CN107133959B - Rapid blood vessel boundary three-dimensional segmentation method and system - Google Patents
Rapid blood vessel boundary three-dimensional segmentation method and system Download PDFInfo
- Publication number
- CN107133959B CN107133959B CN201710438256.4A CN201710438256A CN107133959B CN 107133959 B CN107133959 B CN 107133959B CN 201710438256 A CN201710438256 A CN 201710438256A CN 107133959 B CN107133959 B CN 107133959B
- Authority
- CN
- China
- Prior art keywords
- image
- artifact
- blood vessel
- lumen
- boundary
- 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
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/003—Reconstruction from projections, e.g. tomography
- G06T11/008—Specific post-processing after tomographic reconstruction, e.g. voxelisation, metal artifact correction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10101—Optical tomography; Optical coherence tomography [OCT]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
- Endoscopes (AREA)
Abstract
The invention provides a rapid three-dimensional segmentation method and a rapid three-dimensional segmentation system for a blood vessel boundary, wherein the system comprises the following steps: a preprocessing module; the first coordinate transformation module is used for converting the image into an image under a polar coordinate system; an image artifact identification module: determining an artifact in the image, and generating an artifact mask image; lumen center positioning module: determining the center of the lumen of the blood vessel in the original image; a second coordinate transformation module: converting the image of the determined vascular lumen center into a new image under a polar coordinate system, and removing the artifact in the new image according to the artifact mask image; a lumen boundary extraction module: generating an energy consumption image according to the image without the artifact, and determining the blood vessel boundary of the image under a polar coordinate system; the coordinate inverse transformation module: and converting the blood vessel boundary of the image under the polar coordinate system into a blood vessel boundary curve under the original image coordinate system. The invention can well identify and eliminate guide wire artifacts and vessel bifurcations in the image, thereby realizing rapid and accurate vessel three-dimensional boundary segmentation.
Description
Technical Field
The invention relates to a blood vessel three-dimensional reconstruction method based on an optical coherence tomography image, and belongs to the technical field of medical image processing and medical detection.
Background
Optical Coherence Tomography (OCT) is an emerging high resolution Tomography technique. Similar to the principle of Intravascular Ultrasound (IVUS), the technology takes near infrared as a light source, utilizes the interference principle of optical waves to image, converts optical signals of light beams scanning tissues into electric signals, and displays the electric signals as two-dimensional and three-dimensional images of gray images or pseudo-color images after computer processing. Because the wavelength of light is shorter than that of sound waves, the optical coherence tomography has higher resolution and can reach 10-20 um.
Based on the blood vessel segmentation and reconstruction of the optical coherence tomography image, a high-precision and high-accuracy blood vessel three-dimensional model can be provided so as to improve the accuracy of the later-stage virtual fractional flow reserve calculation. At present, two difficulties exist in blood vessel segmentation based on optical coherence tomography images: how to achieve fast segmentation and reconstruction; how to eliminate the interference of factors such as guide wire artifacts, branches and poor image quality in the segmentation reconstruction process, and the like, thereby realizing full-automatic three-dimensional vessel segmentation.
Due to the difficulties, the image segmentation and reconstruction system of the current optical coherence tomography imaging equipment has a plurality of defects. The current typical blood vessel segmentation methods include the following types:
the method comprises the following steps: the optical coherence tomography equipment manufactured by St Jude company has a blood vessel segmentation and reconstruction function, and mainly comprises the steps of analyzing the gray level change of an OCT image A-Line, extracting special gray level points in the A-Line as lumen boundary points, interpolating a plurality of lumen boundary points to form a lumen boundary curve, obtaining a blood vessel segmentation result, and forming a blood vessel three-dimensional model. The above method has the following technical defects: failing to identify a vessel branch; under the condition of poor image quality, lumen boundary points cannot be identified without any recognition, and segmentation cannot be carried out, so that reconstruction errors of the three-dimensional model of the blood vessel are caused.
Patent document CN 105741251 a: a blood vessel segmentation method of a liver CTA sequence image is disclosed. Firstly, performing contrast enhancement and smooth noise pretreatment on an input three-dimensional liver sequence image, then, enhancing a liver blood vessel and a boundary thereof by adopting OOF and OFA algorithms, and refining the center of the blood vessel; automatically searching seed points of the center line of the blood vessel according to the geometric structure of the blood vessel, extracting the center line of the liver blood vessel and constructing a liver blood vessel tree; and finally, preliminarily segmenting the liver blood vessels by combining a fast marching method, calculating corresponding blood vessels and a background gray level histogram, and realizing accurate segmentation of the liver blood vessels by adopting a graph segmentation algorithm. However, the above invention is not applicable to OCT images, and the morphology and the gradation of blood vessels are not the same between OCT images and CTA images: blood vessels in the CTA image are in a high-gray tree structure depending on blood filled with a contrast agent in a lumen; the OCT image is a single blood vessel image, the blood vessel is tubular, the inner part of the lumen of the blood vessel is low-gray, and the wall of the blood vessel is high-gray.
Patent document CN101283929 a: a method for reconstructing a three-dimensional model of a blood vessel is disclosed. The invention focuses on three-dimensional reconstruction of blood vessels, and the technical scheme is as follows: it combines the vessel cross-section information obtained from the IVUS image sequence with the spatial geometry information of the ultrasound catheter obtained from the three-dimensional reconstruction based on the X-ray contrast image to accurately reconstruct the anatomical structure of the vessel. The adopted blood vessel segmentation method comprises the following steps: the segmentation mode combining the snake model and the dynamic planning has the disadvantages that the blood vessel contour points need to be manually selected in the first frame image, and the snake model is not suitable for OCT images with guide wire artifacts and blood vessel branches. Although the invention can more comprehensively and accurately reflect the true forms of the blood vessels and the plaques which may exist compared with the reconstruction results obtained by respectively and independently utilizing the two images, the accuracy of the three-dimensional reconstruction result of the blood vessels is inferior to that of the invention because the two image sources of the invention have no advantages in image resolution compared with the optical coherence tomography image. In addition, in the aspect of plaque true morphological reaction, the advantages of optical coherence tomography imaging are very obvious, and information such as fiber cap thickness, macrophage infiltration degree and plaque nature which cannot be reflected by IVUS and X-ray contrast images can be reflected.
Patent document CN 106570856 a: the computer processes the received ultrasonic image and is provided with an image cutting module, an ROI extracting module, an image dynamic stretching module, an image filtering module, an LII estimating module, an MAI estimating module, a dynamic planning adjustment estimation boundary module and a post-processing module. The invention applies the computer image processing technology to the measurement of the intima-media thickness in the carotid artery ultrasonic image, and can avoid the defects of low efficiency and instability existing in a manual measurement mode. The level set in this method is used to initialize the extraction of the ROI area, while the present invention is used to find the center of the vessel lumen, with substantial differences in the algorithm for finding the center specifically. The method is greatly different from the application field of the patent, focuses on measuring the thickness of a membrane in a carotid artery ultrasonic image, and is not applied to the field of blood vessel segmentation and reconstruction. Moreover, the system module of the invention and the system module of the invention have great difference, the algorithm implementation aspect is completely different from the data structure construction aspect, and the functions such as guide wire artifact removal, branch identification and the like provided by the invention are not provided by the invention.
Patent document CN 102243759 a: a three-dimensional pulmonary vessel image segmentation method based on a geometric deformation model is disclosed, and the method comprises the following steps: (1) determining a blood vessel segmentation calculation region according to the characteristics of the human physiological structure; (2) calculating the mean value of the blood vessel region, and positioning the internal and external homogeneous regions of the target; (3) calculating the energy of the blood vessel edge, and evolving the curved surface along a second derivative in the gradient direction of the image so as to accurately converge the curved surface to a target edge; (4) on the basis, a three-dimensional blood vessel segmentation curved surface evolution model is established, and the mean value and the edge energy of the internal and external regions of the pulmonary blood vessel are effectively combined; (5) and according to the constructed deformation model, solving by adopting optimized level set evolution, and implicitly solving the curved surface motion according to level set function curved surface evolution. The method modifies a snake model used in a patent document CN 101283929A, creates an image segmentation method based on a geometric deformation model, and solves the problem by using optimization level set evolution. However, the invention is not suitable for segmentation and vessel reconstruction based on optical coherence tomography images. Because the guide wire, the catheter artifact and the branch exist in the OCT image, the catheter cannot be removed through the geometric deformation model, the artifact (non-vascular lumen part) behind the guide wire can be segmented out as a lumen, and the branch cannot be removed.
Although the above device software and patent literature provide a segmentation method for medical images from different angles, different image sources and different calculation methods, one or more technical defects still exist, and are not suitable for the blood vessel boundary segmentation of OCT images. Therefore, it is desirable to provide a new method and system for three-dimensional segmentation of blood vessel boundaries, which can overcome the above-mentioned drawbacks and achieve rapid and accurate three-dimensional segmentation.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, provides a high-efficiency and accurate three-dimensional segmentation method and system for a blood vessel boundary, and avoids the defects of low efficiency and instability of manual segmentation and incapability of eliminating lumen branches in the conventional automatic segmentation method.
The invention provides the following technical scheme:
the invention provides a blood vessel boundary three-dimensional segmentation method, which is characterized by comprising the following steps:
step 1: acquiring an OCT image;
step 2: preprocessing the OCT image;
and step 3: performing polar coordinate conversion on the preprocessed OCT image;
and 4, step 4: removing artifacts in the OCT image converted in the step (3), wherein the artifacts comprise catheter artifacts and guide wire artifacts, and artifact information is saved by an artifact mask image;
and 5: determining the center of the lumen of the blood vessel in the OCT image;
step 6: taking the lumen center as a polar coordinate origin, converting the image into a polar coordinate image to obtain a new image, and removing the artifact in the new image by using the artifact mask image obtained in the step 4, wherein the artifact comprises a catheter artifact and a guide wire artifact;
and 7: generating an energy consumption image according to the result of the step 6, and extracting a curve with the lowest total energy consumption, wherein the curve corresponds to the blood vessel boundary of the OCT image;
and 8: and (4) converting the blood vessel boundary obtained in the step (7) back to a blood vessel boundary curve under a Cartesian coordinate system through coordinate inverse transformation.
Preferably, the acquisition of the OCT image may be read from a local disk or transmitted directly by the OCT apparatus.
Preferably, the step 2 includes:
201: by usingA fixed mask I0Removing image pickup information on the periphery of the image;
202: performing three-dimensional Gaussian filtering on the image; variance of Gaussian filtering is σi＝ai*Spacing[i]，i∈[1,3]Represents the dimension, aiIs a constant, Spacing [ i ]]The physical distance between pixels in dimension i.
Preferably, the step 4 further comprises: since the catheter is fixed in physical diameter and always located in the center of the image, the catheter artifact in the image can be removed by modifying the gray value within the fixed radius of the right center of the OCT image to be 0.
Further preferably, the catheter artifact information is saved by modifying the gray scale values within a fixed radius from the center in the artifact mask image.
Preferably, the step 4 further comprises: before removing the guide wire artifact, the converted OCT image is accumulated and subjected to dimension reduction along the direction of a polar coordinate axis. Further preferably, the dimension reduction by accumulation may be performed along the polar coordinate axis to obtain an accumulation result T (θ, z):
wherein, the Image1(θ, ρ, z) is an image in polar coordinates. Since the low-gray region is always behind the guide wire artifact, the guide wire appears as a dark band in T (θ, z).
Preferably, for a single guide wire, extracting the upper edge and the lower edge of a dark band in T (theta, z), wherein the curves of the upper edge and the lower edge correspond to the upper angle and the lower angle of a guide wire artifact in an image, and setting the pixel at the part of angle to be 0 so as to remove the guide wire artifact; extracting the upper edge and the lower edge, and completing the extraction through image convolution;
t (theta, z) is the polar coordinate image after accumulated dimension reduction.
Further preferably, in the case of a single guide wire, the guide wire artifact in the original image can be eliminated by extracting curves of upper and lower edges of a dark band in T (θ, z), and the extraction method is similar to the step 7 lumen boundary extraction.
Preferably, for a plurality of guide wires, firstly, the T (theta, z) image is processed by top-hat transformation, opening operation and closing operation, and the post-image is binarized;
secondly, removing a communication area with a smaller area from the image obtained after binarization, and reserving a dark band;
thirdly, extracting the upper edge and the lower edge of the dark band, wherein the curves of the upper edge and the lower edge correspond to the upper angle and the lower angle of the guide wire artifact in the image, modifying the pixel gray level at the part of the angle in the artifact mask image, and storing the guide wire artifact result;
t (theta, z) is the polar coordinate image after accumulated dimension reduction.
Further preferably, for the case where there are a plurality of guide wires, the dark band in T (θ, z) will be extracted by a series of morphological operations plus binarization and connected component filtering. The morphological operations comprise: top hat transformation, opening operation and closing operation, wherein Otsu binarization is adopted for binarization. The top hat transformation is used for removing the condition that the gray levels of different areas of the image are uneven, the opening operation and the closing operation mainly eliminate the influence of the bracket in the image, and after binaryzation, a dark zone formed by the guide wire and dark spots formed by some brackets become 0. And removing dark spots formed by the stent by judging the connected components, and reserving dark bands. After the result of the dark band is obtained, the method for storing the guide wire artifact in the original image in the artifact mask image is the same as that of a single guide wire.
Theoretically, the method for removing the multiple guide wires is also applicable to a single guide wire, but the method for removing the single guide wire considers the continuous characteristics of the upper layer and the lower layer of the guide wire, and the effect is better. And finally, storing the obtained artifact result through an artifact mask.
Preferably, the vessel lumen center in the OCT image is determined by a fast wavefront evolution method. The rapid wave front evolution method can adopt a basic algorithm formula without improvement and adjustment, and can also be applied to the technical scheme of the invention.
Preferably, in step 7, the obtaining of the blood vessel boundary includes:
step 701, enhancing a boundary curve in an image through edge extraction;
step 702, converting the image after edge extraction into an energy consumption image; the energy consumption image is obtained by:
wherein Temp (i, j, k) is a mask, and Image2 is an Image after edge extraction;
and step 703, searching a curve with the lowest total energy consumption in the CostImage (x, y, z) by a dynamic programming method.
More preferably, the lumen centering can also be performed in the following specific manner:
and performing primary lumen segmentation by adopting a rapid wavefront evolution method, and solving the center of a segmentation result to serve as the center of the lumen. The center of the segmentation result may be determined by finding the center of the smallest inscribed circle, for example, but may also be determined by other methods, such as circumscribed polygon, centroid finding method, and the like.
Finally, curve fitting is performed on the central curve in consideration of continuity of the lumen center, and for example, Bezier curve fitting may be used.
More preferably, the lumen boundary extraction may be performed in the following specific manner:
after the coordinate transformation is completed, a closed and circular lumen boundary in each layer of the original image becomes a curve traversing from the left side to the right side of the polar coordinate system image. Lumen boundary extraction also becomes extracting the left-hand to right-hand curve of the polar coordinate system image. The boundary curve calculation can be realized by firstly extracting a reinforced boundary curve through the image edge, converting the reinforced boundary curve into an energy consumption image and then calculating the boundary curve by searching the curve with the minimum total energy consumption. The method comprises the following specific steps:
more preferably, the energy consumption image may be obtained as follows:
from the lower end of the image up, when the lumen boundary is reached, the gray level suddenly rises. Therefore, edge enhancement can be performed by mask convolution to obtain an energy consumption image CostImage (x, y, z):
wherein Temp (i, j, k) is a mask with a size of l m n. CostImage (x, y, z) considers the three-dimensional space continuity of the vessel lumen and can restrict the generation of branches.
More preferably, the lumen boundary curve may be obtained as follows:
the lumen boundary finding is completed by finding a curve with the lowest total energy consumption in the CostImage (x, y, z), and can be quickly completed by dynamic planning.
The dynamically planned state transition equation is:
where the curve extends from START (set of starting points), f (i, j, k) denotes: of all the curves from the start point set to the coordinates (i, j, k), the curve with the lowest total energy consumption has an energy consumption f (i, j, k). N (i, j, k) represents a set of points of the curve from the start point set to the coordinates (i, j, k) that are located at a point before the coordinates (i, j, k), i.e., a transition rule. The transfer rule meets the following requirements:
the directional state can be achieved, and the state can be shifted only clockwise or anticlockwise, namely, the state can be shifted only from left to right or from right to left all the time.
The neighborhood transfer can only be transferred from the current position to the adjacent position, the size of the neighborhood is not limited, and the neighborhood can be adjusted according to the image.
cost (i, j, k) represents the energy consumption function at (i, j, k):
cost(i,j,k)＝CostImage(i,j,k)+m(i,j,k)
here, CostImage (i, j, k) is the energy consumption image obtained in step a.
m (i, j, k) represents the energy consumption artificially set at (i, j, k), and can be used as the embodiment of the energy consumption function when the user adjusts the boundary. extra (i, j, k, i ', j', k ') represents the transfer cost, representing the additional energy consumption transferred from coordinate (i', j ', k') to coordinate (i, j, k). This term constrains state transitions to some extent and can constrain branch generation.
When f (i, j, k) is updated, it is recorded from which coordinate (i ', j ', k ') the coordinate (i, j, k) is updated.
The final vessel boundary Curve with the lowest total energy consumption is generated as follows: and finding the coordinate (i, j, k) with the minimum f (i, j, k) in the transition destination set, and then sequentially pushing back the coordinate (i ', j ', k ') until returning to the starting point set.
In addition, the invention also provides a rapid three-dimensional segmentation system for the blood vessel boundary, which is characterized by comprising:
the preprocessing module is used for preprocessing the acquired image;
a coordinate transformation module 1 (i.e. a first coordinate transformation module) which takes the right center of the image as a polar coordinate origin and converts the image into an image under a polar coordinate system;
an image artifact removal module: determining an artifact in the image, and removing the artifact from the image which is determined to be the center of the lumen and converted into a polar coordinate system; the artifacts include catheter, guide wire artifacts;
lumen center positioning module: determining the center of the lumen of the blood vessel in the original image;
coordinate transformation module 2 (i.e., second coordinate transformation module): converting the image for determining the center of the lumen of the blood vessel into an image under a polar coordinate system by taking the center of the lumen as a polar coordinate origin, and removing the artifact by combining an artifact mask image;
a lumen boundary extraction module: generating an energy consumption image according to the image without the artifact, and extracting a curve with the lowest total energy consumption to determine the blood vessel boundary of the image under a polar coordinate system;
the coordinate inverse transformation module: and converting the blood vessel boundary of the image under the polar coordinate system into a blood vessel boundary curve under the original image coordinate system.
Preferably, the system further comprises an OCT image acquisition module for acquiring an OCT image.
Preferably, the image preprocessing module transmits the processed image to the coordinate transformation module 1 and the lumen center positioning module.
Preferably, the image artifact identification module identifies an artifact and generates an artifact mask image, which is transmitted to the lumen boundary extraction module.
Preferably, the coordinate transformation module 2 generates an OCT image with the lumen center as the origin of polar coordinates in a new polar coordinate system according to the artifact mask image, the lumen center curve, and the preprocessed image, and transmits the OCT image to the lumen boundary extraction module.
Preferably, the coordinate transformation module 1 (i.e., the first coordinate transformation module) and the coordinate transformation module 2 (i.e., the second coordinate transformation module) may be provided as the same coordinate transformation module.
After the OCT image acquisition module acquires the image, the image is transmitted to the preprocessing module, the preprocessing module preprocesses the image, and the generated new image is transmitted to the coordinate transformation module 1 and the lumen center positioning module. The coordinate transformation module 1 converts the original OCT image into an OCT image under a polar coordinate system and transmits the OCT image to the image artifact identification module. The image artifact identification module removes the guide wire artifact of the catheter, generates an artifact mask image and transmits the artifact mask image to the lumen boundary extraction module. The lumen center positioning module generates a lumen center after obtaining the image from the preprocessing module, and transmits a lumen center curve to the coordinate transformation module 2. The coordinate transformation module 2 generates an OCT image which takes the lumen center as the origin of polar coordinates under a new polar coordinate system according to the artifact mask image, the lumen center curve and the OCT image, and transmits the image to the lumen boundary extraction module. And after extracting the lumen boundary curve, the lumen boundary extraction module transmits the curve data to the coordinate inverse transformation module, and the coordinate inverse transformation module converts the curve in the polar coordinate system into the curve in the original coordinate system.
The invention has the characteristics and beneficial effects that:
the method combining the minimum energy consumption and the wave front evolution is utilized to realize the full-automatic, rapid and accurate three-dimensional segmentation of the blood vessel boundary. The method solves the problems of difficulty in guide wire artifact identification, difficulty in blood vessel branch identification, incapability of segmenting images with poor quality and the like in the prior segmentation technology, improves the segmentation speed and ensures the segmentation accuracy.
Description of the drawings:
FIG. 1 is an algorithmic flow diagram of an embodiment of the present invention;
FIG. 2 is a raw OCT image of an embodiment of the invention;
FIG. 3 is an exemplary illustration of a fixed mask according to an embodiment of the invention;
FIG. 4 is a coordinate transformed OCT image of an embodiment of the invention;
FIG. 5 is a diagram illustrating accumulated dimension reduction according to an embodiment of the present invention;
FIG. 6 is an exemplary diagram of upper and lower edge extraction results according to an embodiment of the present invention;
FIG. 7 is a schematic illustration of an artifact mask according to an embodiment of the present invention;
FIG. 8 is an OCT image after obtaining the blood vessel boundary according to the embodiment of the invention;
fig. 9 is a schematic structural diagram of a rapid three-dimensional vessel boundary segmentation system according to an embodiment of the present invention.
Detailed Description
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It will be appreciated by those of skill in the art that the following specific examples or embodiments are a series of presently preferred arrangements of the invention to further explain the principles of the invention, and that such arrangements may be used in conjunction or association with one another, unless it is expressly stated that some or all of the specific examples or embodiments are not in association or association with other examples or embodiments. Meanwhile, the following specific examples or embodiments are only provided as an optimized arrangement mode and are not to be understood as limiting the protection scope of the present invention.
Example 1:
fig. 1 is a flow chart of a fast three-dimensional segmentation method for blood vessel boundaries according to the present invention. The method comprises the steps of firstly acquiring an original OCT image and preprocessing the original OCT image, wherein preprocessing comprises filtering and noise reduction and removing useless information on the periphery of the image by using a fixed mask. Then, the images are subjected to polar coordinate transformation, and guide wire and catheter artifacts in the original OCT images are removed according to the obtained images. Next, the center of the lumen in the original OCT image is found using a fast wavefront diffusion algorithm. And then, polar coordinate transformation is carried out according to the obtained lumen center, and the lumen boundary is rapidly extracted by using a dynamic programming method. And finally, inversely transforming the extracted boundary curve under the polar coordinate system back to the lower boundary curve of the OCT original image coordinate system.
The following describes the fast three-dimensional segmentation method of the blood vessel boundary according to the present invention in an example with reference to fig. 1.
Assume that the original image size obtained is
1. Acquiring an OCT image:
an OCT original image is obtained from the OCT imaging device, the size of the image is 700 × 270, the physical distance between pixels is (0.014mm,0.014mm,0.2mm), and the OCT image and parts in the image are shown in fig. 2.
2. Pretreatment:
first, a fixed mask I is used0Image pickup information on the periphery of the image is removed. Mask I0Is a circle with the diameter of 700, the center of the circle is positioned at the center of the original image, the inner gray scale of the circle is 255, and the outer gray scale of the circle is 0.
Gaussian filtering is carried out on the image, and the variance of the Gaussian filtering is sigmai＝ai*Spacing[i]，i∈[1,3]Represents the dimension, aiIs a constant, Spacing [ i ]]The physical distance between pixels in dimension i. Wherein a is1＝2，a2＝2，a3＝0。
3. And (3) carrying out coordinate transformation on the OCT original image:
converting the original image into a polar coordinate system image with the positive center of the image as a polar coordinate origin:
Image(x,y,z)→Image1(θ,ρ,z)
the polar coordinate system is down-sampled in the polar axis direction, and the size of the image obtained by final transformation is 180 × 100 × 270.
4. Removing catheter and guide wire artifacts in the image:
the catheter has a physical diameter of 1mm, so a circle with an image center diameter of 1 ÷ 0.014 ≈ 71 is used as the catheter location. The catheter artifact information is saved in an artifact mask map.
The dimension reduction and accumulation process in the guide wire artifact processing process is shown in fig. 5. Dimension reduction accumulation is carried out along the direction of a polar coordinate axis to obtain an accumulation result T (theta, z):
since the low-gray region is always behind the guide wire artifact, the guide wire appears as a dark band in T (θ, z). The size of the dimensionality reduction accumulation result T (θ, z) is 180 × 270.
For the case of a single guide wire, the guide wire artifact in the original image can be eliminated by extracting the upper and lower edges of the dark band in T (θ, z). The extraction of the upper and lower edges is done by image convolution:
wherein, TempUp (i, j) is an upper edge operator:
TempDown (i, j) is the lower edge operator:
the upper and lower edge extraction results are shown in fig. 6.
The extracted upper and lower edge curves correspond to the upper and lower angles of the guide wire artifact of the original image, and the guide wire artifact result is saved by modifying pixels in the artifact mask at the part of the angles.
For the case of a few images with multiple guide wires, dark bands in T (theta, z) will be extracted through a series of morphological operations plus binarization and connected component screening. The morphological operations comprise: top hat transformation, opening operation and closing operation, wherein Otsu binarization is adopted for binarization.
The top hat transformation is used for removing the condition of uneven gray level of different areas of the image, and the size of an operator is 15 x 15.
The open operation and the close operation mainly eliminate the influence of the stent in the image, and the size of the operator is 1 x 5.
After Otsu binarization of the results, dark bands formed by the guide wires and some dark spots formed by the scaffold will become 0. And removing the connected region with a smaller area through connected component detection, reserving a dark band, and taking the threshold value as 50. After the result of the dark band is obtained, the method for removing the guide wire artifact in the original image is the same as that of a single guide wire.
Theoretically, the method for removing the multiple guide wires is also applicable to a single guide wire, but the method for removing the single guide wire considers the continuous characteristics of the upper layer and the lower layer of the guide wire, and the effect is better.
5. Positioning the center of the tube cavity:
the lumen center positioning is mainly completed by a rapid wavefront evolution algorithm. The fast wavefront evolution algorithm is a solution method for solving the Eikonal equation. The Eikonal equation approximates the wave equation, which has the form:
the physical meaning of the Eikonal equation to solve for T (x) is that the wave is from the source point x0The minimum time required to reach point x within the computational domain Ω is diffused by the velocity f (x).
If the velocity F (x) satisfies constant positive or constant negative, the equation is converted into
The velocity function f (x) is a function related to the gray value, uniformity, and boundary characteristics of the image.
The fast wavefront evolution algorithm is calculated as follows:
a. initialization
1) Adding the catheter region into an Alive set, and setting the wave front arrival time T of the points in the Alive set to be 0, namely T (x, y, z) is 0;
2) and adding the neighborhood coordinates of the Alive interior points into the set Trial, and solving the arrival time T according to the equation (2).
3) The remaining points are added to the set Far and T is set to ∞.
b. Evolution of wavefront
1) And adding the coordinate (i, j, k) with the minimum T in the Trial set into the Alive set, and deleting the coordinate (i, j, k) from the Trail set.
2) The neighborhoods at (i, j, k) of the set Far and the set Trial are updated: if it is in the Far set, it is added to the Trial set and removed from the Far set.
3) The T of all neighborhoods of (i, j, k) are recalculated according to equation (2).
4) Return to step 1)
After the calculation of the fast wave front evolution algorithm is completed, T is taken as TsServes as a lumen diffusion boundary. The center of the inscribed circle is obtained and used as the center of the tube cavity. Wherein T issDenotes diffusion cutoff time, and takes Ts＝5.2。
And finally, Bezier curve fitting is carried out on the obtained lumen center curve.
6. And (3) coordinate transformation:
converting the original image into a polar coordinate system image with the lumen center as the origin of polar coordinates obtained in the previous step:
Image(x,y,z)→Image2(θ,ρ,z)
the polar coordinate system is down-sampled in the polar axis direction, and the size of the image obtained by final transformation is 180 × 100 × 270.
7. Extracting the lumen boundary:
after the coordinate conversion is completed in the previous step, a closed circular lumen boundary in each layer of the original image becomes a curve which traverses from the left side to the right side of the polar coordinate system image. Lumen boundary extraction also becomes extracting the left-hand to right-hand curve of the polar coordinate system image. The boundary curve calculation can be realized by firstly extracting a reinforced boundary curve through the image edge, converting the reinforced boundary curve into an energy consumption image and then calculating the boundary curve by searching the curve with the minimum total energy consumption. The method comprises the following specific steps:
a. energy consumption image acquisition
From the lower end of the image up, when the lumen boundary is reached, the gray level suddenly rises. Therefore, edge enhancement can be performed by mask convolution to obtain an energy consumption image CostImage (x, y, z):
wherein Temp (i, j, k) is a mask with a size of l m n.
The value is Temp (i, j, k) — (1, -2, -1; 0,0, 0; 1,2, 1;
-1,-2,-1；0,0,0；1,2,1；
-1,-2,-1；0,0,0；1,2,1；)
CostImage (x, y, z) considers the three-dimensional space continuity of the vessel lumen and can restrict the generation of branches.
b. Lumen boundary curve extraction
The lumen boundary finding is completed by finding a curve with the lowest total energy consumption in the CostImage (x, y, z), and can be quickly completed by dynamic planning.
The dynamically planned state transition equation is:
where the curve extends from START (starting point set, taking all coordinates where i ═ 0), and f (i, j, k) denotes: in all curves from the set of starting points to the coordinates (i, j, k), the total energy consumption is the lowest f (i, j, k). N (i, j, k) represents a set of points located before the coordinates (i, j, k) from the start point set to the set of curves of the coordinates (i, j, k), i.e., a transition rule. The transfer rule meets the following requirements:
1. the directional state can be achieved, and the state can be shifted only clockwise or anticlockwise, namely, the state can be shifted only from left to right or from right to left all the time.
2. The neighborhood transfer can only be transferred from the current position to the adjacent position, the size of the neighborhood is not limited, and the neighborhood can be adjusted according to the image.
N(i,j,k)＝{(i-1,j-3,k),(i-1,j-2,k),(i-1,j-1,k),(i-1,j,k),(i-1,j+1,k),(i-1,j+2,k),(i-1,j+3,k)}
cost (i, j, k) represents the energy consumption function at (i, j, k):
cost(i,j,k)＝CostImage(i,j,k)+m(i,j,k)
here, CostImage (i, j, k) is the energy consumption image obtained in step a.
m (i, j, k) represents the energy consumption artificially set at (i, j, k), and can be used as the embodiment of the energy consumption function when the user adjusts the boundary.
extra (i, j, k, i ', j', k ') represents the transfer cost, representing the additional energy consumption transferred from coordinate (i', j ', k') to coordinate (i, j, k). This term constrains state transitions to some extent and can constrain branch generation.
d(i,j,k,i′,j′,k′)＝h(i-i′,j-j′,k-k′)
When f (i, j, k) is updated, it is recorded from which coordinate (i ', j ', k ') the coordinate (i, j, k) is updated.
The final vessel boundary Curve with the lowest total energy consumption is generated as follows: finding the coordinate (i) with the minimum f (i, j, k) in the transition destination setmin,jmin,kmin) And then push back its last coordinate (i ', j ', k ') in turn until the starting point set is returned.
The specific steps of the coordinate inverse transformation are as follows:
and sequentially converting each coordinate in the blood vessel boundary Curve Curve from the polar coordinate system back to the OCT original image coordinate system to generate a boundary Curve in the OCT original image coordinate system. The resulting boundary curve is shown in fig. 8.
Example 2:
next, a rapid three-dimensional blood vessel boundary segmentation system according to the present embodiment will be described with reference to fig. 9. The system comprises:
the preprocessing module is used for preprocessing the acquired image;
the coordinate transformation module 1 is used for transforming the image into an image under a polar coordinate system by taking the right center of the image as a polar coordinate origin;
an image artifact removal module: determining an artifact in the image, and removing the artifact from the image which is determined to be the center of the lumen and converted into a polar coordinate system; the artifacts include catheter, guide wire artifacts;
lumen center positioning module: determining the center of the lumen of the blood vessel in the original image;
the coordinate transformation module 2: converting the image for determining the center of the lumen of the blood vessel into an image under a polar coordinate system by taking the center of the lumen as an origin of polar coordinates;
a lumen boundary extraction module: generating an energy consumption image according to the image without the artifact, and extracting a curve with the lowest total energy consumption to determine the blood vessel boundary of the image under a polar coordinate system;
the coordinate inverse transformation module: and converting the blood vessel boundary of the image under the polar coordinate system into a blood vessel boundary curve under the original image coordinate system.
Preferably, the system further comprises an OCT image acquisition module for acquiring an OCT image.
Preferably, the image preprocessing module transmits the processed image to the coordinate transformation module 1 and the lumen center positioning module.
Preferably, the image artifact identification module identifies an artifact and generates an artifact mask image, which is transmitted to the lumen boundary extraction module.
Preferably, the coordinate transformation module 2 generates an OCT image with the lumen center as the origin of polar coordinates in a new polar coordinate system according to the artifact mask image, the lumen center curve, and the preprocessed image, and transmits the OCT image to the lumen boundary extraction module.
Preferably, the coordinate transformation module 1 and the coordinate transformation module 2 may be configured as the same coordinate transformation module.
When the system operates, the OCT image acquisition module acquires an image and transmits the image to the preprocessing module, the preprocessing module preprocesses the image and transmits a generated new image to the coordinate transformation module 1 and the lumen center positioning module. The coordinate transformation module 1 converts the original OCT image into an OCT image under a polar coordinate system and transmits the OCT image to the image artifact identification module. The image artifact identification module removes the guide wire artifact of the catheter, generates an artifact mask image and transmits the artifact mask image to the lumen boundary extraction module. The lumen center positioning module generates a lumen center after obtaining the image from the preprocessing module, and transmits a lumen center curve to the coordinate transformation module 2. The coordinate transformation module 2 generates an OCT image which takes the lumen center as the origin of polar coordinates under a new polar coordinate system according to the artifact mask image, the lumen center curve and the OCT image, and transmits the image to the lumen boundary extraction module. And after extracting the lumen boundary curve, the lumen boundary extraction module transmits the curve data to the coordinate inverse transformation module, and the coordinate inverse transformation module converts the curve in the polar coordinate system into the curve in the original coordinate system.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A method of three-dimensional segmentation of a vessel boundary, the method comprising:
step 1: acquiring an OCT image;
step 2: preprocessing the OCT image;
and step 3: performing polar coordinate conversion on the preprocessed OCT image;
and 4, step 4: identifying artifacts in the OCT image converted in the step 3, wherein the artifacts comprise catheter artifacts and guide wire artifacts;
and 5: determining the center of the lumen of the blood vessel in the OCT image;
step 6: taking the lumen center as a polar coordinate origin, converting the image into a polar coordinate image to obtain a new image, and removing the artifact in the new image by using the artifact obtained in the step 4, wherein the artifact comprises a catheter artifact and a guide wire artifact;
and 7: generating an energy consumption image according to the result of the step 6, and extracting a curve with the lowest total energy consumption, wherein the curve corresponds to the blood vessel boundary of the OCT image;
and 8: and (4) converting the blood vessel boundary obtained in the step (7) into a blood vessel boundary curve in a Cartesian coordinate system through coordinate inverse transformation.
2. The method of claim 1, wherein the step 2 further comprises:
step 201: using a fixed mask I0Removing image pickup information on the periphery of the image;
step 202: performing three-dimensional Gaussian filtering on the image; variance of Gaussian filtering is σi＝ai*Spacing[i]，i∈[1,3]Represents the dimension, aiIs a constant, Spacing [ i ]]The physical distance between pixels in dimension i.
3. The method of claim 1, wherein the step 4 further comprises:
and modifying the gray value of the corresponding position in the artifact mask image, and storing the guide wire artifact information.
4. The method of claim 1,2 or 3, wherein the step 4 further comprises: before removing the guide wire artifact, the converted OCT image is accumulated and subjected to dimension reduction along the direction of a polar coordinate axis.
5. The method according to claim 4, characterized in that for a single guide wire, extracting the upper and lower edges of the dark band in T (theta, z), wherein the curves of the upper and lower edges correspond to the upper and lower angles of the guide wire artifact in the image, and setting the pixel at the part of the angle to 0 to remove the guide wire artifact; extracting the upper edge and the lower edge, and completing the extraction through image convolution;
t (theta, z) is the polar coordinate image after accumulated dimension reduction.
6. The method according to claim 4, wherein for a plurality of guide wires, first, top-hat conversion, opening operation, and closing operation processing are performed on a T (θ, z) image, and the processed image is binarized;
secondly, removing a communication area with a smaller area from the image obtained after binarization, and reserving a dark band;
thirdly, extracting the upper edge and the lower edge of the dark band, wherein the curves of the upper edge and the lower edge correspond to the upper angle and the lower angle of the guide wire artifact in the image, modifying the pixel gray level at the part of the angle in the artifact mask image, and storing the guide wire artifact result;
and the T (theta, z) image is a polar coordinate image after accumulated dimension reduction.
7. The method of claim 1,2 or 3, wherein in the step 5, the lumen center of the blood vessel in the OCT image is determined by a fast wavefront evolution method.
8. The method according to claim 1,2 or 3, wherein in step 7, the obtaining of the vessel boundary comprises:
step 701, enhancing a boundary curve in an image through edge extraction;
step 702, converting the image after edge extraction into an energy consumption image; the energy consumption image is obtained by:
wherein Temp (i, j, k) is a mask, and Image2 is an Image after edge extraction;
and step 703, searching a curve with the lowest total energy consumption in the CostImage (x, y, z) by a dynamic programming method.
9. A rapid vessel boundary three-dimensional segmentation system, the system comprising:
the preprocessing module is used for preprocessing the acquired image;
the first coordinate transformation module is used for transforming the image into an image under a polar coordinate system by taking the right center of the image as a polar coordinate origin;
an image artifact identification module: determining an artifact in the image, and generating an artifact mask image; the artifacts include catheter, guide wire artifacts;
lumen center positioning module: determining the center of the lumen of the blood vessel in the original image;
the second coordinate transformation module 2: converting the image for determining the center of the lumen of the blood vessel into an image under a polar coordinate system by taking the center of the lumen as a polar coordinate origin, and removing the artifact by combining an artifact mask image;
a lumen boundary extraction module: generating an energy consumption image according to the image without the artifact, and extracting a curve with the lowest total energy consumption to determine the blood vessel boundary of the image under a polar coordinate system;
the coordinate inverse transformation module: and converting the blood vessel boundary of the image under the polar coordinate system into a blood vessel boundary curve under the original image coordinate system.
10. The system of claim 9, further comprising an OCT image acquisition module to acquire an OCT image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710438256.4A CN107133959B (en) | 2017-06-12 | 2017-06-12 | Rapid blood vessel boundary three-dimensional segmentation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710438256.4A CN107133959B (en) | 2017-06-12 | 2017-06-12 | Rapid blood vessel boundary three-dimensional segmentation method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107133959A CN107133959A (en) | 2017-09-05 |
| CN107133959B true CN107133959B (en) | 2020-04-28 |
Family
ID=59734620
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710438256.4A Active CN107133959B (en) | 2017-06-12 | 2017-06-12 | Rapid blood vessel boundary three-dimensional segmentation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107133959B (en) |
Families Citing this family (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107949311B (en) | 2015-04-16 | 2021-04-16 | Gentuity有限责任公司 | Low-light level probe for neurology |
| US10631718B2 (en) | 2015-08-31 | 2020-04-28 | Gentuity, Llc | Imaging system includes imaging probe and delivery devices |
| EP3700406A4 (en) | 2017-11-28 | 2021-12-29 | Gentuity LLC | Imaging system |
| CN108182680B (en) * | 2017-12-28 | 2021-12-28 | 中科微光医疗研究中心(西安)有限公司 | IVOCT image-based angle automatic identification method for bifurcated vessels |
| CN108230341B (en) * | 2018-03-07 | 2021-12-17 | 汕头大学 | Eyeground image blood vessel segmentation method based on layered matting algorithm |
| CN108765432B (en) * | 2018-05-07 | 2020-08-07 | 山东大学 | Automatic carotid intima-media boundary segmentation method and system |
| CN108735270A (en) * | 2018-05-25 | 2018-11-02 | 杭州脉流科技有限公司 | Blood flow reserve score acquisition methods, device, system and computer storage media based on dimensionality reduction model |
| CN111093525A (en) * | 2018-08-07 | 2020-05-01 | 温州医科大学 | Optical coherence tomography image processing method |
| CN109118508B (en) * | 2018-08-31 | 2022-03-25 | 成都美律科技有限公司 | IVOCT image blood vessel wall lumen contour extraction method |
| CN109389568B (en) * | 2018-10-25 | 2022-04-01 | 中国科学院上海光学精密机械研究所 | Method for automatically measuring skin thickness in skin optical coherence tomography image |
| CN110060261B (en) * | 2019-04-17 | 2021-04-02 | 北京理工大学 | Blood vessel segmentation method based on optical coherence tomography system |
| CN110717888B (en) * | 2019-09-06 | 2021-03-12 | 天津大学 | Automatic identification method for intravascular Optical Coherence Tomography (OCT) vessel wall inner contour |
| CN111161342B (en) * | 2019-12-09 | 2023-08-29 | 杭州脉流科技有限公司 | Method, apparatus, device, system and readable storage medium for obtaining fractional flow reserve based on coronary angiography image |
| CN111523538B (en) * | 2020-04-14 | 2021-09-17 | 博动医学影像科技（上海）有限公司 | Blood vessel image processing method and system, computing device and storage medium |
| CN112509020B (en) * | 2020-12-07 | 2021-09-28 | 昆山戎影医疗科技有限公司 | Registration method and device for blood vessel external image and intracavity image and computing equipment |
| CN112766046B (en) * | 2020-12-28 | 2024-05-10 | 深圳市捷顺科技实业股份有限公司 | Target detection method and related device |
| CN112635029B (en) * | 2020-12-31 | 2024-09-17 | 上海联影智能医疗科技有限公司 | Medical image processing method, device, terminal and storage medium |
| CN113222956B (en) * | 2021-05-25 | 2023-09-15 | 南京大学 | Method for identifying plaque in blood vessel based on ultrasonic image |
| CN113436099B (en) * | 2021-06-25 | 2022-03-22 | 天津大学 | Intravascular optical coherence tomography two-stage catheter artifact removal method |
| CN113902690A (en) * | 2021-09-24 | 2022-01-07 | 杭州脉流科技有限公司 | Method, device, computing equipment and storage medium for computing fractional flow reserve based on intravascular images |
| CN114170258A (en) * | 2021-12-10 | 2022-03-11 | 深圳市赛禾医疗技术有限公司 | Image segmentation method and device, electronic equipment and storage medium |
| CN114445515B (en) * | 2022-02-14 | 2024-10-18 | 深圳市赛禾医疗技术有限公司 | Image artifact removal method and device, electronic equipment and storage medium |
| CN114820600B (en) * | 2022-06-27 | 2022-09-27 | 天津恒宇医疗科技有限公司 | Coronary artery intravascular stent detection method and system based on OCT image |
| CN116563414B (en) * | 2023-07-11 | 2023-09-12 | 天津博霆光电技术有限公司 | OCT-based cardiovascular imaging fibrillation shadow eliminating method and equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8750615B2 (en) * | 2010-08-02 | 2014-06-10 | Case Western Reserve University | Segmentation and quantification for intravascular optical coherence tomography images |
| WO2014117108A1 (en) * | 2013-01-25 | 2014-07-31 | Duke University | Segmentation and identification of closed-contour features in images using graph theory and quasi-polar transform |
| CN104766304A (en) * | 2015-02-26 | 2015-07-08 | 浙江工业大学 | Blood vessel registering method based on multi-sequence medical images |
| CN104851103A (en) * | 2015-05-29 | 2015-08-19 | 西安交通大学医学院第一附属医院 | Choroid membrane blood vessel extraction method based on SD-OCT retina image |
| CN104966288A (en) * | 2015-06-11 | 2015-10-07 | 南京沃福曼医疗科技有限公司 | Method for automatic searching of inner wall of blood vessel based on image processing |
| CN105825488A (en) * | 2016-05-30 | 2016-08-03 | 天津大学 | Cardiovascular optical coherence tomography image enhancement method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5778579B2 (en) * | 2008-10-14 | 2015-09-16 | ライトラボ・イメージング・インコーポレーテッド | Stent strut detection using optical coherence tomography and related methods for measurement and display |
-
2017
- 2017-06-12 CN CN201710438256.4A patent/CN107133959B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8750615B2 (en) * | 2010-08-02 | 2014-06-10 | Case Western Reserve University | Segmentation and quantification for intravascular optical coherence tomography images |
| WO2014117108A1 (en) * | 2013-01-25 | 2014-07-31 | Duke University | Segmentation and identification of closed-contour features in images using graph theory and quasi-polar transform |
| CN104766304A (en) * | 2015-02-26 | 2015-07-08 | 浙江工业大学 | Blood vessel registering method based on multi-sequence medical images |
| CN104851103A (en) * | 2015-05-29 | 2015-08-19 | 西安交通大学医学院第一附属医院 | Choroid membrane blood vessel extraction method based on SD-OCT retina image |
| CN104966288A (en) * | 2015-06-11 | 2015-10-07 | 南京沃福曼医疗科技有限公司 | Method for automatic searching of inner wall of blood vessel based on image processing |
| CN105825488A (en) * | 2016-05-30 | 2016-08-03 | 天津大学 | Cardiovascular optical coherence tomography image enhancement method |
Non-Patent Citations (2)
| Title |
|---|
| The Effect of the Sensor Position of the Pressure Wire Distal to a Coronary;B Tar等;《Computing in Cardiology 2013》;20131231;第1099-1102页 * |
| 一种面向医学图像非刚性配准的多维特征度量方法;陆雪松等;《自动化学报》;20160930;第1413-1420页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107133959A (en) | 2017-09-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107133959B (en) | Rapid blood vessel boundary three-dimensional segmentation method and system | |
| CN110706246B (en) | Blood vessel image segmentation method and device, electronic equipment and storage medium | |
| US9119559B2 (en) | Method and system of generating a 3D visualization from 2D images | |
| JP6243763B2 (en) | Image processing apparatus, method of operating image processing apparatus, and program | |
| CN109478327B (en) | Method for automatic detection of systemic arteries in Computed Tomography Angiography (CTA) of arbitrary field of view | |
| CN101964118B (en) | Three-dimensional segmentation method for intravascular ultrasound image sequence | |
| CN110047086B (en) | Automatic carotid intimal thickness measuring method and system | |
| CN111784720B (en) | DSA and IVOCT blood vessel image fusion method | |
| WO2015136853A1 (en) | Image processing device, image processing method, and program | |
| CN116503607B (en) | CT image segmentation method and system based on deep learning | |
| CN109410181B (en) | Heart image segmentation method and device | |
| CN117078711A (en) | Medical image segmentation method, system, electronic device and storage medium | |
| CN117974652B (en) | Ultrasonic image auxiliary positioning method based on machine vision | |
| CN117115150A (en) | Method, computing device and medium for determining branch vessels | |
| CN114693710A (en) | Blood vessel lumen intimal contour extraction method and device, ultrasonic equipment and storage medium | |
| CN113902690A (en) | Method, device, computing equipment and storage medium for computing fractional flow reserve based on intravascular images | |
| CN109674493B (en) | Method, system and equipment for medical ultrasonic automatic tracking of carotid artery blood vessel | |
| CN112651984B (en) | Vascular lumen intima contour extraction method and device, ultrasonic equipment and storage medium | |
| CN111627023B (en) | Method and device for generating coronary artery projection image and computer readable medium | |
| CN115564817B (en) | Three-dimensional blood vessel image determination method, device, equipment and storage medium | |
| CN114820678B (en) | Automatic extraction method and system for inner contour of blood vessel wall based on OCT image | |
| CN109345498B (en) | Coronary artery segmentation method fusing dual-source CT data | |
| Hammouche et al. | Space curve approach for IVUS image segmentation | |
| Moraes et al. | Automatic IOCT lumen segmentation using wavelet and Mathematical Morphology | |
| JP2022189770A (en) | Systems and methods for endovascular device detection and crimping measurement |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230626 Address after: 200235 A6, Building 1, No. 900 Yishan Road, Xuhui District, Shanghai Patentee after: Shanghai Bodong Medical Technology Co.,Ltd. Address before: 200240 No. 800, Dongchuan Road, Shanghai, Minhang District Patentee before: SHANGHAI JIAO TONG University |
|
| TR01 | Transfer of patent right |