CN109124662B - Rib center line detection device and method - Google Patents
Rib center line detection device and method Download PDFInfo
- Publication number
- CN109124662B CN109124662B CN201810770823.0A CN201810770823A CN109124662B CN 109124662 B CN109124662 B CN 109124662B CN 201810770823 A CN201810770823 A CN 201810770823A CN 109124662 B CN109124662 B CN 109124662B
- Authority
- CN
- China
- Prior art keywords
- rib
- region
- center line
- point
- preset
- 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
- 238000000034 method Methods 0.000 title claims abstract description 59
- 238000001514 detection method Methods 0.000 title claims abstract description 39
- 230000005484 gravity Effects 0.000 claims abstract description 45
- 238000009499 grossing Methods 0.000 claims description 25
- 238000004891 communication Methods 0.000 claims description 14
- 230000009466 transformation Effects 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 11
- 238000012937 correction Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000010801 machine learning Methods 0.000 claims description 7
- 238000003062 neural network model Methods 0.000 claims description 6
- 238000003672 processing method Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 abstract description 13
- 230000002159 abnormal effect Effects 0.000 abstract description 6
- 238000012545 processing Methods 0.000 abstract description 5
- 230000000694 effects Effects 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000011426 transformation method Methods 0.000 description 2
- 206010017076 Fracture Diseases 0.000 description 1
- 206010027476 Metastases Diseases 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 208000027790 Rib fracture Diseases 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 210000000988 bone and bone Anatomy 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 210000000038 chest Anatomy 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 231100001063 rib abnormality Toxicity 0.000 description 1
- 210000000278 spinal cord Anatomy 0.000 description 1
- 238000013185 thoracic computed tomography Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
Landscapes
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Pathology (AREA)
- Heart & Thoracic Surgery (AREA)
- High Energy & Nuclear Physics (AREA)
- Physics & Mathematics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Optics & Photonics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Radiology & Medical Imaging (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the technical field of image processing, and particularly provides a rib center line detection device and method, aiming at solving the technical problems of poor robustness and low calculation speed in the prior art. To this end, the invention provides a rib center line detection method, which comprises the following steps: identifying the region of the ribs in the target image based on a preset rib identification model, and acquiring a rib region corresponding to each rib according to the region of the ribs; and acquiring a plurality of specific planes corresponding to each rib region in the three-dimensional space image corresponding to the target image, calculating the gravity centers of the corresponding rib regions according to the specific planes, and connecting all the gravity centers of the rib regions to obtain the center lines of the corresponding ribs. Based on the steps, the rib center line detection method provided by the invention can accurately and efficiently identify the region of the rib in the target image, can adapt to various abnormal conditions, and has the advantages of good robustness and high calculation speed. The device of the present invention also has the above-described advantageous effects.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a rib center line detection device and method.
Background
Physicians in the imaging department usually detect rib fractures and rib metastases of cancer by means of CT images. The number of CT images per patient may reach hundreds, and in order to check the focus and abnormality existing in the CT images, doctors often need to check the CT images layer by layer, and track the dynamic changes of each rib on different layers of slices so as to visually track the changes of the rib sectional area. However, manual examination of CT images is time consuming and is prone to errors and missing rib abnormalities.
Extracting the rib centerline can be used to enhance visualization of the unfolded rib, allowing the physician to more efficiently and effectively examine rib-related conditions, and can also be used as a reference for mapping between serial thoracic CT scans for organ localization, lesion registration, and guided interval change analysis.
The existing rib centerline detection method may include the following methods:
1. the ribs are modeled as elongated tubular structures and the spine voxel detection is performed using Hessian matrix (Hessian) or structure tensor eigensystem analysis, the tubular structures being detected by computing eigenvalues of the Hessian matrix at each point in 3D space. However, this method is computationally expensive and may not achieve consistent results for all patients.
2. Rib centerline matching by template: the section of the rib usually has a certain rule, the spinal cord at the center of the rib is dark, the bone at the edge of the rib is bright, the edge of the rib is in an ellipse shape, a template can be designed according to the rule to match the rib so as to find a central line, and then the central line is corrected according to the requirement. However, in many cases, the rib marrow may be darker than the rib boundaries, and thus the rib center point cannot be consistently detected as a dorsal voxel.
3. Identifying the centerline by a machine learning method: and extracting Haar characteristics of the CT image, training a prediction model of a central line through artificially labeled data, and predicting the probability of each pixel as the central line.
4. Tracking the center line by a curve tracking method: the seeds are manually set or automatically selected and the detected rib center point is tracked from one slice to the next using a tracking method such as kalman filtering. However, this method requires an artificial initial seed point, which is highly sensitive to local blurring or discontinuities caused by rib lesions such as fractures.
The prior art scheme has poor robustness, and medical images, particularly data containing focuses, are frequently varied, and abnormal conditions cannot be well handled by manually set rules; the calculation speed is slow, and the method in the prior art needs to set more rules and steps due to the need of processing more complex abnormal conditions, so that the calculation speed is slow.
Therefore, how to propose a technical solution to solve the above problems is a problem that needs to be solved by those skilled in the art at present.
Disclosure of Invention
In order to solve the above-mentioned problems of the prior art, i.e. to solve the problems of the prior art that the robustness is poor and the calculation speed is slow, a first aspect of the present invention provides a rib centerline detection apparatus, comprising:
the rib identification module is configured to identify the region where the ribs are located in the target image based on a preset rib identification model, and acquire a rib region corresponding to each rib according to the region where the ribs are located;
a rib center line acquisition module configured to acquire a plurality of specific planes corresponding to each rib region in a three-dimensional space image corresponding to the target image, calculate the center of gravity of the corresponding rib region according to the specific planes, and connect all the centers of gravity of each rib region to obtain the center line of the corresponding rib;
the rib recognition model is a neural network model which is constructed on the basis of a pre-calibrated rib image and by utilizing a machine learning algorithm; the specific plane is a plane with a plane angle being a preset polar angle and being parallel to a preset coordinate axis, and the angle value of the polar angle depends on a polar angle range of a two-dimensional projection point obtained by performing polar coordinate transformation on the two-dimensional projection point in the corresponding rib region according to the center of gravity of the region where the rib is located.
In a preferred embodiment of the above apparatus, the rib centerline acquisition module is further configured to perform the following operations:
projecting the gravity center of the region where the ribs are located and each rib region to a two-dimensional plane to respectively obtain a gravity center projection point and a rib projection region of each rib region;
performing polar coordinate transformation on the two-dimensional projection point in the projection area of each rib by taking the gravity center projection point as a center;
and acquiring a polar angle range of the two-dimensional projection point corresponding to the corresponding rib region according to the polar coordinate of each two-dimensional projection point in each rib projection region.
In a preferred embodiment of the above apparatus, the rib centerline acquisition module is further configured to perform the following operations:
and extracting a plurality of polar angles corresponding to each rib region according to the polar angle range of the two-dimensional projection point corresponding to each rib region and a preset angle step length.
In a preferred embodiment of the above device, the rib center line detection device further includes a center line correction module, and the center line correction module is configured to:
and smoothing the center line acquired by the rib center line acquisition module by using a preset curve smoothing method.
In a preferred embodiment of the above apparatus, the curve smoothing method is a curve processing method based on a preset markov random field model.
In a preferred embodiment of the foregoing apparatus, the centerline correction module is further configured to smooth the centerline according to a method shown in the following formula:
wherein E (X) represents an energy function of the Markov random field model, wherein X represents a dimension (X, y or z axis) in which all points on the centerline lie, and X representsiRepresenting the coordinates of the ith point on the central line in the X dimension, NiRepresents the number of points adjacent to the ith point on the central line, xjRepresents the sum of x on the center lineiCoordinates of the j-th point adjacent to the point, x'iRepresenting the initial coordinate of the ith point in the X dimension, and gamma represents X'iThe occupied weight, d, represents a preset distance.
In a preferred embodiment of the above apparatus, the rib identification module is further configured to perform the following operations:
and acquiring a plurality of communication areas contained in the area where the ribs are located in the three-dimensional space image corresponding to the target image, and acquiring the rib area corresponding to each rib according to each communication area.
In a preferred embodiment of the above apparatus, the rib identification module is further configured to perform the following operations:
acquiring all three-dimensional space points with Euclidean distances smaller than a preset distance in the region of the rib;
and acquiring a communication area of each rib according to the acquired three-dimensional space point.
The second aspect of the present invention provides a rib centerline detection method, including:
identifying the region of a rib in a target image based on a preset rib identification model, and acquiring a rib region corresponding to each rib according to the region of the rib;
acquiring a plurality of specific planes corresponding to each rib region in a three-dimensional space image corresponding to the target image, calculating the gravity center of the corresponding rib region according to the specific planes, and connecting all the gravity centers of each rib region to obtain the center line of the corresponding rib;
the rib recognition model is a neural network model which is constructed on the basis of a pre-calibrated rib image and by utilizing a machine learning algorithm; the specific plane is a plane with a plane angle being a preset polar angle and being parallel to a preset coordinate axis, and the angle value of the polar angle depends on a polar angle range of a two-dimensional projection point obtained by performing polar coordinate transformation on the two-dimensional projection point in the corresponding rib region according to the center of gravity of the region where the rib is located.
In a preferred embodiment of the above method, before the step of "acquiring a plurality of specific planes corresponding to each rib region in a three-dimensional space image corresponding to the target image, and calculating the center of gravity of the corresponding rib region according to the specific planes", the method further includes:
projecting the gravity center of the region where the ribs are located and each rib region to a two-dimensional plane to respectively obtain a gravity center projection point and a rib projection region of each rib region;
performing polar coordinate transformation on the two-dimensional projection point in the projection area of each rib by taking the gravity center projection point as a center;
and acquiring a polar angle range of the two-dimensional projection point corresponding to the corresponding rib region according to the polar coordinate of each two-dimensional projection point in each rib projection region.
In a preferred technical solution of the above method, after the step of "obtaining a polar angle range of the two-dimensional projection point corresponding to the corresponding rib region according to the polar coordinate of each two-dimensional projection point in each rib projection region", the method further includes:
and extracting a plurality of polar angles corresponding to each rib region according to the polar angle range of the two-dimensional projection point corresponding to each rib region and a preset angle step length.
In a preferred embodiment of the above method, after the step of "connecting all the centers of gravity of each rib region to obtain the center line of the corresponding rib", the method further includes:
and smoothing the center line acquired by the rib center line acquisition module by using a preset curve smoothing method.
In a preferred embodiment of the above method, the curve smoothing method is a curve processing method based on a preset markov random field model.
In a preferred technical solution of the above method, the step of "smoothing the center line acquired by the rib center line acquisition module by using a preset curve smoothing method" includes smoothing the center line by using a method shown in the following formula:
wherein E (X) represents an energy function of the Markov random field model, the X representing the centerA certain dimension (x, y or z axis) in which all points on the line lie, xiRepresenting the coordinates of the ith point on the central line in the X dimension, NiRepresents the number of points adjacent to the ith point on the central line, xjRepresents the sum of x on the center lineiCoordinates of the j-th point adjacent to the point, x'iRepresenting the initial coordinate of the ith point in the X dimension, and gamma represents X'iThe occupied weight, d, represents a preset distance.
In a preferred technical solution of the above method, the step of "obtaining a rib region corresponding to each rib according to a region where the rib is located" includes:
and acquiring a plurality of communication areas contained in the area where the ribs are located in the three-dimensional space image corresponding to the target image, and acquiring the rib area corresponding to each rib according to each communication area.
In a preferred technical solution of the above method, the step of "acquiring a plurality of connected regions included in a region where the rib is located in the three-dimensional space image corresponding to the target image" includes:
acquiring all three-dimensional space points with Euclidean distances smaller than a preset distance in the region of the rib;
and acquiring a communication area of each rib according to the acquired three-dimensional space point.
A third aspect of the present invention provides a storage device in which a plurality of programs are stored, the programs being adapted to be loaded by a processor to perform the rib centerline detection method according to any one of the above.
A fourth aspect of the present invention provides a control apparatus comprising a processor and a storage device adapted to store a plurality of programs, the programs being adapted to be loaded by the processor to perform the rib centerline detection method of any one of the above.
Compared with the closest prior art, the technical scheme at least has the following beneficial effects:
1. the rib recognition model can accurately and efficiently recognize the region where the ribs are located in the target image, can adapt to various abnormal conditions in the target image, and is more robust in the recognized region where the ribs are located;
2. after the accurate rib region in the target image is obtained based on the rib recognition model, the calculation problem of the central line of the rib region can be converted from the calculation problem of a three-dimensional plane to the calculation problem of a two-dimensional plane by a polar coordinate transformation method, so that the calculation processing process is simplified, and the calculation speed is increased.
Drawings
FIG. 1 is a schematic diagram illustrating the main steps of a rib centerline detection method according to an embodiment of the present invention;
FIG. 2 is a schematic view of a single rib cage access area according to one embodiment of the present invention;
FIG. 3 is a schematic view of a single rib centerline according to one embodiment of the present invention;
FIG. 4 is a schematic view of a modified centerline according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a rib centerline detection apparatus according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention
Preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only for explaining the technical principle of the present invention, and are not intended to limit the scope of the present invention.
Referring to fig. 1, fig. 1 illustrates the main steps of the rib centerline detection method in the present embodiment. As shown in fig. 1, the method for detecting the center line of the rib bone in the present embodiment includes the following steps:
step S101: and identifying the region of the rib in the target image based on a preset rib identification model.
In the embodiment of the invention, the rib recognition model can recognize the region where the rib is located from the target image. Specifically, the target image may be a CT image, and the rib recognition model may be a neural network model constructed based on a pre-calibrated rib image and using a machine learning algorithm, and the model may adopt a UNet-like network structure including 4 upsampling structures and 4 downsampling structures. Inputting the target image into a rib recognition model, wherein a convolution layer of the rib recognition model can adopt 3-by-3 convolution to check the input target image for convolution operation, and extracting the characteristics of the input convolution layer target image; the rib recognition model can use a ReLU activation function to remove redundancy in image data and retain the characteristics of the data to the maximum extent; the pooling layer of the rib recognition model can reduce the feature map of the input pooling layer by using Maxpooling with stride 2 and upsample with size 2, thereby reducing the feature vector output by the convolutional layer; and finally, performing convolution operation by using a convolution kernel of 1 × 1, connecting all the features through a full connection layer, inputting the connected features into a softmax classifier, and outputting the region where the ribs are located in the target image.
In order that the rib recognition model can better recognize the region where the ribs are located in the target image, the rib recognition model can be trained through the pre-calibrated rib image. Specifically, the original target image may be subjected to operations such as image translation, image rotation, image scaling, and image flipping, and pixels of the target image may be geometrically transformed without changing the content of the target image. After the pixels of the image are subjected to geometric transformation, the image subjected to the geometric transformation is subjected to operations such as gamma transformation, random clipping, Gaussian blur, normalization and the like, so that the rib recognition model to be trained can better recognize the region where the rib is located. In the training process, a dice loss function can be adopted to judge the proximity degree between the region where the ribs output by the rib recognition model to be trained and the region where the real ribs are located. The trained rib recognition model can accurately recognize the region where the rib is located from the target image, and subsequently, the center line of the rib can be quickly acquired based on the accurate region where the rib is located, so that the operation speed is improved. In addition, the rib identification model can adapt to various abnormal conditions, and the identified region of the rib is more robust.
The rib recognition model in the embodiment of the present invention may adopt a UNet-like network structure, a VGG network structure, a ResNet network structure, or the like, but the present invention is not limited herein.
Step S102: the method comprises the steps of obtaining a plurality of connected regions contained in a region where a rib is located in a three-dimensional space image corresponding to a target image, and obtaining a rib region corresponding to each rib according to each connected region.
With continued reference to FIG. 2, FIG. 2 illustratively depicts a schematic view of the communication zone of a single rib. And identifying a plurality of connected regions contained in the region where the ribs are located in the target image through a rib identification model, wherein each connected region corresponds to one rib. In order to obtain the center line of each rib, each rib in the region of the rib may be separated and analyzed individually. Specifically, three-dimensional space points in the region of the rib can be obtained, and whether the euclidean distance between any two points in the three-dimensional space points is smaller than or equal to a preset distance (for example, the euclidean distance between the two points is 1) is judged, if yes, the two points can be judged to be connected, and a plurality of connected points can form a connected region of a single rib.
Step S103: and acquiring a plurality of specific planes corresponding to each rib region, calculating the gravity center of the corresponding rib region according to the specific planes, and connecting all the gravity centers of each rib region to obtain the center line of the corresponding rib.
In the embodiment of the invention, after the region where the ribs are located is obtained, the first gravity center of the region where the ribs are located and each rib region are projected to the two-dimensional plane, the first gravity center projection point and the rib projection region of each rib region are respectively obtained, the first gravity center projection point is recorded as C, and the two-dimensional plane can be an xy plane. After each rib region is projected to a two-dimensional plane, angles of a starting point and an end point of the rib projection region are obtained and can be respectively recorded as alpha and beta, polar coordinate transformation is carried out on polar angle ranges (alpha, beta) of the rib projection region according to a preset angle step length by taking a first gravity center projection point C as a center, and a plurality of polar angles corresponding to each rib region are obtained. Assuming that the current angle is theta, the preset angle step is omega, the preset angle step has 10 corresponding values in the current angle theta, solving the corresponding second barycenter of the 10 values, and so on, and obtaining a plurality of second barycenters in the polar angle range (alpha, beta).
And acquiring a plurality of specific planes corresponding to each rib region in the three-dimensional space image corresponding to the target image, taking the current angle as theta as an example, constructing a plane parallel to the z axis based on a second gravity center corresponding to the angle, wherein the plane is the specific plane, the intersection point of the plane and the rib region is the z-axis coordinate of the second gravity center, and connecting the second gravity centers of each rib region to obtain the center line of the rib. With continued reference to FIG. 3, FIG. 3 illustrates a schematic view of a centerline of a single rib. The method can convert the central line of the rib region from the calculation problem of a three-dimensional plane to the calculation problem of a two-dimensional plane by a polar coordinate transformation method, thereby simplifying the algorithm processing process and improving the calculation speed.
Step S104: and smoothing the central line by using a preset curve smoothing method.
The calculated center line is calculated separately and independently for each rib, correlation between different ribs is not considered, and some points on the finally calculated center line may not be on the center line, so the center line is not smooth. In addition, the rib region identified by the rib identification model may be deviated, and abnormal points may be present, which may be far from the normal position. In order to correct the center line, the center line may be smoothed by a predetermined curve smoothing method. Specifically, the curve smoothing method may be to modify the centerline based on a preset markov random field model, and the centerline may be modified according to the method shown in formula (1):
wherein E (X) represents an energy function of the Markov random field model, X represents a certain dimension (X, y or z axis) of all points on the centerline, and X representsiRepresenting the coordinates of the ith point on the centerline in the X dimension, NiIndicates the number of points adjacent to the ith point on the center line, xjRepresents the sum of x on the center lineiCoordinates of the j-th point adjacent to the point, x'iDenotes the initial coordinate of the ith point in the X dimension, and gamma denotes X'iThe occupied weight, d, represents a preset distance.
The first part of the energy function (the part to the left of the plus sign) requires that the distance between adjacent center points is short, and the number of the adjacent center points is a factor for determining the smoothness degree of the curve; the second part of the energy function (the part to the right of the plus sign) requires that the centre point is not too far from the initial point, d represents a predetermined distance, which avoids outliers that are too far from the normal centre line, and if the points are too far from the centre line, the distance from the centre line is also simply d. By constraining the energy function, the centerline can be made smoother and insensitive to outliers. With continued reference to FIG. 4, FIG. 4 illustratively provides a schematic representation of the corrected centerline.
Although the foregoing embodiments describe the steps in the above sequential order, those skilled in the art will understand that, in order to achieve the effect of the present embodiments, the steps may not be executed in such an order, and may be executed simultaneously (in parallel) or in an inverse order, and these simple variations are within the scope of the present invention.
Based on the rib center line detection method embodiment, the invention also provides a rib center line detection device. The rib center line detection device will be described below with reference to the drawings.
Referring to fig. 5, fig. 5 is a view schematically showing the main structure of the rib center line detecting apparatus according to the present embodiment. As shown in fig. 5, the rib center line detection apparatus of the present embodiment includes a rib identification module 1 and a rib center line acquisition module 2.
The rib recognition module 1 is configured to recognize a region where ribs are located in a target image based on a preset rib recognition model, and acquire a rib region corresponding to each rib according to the region where the ribs are located;
the rib center line acquisition module 2 is configured to acquire a plurality of specific planes corresponding to each rib region in a three-dimensional space image corresponding to a target image, calculate the center of gravity of the corresponding rib region according to the specific planes, and connect all the centers of gravity of each rib region to obtain the center line of the corresponding rib;
the rib recognition model is a neural network model constructed based on a pre-calibrated rib image and by utilizing a machine learning algorithm; the specific plane is a plane with a plane angle as a preset polar angle and parallel to a preset coordinate axis, and the angle value of the polar angle depends on the polar angle range of the two-dimensional projection point obtained by performing polar coordinate transformation on the two-dimensional projection point in the corresponding rib region according to the center of gravity of the rib region.
In a preferred implementation of the embodiment of the present invention, the rib center line obtaining module 2 is further configured to perform the following operations:
projecting the gravity center of the region where the ribs are located and each rib region to a two-dimensional plane to respectively obtain a gravity center projection point and a rib projection region of each rib region;
performing polar coordinate transformation on the two-dimensional projection point in the projection area of each rib by taking the gravity center projection point as a center;
and acquiring the polar angle range of the two-dimensional projection point corresponding to the corresponding rib region according to the polar coordinates of each two-dimensional projection point in each rib projection region.
In a preferred implementation of the embodiment of the present invention, the rib center line obtaining module 2 is further configured to perform the following operations:
and extracting a plurality of polar angles corresponding to each rib region according to the polar angle range of the two-dimensional projection point corresponding to each rib region and a preset angle step length.
In a preferred implementation of the embodiment of the present invention, the rib centerline detection apparatus further includes a centerline correction module, and the centerline correction module is configured to:
and smoothing the central line acquired by the rib central line acquisition module 2 by using a preset curve smoothing method.
In a preferred implementation of the embodiment of the invention, the curve smoothing method is a curve processing method based on a preset markov random field model.
In a preferred implementation of the embodiment of the present invention, the centerline correction module is further configured to smooth the centerline according to a method shown in the following formula:
wherein E (X) represents an energy function of the Markov random field model, X represents a certain dimension (X, y or z axis) of all points on the centerline, and X representsiRepresenting the coordinates of the ith point on the centerline in the X dimension, NiIndicates the number of points adjacent to the ith point on the center line, xjRepresents the sum of x on the center lineiCoordinates of the j-th point adjacent to the point, x'iDenotes the initial coordinate of the ith point in the X dimension, and gamma denotes X'iThe occupied weight, d, represents a preset distance.
In a preferred implementation of the embodiment of the invention, the rib identification module 1 is further configured to perform the following operations:
and acquiring a plurality of connected regions contained in the region where the rib is located in the three-dimensional space image corresponding to the target image, and acquiring the rib region corresponding to each rib according to each connected region.
In a preferred implementation of the embodiment of the invention, the rib identification module 1 is further configured to perform the following operations:
acquiring all three-dimensional space points with Euclidean distances smaller than a preset distance in an area where ribs are located;
and acquiring a communication area of each rib according to the acquired three-dimensional space point.
Further, based on the above method embodiments, the present invention also provides a storage device, wherein the storage device stores a plurality of programs and the programs can be adapted to be loaded by a processor to execute the rib centerline detection method of the above method embodiments.
Still further, based on the above method embodiment, the present invention further provides a control apparatus, which includes a processor and a storage device; the storage device may be adapted to store a plurality of programs and these programs may be adapted to be loaded by the processor to perform the rib centerline detection method of the above-described method embodiments.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working process and the related descriptions of the embodiments of the apparatus according to the present invention may refer to the corresponding process in the foregoing method embodiments, and have the same beneficial effects as the foregoing method, and will not be described herein again.
Those of skill in the art will appreciate that the method steps of the examples described in connection with the embodiments disclosed herein may be embodied in electronic hardware, computer software, or combinations of both, and that the components and steps of the examples have been described above generally in terms of their functionality in order to clearly illustrate the interchangeability of electronic hardware and software. Whether such functionality is implemented as electronic hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing or implying any particular order or sequence. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein.
So far, the technical solutions of the present invention have been described in connection with the preferred embodiments shown in the drawings, but it is easily understood by those skilled in the art that the scope of the present invention is obviously not limited to these specific embodiments. Equivalent changes or substitutions of related technical features can be made by those skilled in the art without departing from the principle of the invention, and the technical scheme after the changes or substitutions can fall into the protection scope of the invention.
Claims (16)
1. A rib centerline detection device, comprising:
the rib identification module is configured to identify the region where the ribs are located in a target image based on a preset rib identification model, acquire a plurality of communication regions included in the region where the ribs are located in a three-dimensional space image corresponding to the target image, and acquire a rib region corresponding to each rib according to each communication region;
a rib center line acquisition module configured to acquire a plurality of specific planes corresponding to each rib region in a three-dimensional space image corresponding to the target image, calculate the center of gravity of the corresponding rib region according to the specific planes, and connect all the centers of gravity of each rib region to obtain the center line of the corresponding rib;
the rib recognition model is a neural network model which is constructed on the basis of a pre-calibrated rib image and by utilizing a machine learning algorithm; the specific plane is a plane with a plane angle being a preset polar angle and being parallel to a preset coordinate axis, and the angle value of the polar angle depends on a polar angle range of a two-dimensional projection point obtained by performing polar coordinate transformation on the two-dimensional projection point in the corresponding rib region according to the center of gravity of the region where the rib is located.
2. The rib centerline detection apparatus of claim 1, wherein the rib centerline acquisition module is further configured to:
projecting the gravity center of the region where the ribs are located and each rib region to a two-dimensional plane to respectively obtain a gravity center projection point and a rib projection region of each rib region;
performing polar coordinate transformation on the two-dimensional projection point in the projection area of each rib by taking the gravity center projection point as a center;
and acquiring a polar angle range of the two-dimensional projection point corresponding to the corresponding rib region according to the polar coordinate of each two-dimensional projection point in each rib projection region.
3. The rib centerline detection apparatus of claim 2, wherein the rib centerline acquisition module is further configured to:
and extracting a plurality of polar angles corresponding to each rib region according to the polar angle range of the two-dimensional projection point corresponding to each rib region and a preset angle step length.
4. The rib centerline detection device of claim 1, further comprising a centerline correction module configured to:
and smoothing the center line acquired by the rib center line acquisition module by using a preset curve smoothing method.
5. The rib centerline detection device of claim 4, wherein the curve smoothing method is a curve processing method based on a preset Markov random field model.
6. The rib centerline detection device of claim 5,
the center line correction module is further configured to smooth the center line according to a method shown in the following formula:
wherein E (X) represents an energy function of the Markov random field model, and X represents a dimension (X, y, or z axis) in which all points on the centerline lieDimension of) x, xiRepresenting the coordinates of the ith point on the central line in the X dimension, NiRepresents the number of points adjacent to the ith point on the central line, xjRepresents the sum of x on the center lineiCoordinates of the j-th point adjacent to the point, x'iRepresenting the initial coordinate of the ith point in the X dimension, and gamma represents X'iThe occupied weight, d, represents a preset distance.
7. The rib centerline detection apparatus of any of claims 1 to 5, wherein the rib identification module is further configured to:
acquiring all three-dimensional space points with Euclidean distances smaller than a preset distance in the region of the rib;
and acquiring a communication area of each rib according to the acquired three-dimensional space point.
8. A rib center line detection method is characterized by comprising the following steps:
identifying the region where the ribs are located in a target image based on a preset rib identification model, acquiring a plurality of communication regions contained in the region where the ribs are located in a three-dimensional space image corresponding to the target image, and acquiring the rib region corresponding to each rib according to each communication region;
acquiring a plurality of specific planes corresponding to each rib region in a three-dimensional space image corresponding to the target image, calculating the gravity center of the corresponding rib region according to the specific planes, and connecting all the gravity centers of each rib region to obtain the center line of the corresponding rib;
the rib recognition model is a neural network model which is constructed on the basis of a pre-calibrated rib image and by utilizing a machine learning algorithm; the specific plane is a plane with a plane angle being a preset polar angle and being parallel to a preset coordinate axis, and the angle value of the polar angle depends on a polar angle range of a two-dimensional projection point obtained by performing polar coordinate transformation on the two-dimensional projection point in the corresponding rib region according to the center of gravity of the region where the rib is located.
9. The rib center line detection method according to claim 8, wherein before the step of "acquiring a plurality of specific planes corresponding to each of the rib regions in a three-dimensional space image corresponding to the target image, and calculating the center of gravity of the corresponding rib region from the specific planes", the method further comprises:
projecting the gravity center of the region where the ribs are located and each rib region to a two-dimensional plane to respectively obtain a gravity center projection point and a rib projection region of each rib region;
performing polar coordinate transformation on the two-dimensional projection point in the projection area of each rib by taking the gravity center projection point as a center;
and acquiring a polar angle range of the two-dimensional projection point corresponding to the corresponding rib region according to the polar coordinate of each two-dimensional projection point in each rib projection region.
10. The rib center line detection method according to claim 9, wherein after the step of obtaining the polar angle range of the two-dimensional projection point corresponding to the corresponding rib region according to the polar coordinates of each two-dimensional projection point in each rib projection region, the method further comprises:
and extracting a plurality of polar angles corresponding to each rib region according to the polar angle range of the two-dimensional projection point corresponding to each rib region and a preset angle step length.
11. The rib center line detection method according to claim 8, wherein after the step of connecting all the centers of gravity of each rib region to obtain the center line of the corresponding rib, the method further comprises:
and smoothing the center line acquired by the rib center line acquisition module by using a preset curve smoothing method.
12. The rib centerline detection method of claim 11, wherein the curve smoothing method is a curve processing method based on a preset markov random field model.
13. The rib center line detection method according to claim 12, wherein the step of smoothing the center line acquired by the rib center line acquisition module by using a preset curve smoothing method comprises smoothing the center line according to a method shown in the following formula:
wherein E (X) represents an energy function of the Markov random field model, wherein X represents a dimension (X, y or z axis) in which all points on the centerline lie, and X representsiRepresenting the coordinates of the ith point on the central line in the X dimension, NiRepresents the number of points adjacent to the ith point on the central line, xjRepresents the sum of x on the center lineiCoordinates of the j-th point adjacent to the point, x'iRepresenting the initial coordinate of the ith point in the X dimension, and gamma represents X'iThe occupied weight, d, represents a preset distance.
14. The rib center line detection method according to any one of claims 8 to 12, wherein the step of obtaining a plurality of connected regions included in a region where the rib is located in the three-dimensional space image corresponding to the target image comprises:
acquiring all three-dimensional space points with Euclidean distances smaller than a preset distance in the region of the rib;
and acquiring a communication area of each rib according to the acquired three-dimensional space point.
15. A storage device having a plurality of programs stored therein, wherein the programs are adapted to be loaded by a processor to perform the rib centerline detection method according to any one of claims 8 to 14.
16. A control apparatus comprising a processor and a storage device adapted to store a plurality of programs, characterized in that the programs are adapted to be loaded by the processor to perform the rib centerline detection method according to any one of claims 8 to 14.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810770823.0A CN109124662B (en) | 2018-07-13 | 2018-07-13 | Rib center line detection device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810770823.0A CN109124662B (en) | 2018-07-13 | 2018-07-13 | Rib center line detection device and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109124662A CN109124662A (en) | 2019-01-04 |
| CN109124662B true CN109124662B (en) | 2021-12-03 |
Family
ID=64800670
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810770823.0A Active CN109124662B (en) | 2018-07-13 | 2018-07-13 | Rib center line detection device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109124662B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109830289B (en) * | 2019-01-18 | 2021-04-06 | 上海皓桦科技股份有限公司 | Rib image display device |
| CN111985512B (en) * | 2019-05-22 | 2024-02-02 | 杭州依图医疗技术有限公司 | Method and device for establishing human rib model |
| CN111986138B (en) * | 2019-05-22 | 2023-08-25 | 杭州依图医疗技术有限公司 | Method and device for acquiring rib positioning |
| CN110807770A (en) * | 2019-10-30 | 2020-02-18 | 杭州依图医疗技术有限公司 | Medical image processing, recognizing and displaying method and storage medium |
| CN111080573B (en) * | 2019-11-19 | 2024-02-27 | 上海联影智能医疗科技有限公司 | Rib image detection method, computer device and storage medium |
| CN112862793B (en) * | 2021-02-20 | 2024-08-16 | 浙江医准智能科技有限公司 | Rib centerline extraction method and device based on rib three-dimensional morphology and distribution |
| CN113643176B (en) * | 2021-07-28 | 2024-05-28 | 东软医疗系统股份有限公司 | Rib display method and device |
| CN114638828B (en) * | 2022-05-18 | 2022-09-23 | 数聚(山东)医疗科技有限公司 | Radiological image intelligent segmentation method based on computer vision |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004167109A (en) * | 2002-11-21 | 2004-06-17 | Osaka Industrial Promotion Organization | Three-dimensional measuring method, three-dimensional measuring system, image processor, and computer program |
| CN101452577A (en) * | 2008-11-26 | 2009-06-10 | 沈阳东软医疗系统有限公司 | Rib auto-demarcating method and device |
| CN102567625A (en) * | 2010-10-29 | 2012-07-11 | 西门子公司 | Automated rib ordering and pairing |
| CN103054602A (en) * | 2011-09-19 | 2013-04-24 | 西门子公司 | Method and system for up-vector detection for ribs in computed tomography volume |
| CN103218800A (en) * | 2011-09-27 | 2013-07-24 | 西门子公司 | Method and system for automatic rib centerline extraction using learning base deformable template matching |
| CN106611411A (en) * | 2015-10-19 | 2017-05-03 | 上海联影医疗科技有限公司 | Rib segmentation method in medical image and medical image processing device |
| CN106683090A (en) * | 2016-12-31 | 2017-05-17 | 上海联影医疗科技有限公司 | Rib positioning method in medical image and system thereof |
| CN107798682A (en) * | 2017-08-31 | 2018-03-13 | 深圳联影医疗科技有限公司 | Image segmentation system, method, apparatus and computer-readable recording medium |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7369693B2 (en) * | 2003-06-23 | 2008-05-06 | Siemens Medical Solutions Usa, Inc. | Thoracic cage coordinate system for recording pathologies in lung CT volume data |
| US7627159B2 (en) * | 2004-10-28 | 2009-12-01 | Siemens Medical Solutions Usa, Inc. | 2D visualization for rib analysis |
| US9373181B2 (en) * | 2005-10-17 | 2016-06-21 | Siemens Medical Soultions Usa, Inc. | System and method for enhanced viewing of rib metastasis |
| US8165376B2 (en) * | 2006-12-11 | 2012-04-24 | Siemens Corporation | System and method for automatic detection of rib metastasis in computed tomography volume |
| JP2010158452A (en) * | 2009-01-09 | 2010-07-22 | Fujifilm Corp | Image processing device and method, and program |
| US20130150704A1 (en) * | 2011-12-12 | 2013-06-13 | Shuki Vitek | Magnetic resonance imaging methods for rib identification |
| CN104252708B (en) * | 2013-06-28 | 2017-08-25 | 深圳先进技术研究院 | A kind of x-ray chest radiograph image processing method and system |
| CN107154038B (en) * | 2016-04-22 | 2021-05-14 | 孔德兴 | Rib fracture auxiliary diagnosis method based on rib visualization |
-
2018
- 2018-07-13 CN CN201810770823.0A patent/CN109124662B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004167109A (en) * | 2002-11-21 | 2004-06-17 | Osaka Industrial Promotion Organization | Three-dimensional measuring method, three-dimensional measuring system, image processor, and computer program |
| CN101452577A (en) * | 2008-11-26 | 2009-06-10 | 沈阳东软医疗系统有限公司 | Rib auto-demarcating method and device |
| CN102567625A (en) * | 2010-10-29 | 2012-07-11 | 西门子公司 | Automated rib ordering and pairing |
| CN103054602A (en) * | 2011-09-19 | 2013-04-24 | 西门子公司 | Method and system for up-vector detection for ribs in computed tomography volume |
| CN103218800A (en) * | 2011-09-27 | 2013-07-24 | 西门子公司 | Method and system for automatic rib centerline extraction using learning base deformable template matching |
| CN106611411A (en) * | 2015-10-19 | 2017-05-03 | 上海联影医疗科技有限公司 | Rib segmentation method in medical image and medical image processing device |
| CN106683090A (en) * | 2016-12-31 | 2017-05-17 | 上海联影医疗科技有限公司 | Rib positioning method in medical image and system thereof |
| CN107798682A (en) * | 2017-08-31 | 2018-03-13 | 深圳联影医疗科技有限公司 | Image segmentation system, method, apparatus and computer-readable recording medium |
Non-Patent Citations (3)
| Title |
|---|
| Pulmonary nodule registration in serial CT scans based on rib anatomy and nodule template matching;Jiazheng Shi等;《Medical physics》;20071231;第34卷(第4期);第1336-1347页 * |
| Segmentation of the Posterior Ribs in Chest Radiographs Using Iterated Contextual Pixel Classification;Marco Loog et al;《IEEE TRANSACTIONS ON MEDICAL IMAGING》;20060531;第25卷(第5期);第602-611页 * |
| 一种新的肝脏CTA图像肋骨和脊椎骨分割方法;刘驰等;《电脑知识与技术》;20100228;第6卷(第5期);第1166-1173页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109124662A (en) | 2019-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109124662B (en) | Rib center line detection device and method | |
| US20240119719A1 (en) | Identification and tracking of a predefined object in a set of images from a medical image scanner during a surgical procedure | |
| JP7558243B2 (en) | Feature Point Detection | |
| AU2017292642B2 (en) | System and method for automatic detection, localization, and semantic segmentation of anatomical objects | |
| CN112885453A (en) | Method and system for identifying pathological changes in subsequent medical images | |
| US8285013B2 (en) | Method and apparatus for detecting abnormal patterns within diagnosis target image utilizing the past positions of abnormal patterns | |
| CN102947841A (en) | Spiculated malignant mass detection and classification in radiographic image | |
| US9092867B2 (en) | Methods for segmenting images and detecting specific structures | |
| JP2007307358A (en) | Method, apparatus and program for image treatment | |
| CN111488872B (en) | Image detection method, image detection device, computer equipment and storage medium | |
| US20190392552A1 (en) | Spine image registration method | |
| JP6458166B2 (en) | MEDICAL IMAGE PROCESSING METHOD, DEVICE, SYSTEM, AND PROGRAM | |
| CN114022554A (en) | Massage robot acupuncture point detection and positioning method based on YOLO | |
| JP4964191B2 (en) | Image processing apparatus and method, and program | |
| Ramakrishnan et al. | Automatic three-dimensional rib centerline extraction from CT scans for enhanced visualization and anatomical context | |
| JP5364009B2 (en) | Image generating apparatus, image generating method, and program thereof | |
| CN115861656A (en) | Method, apparatus and system for automatically processing medical images to output an alert | |
| US9286688B2 (en) | Automatic segmentation of articulated structures | |
| CN113962957A (en) | Medical image processing method, bone image processing method, device and equipment | |
| Hsu | A hybrid approach for brain image registration with local constraints | |
| Roozgard et al. | Medical image registration using sparse coding and belief propagation | |
| JP6827707B2 (en) | Information processing equipment and information processing system | |
| US8165375B2 (en) | Method and system for registering CT data sets | |
| Iakovidis et al. | Robust model-based detection of the lung field boundaries in portable chest radiographs supported by selective thresholding | |
| WO2014106747A1 (en) | Methods and apparatus for image processing |
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 |