints to the two focuses as the distance from the points to the ellipse;
and 54, when the distance from the point to the ellipse meets the threshold value, classifying the point as an interior point, and calculating the number of the interior points.
3. The method for automatically identifying an intima of an intravascular ultrasound image according to claim 2, wherein in step 5, obtaining a fitted ellipse by using a random sampling consistency algorithm in combination with a least square method further comprises:
step 55, randomly selecting a predetermined number of points from the minGradPoints point set again, repeating the steps 51-54, and repeating the iteration until the iteration times reach the set times;
and 56, selecting an ellipse corresponding to the iteration with the largest number of inner points as a final once-fit ellipse.
4. The method for automatically identifying an intima of an intravascular ultrasound image according to claim 3, wherein the method for obtaining a quadratic fit ellipse by combining a random sampling consistency algorithm with a least square method in step 7 is the same as the method for obtaining a once fit ellipse by combining a random sampling consistency algorithm with a least square method in step 5.
5. The method for automatically identifying an intima of an intravascular ultrasound image according to claim 1, wherein in step 6, extracting the point of the point set minGradPoints that most matches the once fitted ellipse comprises: and calculating the distance from each point in the point set minGradPoints to the primary fitting ellipse, reserving the points with the distance meeting a set threshold value, and adding the points to the secondary fitting point set.
6. The method for automatically identifying an intima of an intravascular ultrasound image according to claim 5, wherein in step 6, extracting the point of the point set maxGradPoints that most matches the once fitted ellipse comprises:
step 61, taking the point in the maxGradPoints point set, calculating the difference between the gray value of the corresponding pixel in the intravascular ultrasound image of the point and the gray value corresponding to a certain neighborhood pixel around the point, calculating the absolute value mean value of the gray difference, and calculating the nearest distance from the point to a fitting ellipse;
and step 62, setting a mean value threshold and a distance threshold, extracting points meeting the mean value threshold and the distance threshold at the same time, and adding the points to the secondary fitting elliptical point set.
7. The method for automatically identifying an intima of an intravascular ultrasound image according to claim 1, wherein in step 9, obtaining an intima edge profile from the set of discrete points comprises: and (3) designating the number of control points, uniformly sampling the discrete point set, and then performing cubic uniform spline interpolation, so as to output a smooth vascular intima edge profile.
8. The method of automatic identification of the intima of an intravascular ultrasound image according to claim 2, wherein in step 52, calculating an elliptic equation according to the least squares principle comprises:
an ellipse with arbitrary center coordinates and rotation angles is set as follows:a plurality of measuring points are acquired as +.>；
Taking the least time of the loss function according to the least square principleAs a final fit result, wherein the loss function is:
；
the minimum loss function is obtained by solving the following equation setTo obtain the relative parameters of the ellipse:
。
9. the method for automatically identifying an intima of an intravascular ultrasound image according to claim 2, further comprising, before step 5:
step a, calculating the minimum value of the distance from each point in the minGradPoints set to other points in the set, if the minimum value of the distance exceeds a set value, considering the point as an isolated point, and deleting the point from the minGradPoints set;
and b, counting 8 neighborhood pixel gray value distribution of the point at the corresponding position on the ultrasonic image for each pixel point in the minGradPoints set after the isolated point is deleted, and if the number of gray values in the gray value distribution corresponding to the point is smaller than the specified gray threshold value and smaller than the specified number, reserving the pixel point so as to reserve the pixel point of the edge area of the blood vessel intima in the blood vessel ultrasonic image as far as possible.
10. The method for automatically identifying an intima of an intravascular ultrasound image according to claim 1, wherein in step 3, searching for a difference in pixel gradation in the scanning direction comprises:
sequentially calculating the difference of corresponding pixel gray values of the current point and the next point in the intravascular ultrasound image; when the difference exceeds a set threshold, the pixel coordinates of the current point are recorded, and the pixel coordinates of the next point are updated to the pixel coordinates of the current point, and the process is repeated until the next point reaches the image edge.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311654438.7A CN117351526B (en) | 2023-12-05 | 2023-12-05 | Intravascular ultrasound image automatic identification method for intima |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311654438.7A CN117351526B (en) | 2023-12-05 | 2023-12-05 | Intravascular ultrasound image automatic identification method for intima |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117351526A true CN117351526A (en) | 2024-01-05 |
| CN117351526B CN117351526B (en) | 2024-03-22 |
Family
ID=89367121
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311654438.7A Active CN117351526B (en) | 2023-12-05 | 2023-12-05 | Intravascular ultrasound image automatic identification method for intima |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117351526B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102982314A (en) * | 2012-11-05 | 2013-03-20 | 深圳市恩普电子技术有限公司 | Method of identifying, tracing and measuring external and internal membranes of vessel |
| CN103792567A (en) * | 2014-02-23 | 2014-05-14 | 北京航空航天大学 | Practical method for calibrating beam inclination angle of plate component chromatographic scanning device |
| US20160328998A1 (en) * | 2008-03-17 | 2016-11-10 | Worcester Polytechnic Institute | Virtual interactive system for ultrasound training |
| CN106204544A (en) * | 2016-06-29 | 2016-12-07 | 南京中观软件技术有限公司 | A kind of automatically extract index point position and the method and system of profile in image |
| CN106355586A (en) * | 2016-08-31 | 2017-01-25 | 黑龙江拓盟科技有限公司 | Automatic extraction method of human chest organ tissue |
| CN110327064A (en) * | 2019-06-19 | 2019-10-15 | 东软医疗系统股份有限公司 | Source image distance detecting method, device and equipment |
| CN111091599A (en) * | 2019-11-14 | 2020-05-01 | 东南大学 | Multi-camera-projector system calibration method based on sphere calibration object |
| CN112017200A (en) * | 2020-08-07 | 2020-12-01 | 湖北省农业科学院农产品加工与核农技术研究所 | Fish posture judgment method based on visual recognition |
| CN114370828A (en) * | 2021-12-28 | 2022-04-19 | 中国铁路设计集团有限公司 | Shield tunnel diameter convergence and radial slab staggering detection method based on laser scanning |
| WO2023179405A1 (en) * | 2022-03-21 | 2023-09-28 | 深圳市正浩创新科技股份有限公司 | Obstacle identification method, device and storage medium |
-
2023
- 2023-12-05 CN CN202311654438.7A patent/CN117351526B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160328998A1 (en) * | 2008-03-17 | 2016-11-10 | Worcester Polytechnic Institute | Virtual interactive system for ultrasound training |
| CN102982314A (en) * | 2012-11-05 | 2013-03-20 | 深圳市恩普电子技术有限公司 | Method of identifying, tracing and measuring external and internal membranes of vessel |
| CN103792567A (en) * | 2014-02-23 | 2014-05-14 | 北京航空航天大学 | Practical method for calibrating beam inclination angle of plate component chromatographic scanning device |
| CN106204544A (en) * | 2016-06-29 | 2016-12-07 | 南京中观软件技术有限公司 | A kind of automatically extract index point position and the method and system of profile in image |
| CN106355586A (en) * | 2016-08-31 | 2017-01-25 | 黑龙江拓盟科技有限公司 | Automatic extraction method of human chest organ tissue |
| CN110327064A (en) * | 2019-06-19 | 2019-10-15 | 东软医疗系统股份有限公司 | Source image distance detecting method, device and equipment |
| CN111091599A (en) * | 2019-11-14 | 2020-05-01 | 东南大学 | Multi-camera-projector system calibration method based on sphere calibration object |
| CN112017200A (en) * | 2020-08-07 | 2020-12-01 | 湖北省农业科学院农产品加工与核农技术研究所 | Fish posture judgment method based on visual recognition |
| CN114370828A (en) * | 2021-12-28 | 2022-04-19 | 中国铁路设计集团有限公司 | Shield tunnel diameter convergence and radial slab staggering detection method based on laser scanning |
| WO2023179405A1 (en) * | 2022-03-21 | 2023-09-28 | 深圳市正浩创新科技股份有限公司 | Obstacle identification method, device and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117351526B (en) | 2024-03-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lim et al. | Integrated optic disc and cup segmentation with deep learning | |
| CN106599804B (en) | Fovea centralis detection method based on multiple features model | |
| CN110390665B (en) | Knee joint disease ultrasonic diagnosis method based on deep learning multichannel and graph embedding method | |
| CN112257704A (en) | Cervical fluid-based cell digital image classification method based on deep learning detection model | |
| US20120201430A1 (en) | System and method for animal identification using iris images | |
| CN110263755B (en) | Eye ground image recognition model training method, eye ground image recognition method and eye ground image recognition device | |
| CN104637044B (en) | The ultrasonoscopy extraction system of calcified plaque and its sound shadow | |
| US11302094B2 (en) | System and method for segmenting normal organ and/or tumor structure based on artificial intelligence for radiation treatment planning | |
| CN107256410B (en) | Fundus image classification method and device | |
| Sun et al. | Optic disc segmentation from retinal fundus images via deep object detection networks | |
| CN102068281A (en) | Processing method for space-occupying lesion ultrasonic images | |
| CN111353978B (en) | Method and device for identifying heart anatomy structure | |
| CN115170571B (en) | Method for identifying pathological image of hydrothorax and ascites cells, image identification device and medium | |
| CN105530851A (en) | Image processing device, method, and program | |
| CN111932495B (en) | Medical image detection method, device and storage medium | |
| CN112819958B (en) | Engineering geological mapping method and system based on three-dimensional laser scanning | |
| CN111402243A (en) | Macular fovea identification method and terminal | |
| CN114092450A (en) | Real-time image segmentation method, system and device based on gastroscopy video | |
| CN110826565A (en) | Cross-connection-based convolutional neural network tooth mark tongue picture classification method and system | |
| CN117351526B (en) | Intravascular ultrasound image automatic identification method for intima | |
| CN112258536B (en) | Integrated positioning and segmentation method for calluses and cerebellum earthworm parts | |
| CN110276333B (en) | Eye ground identity recognition model training method, eye ground identity recognition method and equipment | |
| JP2004516585A (en) | Image difference | |
| Aich et al. | Optical disc segmentation from color fundus image using contrast limited adaptive histogram equalization and morphological operations | |
| JP6196760B2 (en) | Image processing device |
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 |