CN112184720A - Method and system for segmenting rectus muscle and optic nerve of CT image - Google Patents
Method and system for segmenting rectus muscle and optic nerve of CT image Download PDFInfo
- Publication number
- CN112184720A CN112184720A CN202010891689.7A CN202010891689A CN112184720A CN 112184720 A CN112184720 A CN 112184720A CN 202010891689 A CN202010891689 A CN 202010891689A CN 112184720 A CN112184720 A CN 112184720A
- Authority
- CN
- China
- Prior art keywords
- image
- segmentation
- optic
- shape
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 210000003205 muscle Anatomy 0.000 title claims abstract description 58
- 210000001328 optic nerve Anatomy 0.000 title claims abstract description 44
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000011218 segmentation Effects 0.000 claims abstract description 91
- 230000004400 visual pathway Effects 0.000 claims abstract description 36
- 210000000239 visual pathway Anatomy 0.000 claims abstract description 36
- 238000012549 training Methods 0.000 claims abstract description 25
- 230000004927 fusion Effects 0.000 claims abstract description 20
- 210000003977 optic chiasm Anatomy 0.000 claims abstract description 10
- 239000011159 matrix material Substances 0.000 claims description 12
- 210000004872 soft tissue Anatomy 0.000 claims description 10
- 238000013507 mapping Methods 0.000 claims description 8
- 230000009466 transformation Effects 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 6
- 238000000513 principal component analysis Methods 0.000 claims description 6
- 230000002146 bilateral effect Effects 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 3
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 238000002598 diffusion tensor imaging Methods 0.000 claims description 3
- 238000011524 similarity measure Methods 0.000 claims description 3
- 210000005036 nerve Anatomy 0.000 claims description 2
- 210000005112 optic tract Anatomy 0.000 claims description 2
- 230000004421 optic tracts Effects 0.000 claims description 2
- 230000004413 optic chiasma Effects 0.000 abstract description 4
- 210000001154 skull base Anatomy 0.000 description 4
- 238000013170 computed tomography imaging Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 2
- 210000005252 bulbus oculi Anatomy 0.000 description 2
- 230000006378 damage Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000002675 image-guided surgery Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 210000000056 organ Anatomy 0.000 description 2
- 238000002719 stereotactic radiosurgery Methods 0.000 description 2
- 238000001356 surgical procedure Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 210000003484 anatomy Anatomy 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 210000000988 bone and bone Anatomy 0.000 description 1
- 230000037182 bone density Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 210000000695 crystalline len Anatomy 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 210000003128 head Anatomy 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000004279 orbit Anatomy 0.000 description 1
- 230000001817 pituitary effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/251—Fusion techniques of input or preprocessed data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30041—Eye; Retina; Ophthalmic
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
A segmentation method and a segmentation system for internal rectus muscle and optic nerve of a CT image can effectively position the optic chiasma and optic nerve bundle which are not clearly imaged in the CT, effectively make up the weakness of lack of local information in multi-mode fusion and remarkably improve the segmentation precision. The method comprises the following steps: (1) constructing a statistical shape model: the statistical shape model is composed of a training data set, wherein the shape of the anterior visual pathway and the internal rectus muscle are manually delineated; (2) segmentation based on MR/CT image fusion: obtaining a shape of a reference MR image by fitting a statistical shape model to a segmentation result of the MR image, and fusing a CT image with the MR image through elastic registration to obtain an initial segmentation result of a anterior visual pathway and an internal rectus muscle; (3) multi-feature constraint segmentation refinement: a multi-feature constraining surface is obtained from the target CT image, and structures not visible in the CT image, including the optic beam and the optic cross, are segmented after fitting the initial segmentation results to the surface.
Description
Technical Field
The invention relates to the technical field of medical image processing, in particular to a method for segmenting an internal rectus muscle and an optic nerve of a CT image and a system for segmenting the internal rectus muscle and the optic nerve of the CT image.
Background
Stereotactic Radiosurgery (SRS) and Image-guided surgery (IGS) are two techniques commonly used in the treatment of cranial base tumors. Due to the high bone density, CT is the primary imaging modality in the planning phase and surgery of cranial base surgery. In clinic, a surgeon must rely on abundant clinical experience to accurately locate brain structures in a CT image and avoid the injury of surgical instruments to key structures of the skull base (nerves, eyeballs, muscles in the eye sockets, etc.). However, such procedures are very dangerous to the patient. Therefore, automatic segmentation of the anterior visual pathway (optic nerve, optic tract and optic chiasm) and the internal rectus muscle in CT images is critical to improve the accuracy of the procedure and reduce damage to other anatomical structures.
In recent years, segmentation methods have been widely developed and can be divided into atlas registration and statistical shape model base methods. Bekes et al propose a geometric model-based method to segment the eyeball, lens, optic nerve and optic chiasm in CT images. It requires an interactive selection of seed points to initialize the segmentation. Huo Y et al propose a multi-atlas registration segmentation process that includes two steps: (1) bone structure affine registration to crop a visual pathway region in a target and map set, (2) deformable registration of the cropped region. However, due to the low contrast of soft tissue in CT images, atlas registration based methods cannot accurately segment the visual pathway. Chen and Dawant use a method of multi-atlas registration to segment head and neck organs. The method allows the target volume to be initially aligned with the map-set and then local registration is achieved by defining bounding boxes for each structure. Aghdasi et al apply a predefined anatomical model to segment visual organs and some brain structures in MR images. In addition, some studies show that the segmentation accuracy of smaller structures such as optic nerves can be improved by a multi-atlas registration-based method. Over the past few decades, model-based segmentation methods have been widely developed for anterior visual pathway segmentation. Nobel et al combines deformable models and atlas registration with previous local intensities to segment the pre-visual pathway. The statistical shape model includes an active appearance model and an active shape model, and is effective for solving the problem of segmentation of a structure with poor CT image quality. In summary, SSM (statistical shape model) based methods are better suited for poor image quality than atlas registration based methods.
In some other studies, it is also common to use deep learning to segment the cranial base tissue. Jose Dolz et al extracted enhanced features in MR images and proposed a deep learning classification scheme for optic nerve, optic chiasm and pituitary segmentation. Ren et al propose a strategy of interleaving 3D-CNNs for segmentation of pre-vision paths in CT images. In the field of medical image segmentation, U-Net is also widely applied and provides accurate segmentation. However, in the absence of a significant amount of data, neural network-based methods cannot accurately segment the anterior visual pathway and the internal rectus muscle.
A priori knowledge plays an important role in the segmentation of CT images. For statistical shape models, a model constructed from training data may be considered a priori information. Segmentation based on atlas registration depends on the quality of the target image and the prior information. Although CT images of soft tissue (e.g., anterior visual pathway and internal rectus muscle) suffer from a number of deficiencies, such as low contrast, blurred edges and noise. In this case, segmentation can be obtained by fitting a statistical shape model even if the extracted object boundary is blurred and fragmented. Furthermore, unlike the learning-based approach, the statistical shape model-based approach performs well in segmentation when the size of the training set is small. Because the structures of the anterior visual pathway and the internal rectus muscle in the MR data are complete, a statistical shape model can be constructed as prior information to realize accurate segmentation of the CT image.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a CT image segmentation method for rectus muscle and optic nerve, which can effectively position the optic chiasma and optic nerve bundle which are not clearly imaged in CT, and can effectively make up the defect of lack of local information in multi-modal fusion so as to obviously improve the segmentation precision.
The technical scheme of the invention is as follows: the method for segmenting the internal rectus muscle and the optic nerve of the CT image comprises the following steps:
(1) constructing a statistical shape model: the statistical shape model is composed of a training data set, wherein the shape of the anterior visual pathway and the internal rectus muscle are manually delineated;
(2) segmentation based on MR/CT image fusion: obtaining a shape of a reference MR image by fitting a statistical shape model to a segmentation result of the MR image, and fusing a CT image with the MR image through elastic registration to obtain an initial segmentation result of a anterior visual pathway and an internal rectus muscle;
(3) multi-feature constraint segmentation refinement: a multi-feature constraining surface is obtained from the target CT image, and structures not visible in the CT image, including the optic beam and the optic cross, are segmented after fitting the initial segmentation results to the surface.
The MR data set is used for constructing a prior shape model to assist the segmentation of a structural CT image, and because of the weakness of soft tissue CT imaging, the invention can effectively position the optic chiasma and optic nerve bundle which are not clearly imaged in CT; the multi-feature constrained surface can effectively make up the weakness of lack of local information in multi-modal fusion, so that the segmentation precision is remarkably improved.
Also provided is a system for medial rectus muscle and optic nerve segmentation of CT images, comprising:
a statistical shape model construction module configured to train a shape correspondence of the data set, and construct a statistical shape model of the training shape using principal component analysis;
a MR/CT image fusion-based segmentation module configured to obtain a shape of a reference MR image by fitting a statistical shape model to a segmentation result of the MR image, the CT image being fused with the MR image by elastic registration to obtain an initial segmentation result of the anterior visual pathway and the medial rectus muscle;
a multi-feature constrained segmentation refinement module configured to obtain a multi-feature constrained surface from the target CT image, and, after fitting the initial segmentation results to the surface, segment structures not visible in the CT image, including the optic beam and the optic cross.
Drawings
Fig. 1 is a flowchart of a method of intra-rectus muscle and optic nerve segmentation of a CT image according to the present invention.
Detailed Description
As shown in fig. 1, the method for segmenting the rectus muscle and the optic nerve of the CT image comprises the following steps:
(1) constructing a statistical shape model: the statistical shape model is composed of a training data set, wherein the shape of the anterior visual pathway and the internal rectus muscle are manually delineated;
(2) segmentation based on MR/CT image fusion: obtaining a shape of a reference MR image by fitting a statistical shape model to a segmentation result of the MR image, and fusing a CT image with the MR image through elastic registration to obtain an initial segmentation result of a anterior visual pathway and an internal rectus muscle;
(3) multi-feature constraint segmentation refinement: a multi-feature constraining surface is obtained from the target CT image, and structures not visible in the CT image, including the optic beam and the optic cross, are segmented after fitting the initial segmentation results to the surface.
The MR data set is used for constructing a prior shape model to assist the segmentation of a structural CT image, and because of the weakness of soft tissue CT imaging, the invention can effectively position the optic chiasma and optic nerve bundle which are not clearly imaged in CT; the multi-feature constrained surface can effectively make up the weakness of lack of local information in multi-modal fusion, so that the segmentation precision is remarkably improved.
Preferably, in step (1), in order to construct a statistical shape model, the shape correspondence of the training data set is:
shape correspondence is expressed as a dense mapping between a set of shape points in the MR data set, the correspondence of the two shapes being obtained by pairwise non-rigid registration; for MR data sets
Obtaining unbiased point correspondences by group shape registration, anAnd the similarity measure shape registration obtained by means of grouping level is expressed as formula (1):
where N is the number of training data, d (.) is the Euclidean distance, gijIs the connection between the ith and jth shapes in the dataset;
the connection relation among all shapes is represented by a graphic model, and then the group level registration is realized through the guidance of the graphic model;
obtaining shape correspondence
On this basis, the alignment shape is analyzed using the generalized equation.
Preferably, the step (1) adopts principal component analysis to construct a statistical shape model of the training shape, and performs eigenvalue decomposition on the matrix, wherein
Are vectorized and then arranged together, and the eigenvectors are arranged according to a descending order of eigenvalues, the first few eigenvectors being used to model the shape data, the statistical shape model being formula (2):
where P represents a vectoring matrix vec (P), vectoring average shape
And the principal eigenmodes form a matrix Φ that is pre-computed from the training dataset, where b represents the parameters of the model.
Preferably, in the step (2), the reference MR image I is randomly selected from the training datarefAnd its corresponding segmentation image; by fitting a statistical shape model to the reference MR mapLike IrefIs divided intoTTo obtain the reference shape, this process is called surface fitting and is expressed by equation (3):
wherein DTIs ITThe distance of (a) is transformed,
coordinates of points on the statistical shape model are represented, and diag (λ) represents a diagonal matrix composed of eigenvalues λ; b is constrained in a hyper-rectangle defined by β and λ, where λiIs the i-th element of λ, biIs the ith parameter in b; the first term in equation (3) is from each point on the transformed shape model to the surface ITAnd is used to describe registration errors, the second term being a regularization term for statistical shape model deformation, for penalizing the degree of model deformation.
Preferably, the reference MR image I is obtained in step (2) by elastic registration of 3D imagesrefMapping to target CT image ItarFitting a parameterized deformation field by using a B spline; elastic registration of the two images can be achieved by solving the optimal transformation T and calculating according to equation (4):
after obtaining the optimized transformation, obtaining a deformation field between the MR and CT images; then, the reference shape is transformed into a target image to realize the fusion of the MR and CT images; the corresponding result is considered as the initial segmentation result of the anterior visual pathway and the internal rectus muscle.
Preferably, the anterior visual pathway and the internal rectus muscle in the step (3) are soft tissues corresponding to a specific gray window in the CT image, and according to this feature, a good enhancement effect is obtained by setting appropriate upper and lower thresholds; bilateral filtering is then used to reduce noise in the enhanced image, and the Sobel operator is employed to extract boundary information of the anterior visual pathway and the internal rectus muscle.
Preferably, in the step (3), after fitting the optic nerve and internal rectus muscle models to the multi-feature constraint surface, driving the optic nerve and optic cross model; driving a statistical shape model through an optimization formula (3) to enable the space position I between the converted model and the multi-feature constraint surfaceSThe consistency is achieved; finally, the segmentation of optic nerve and internal rectus muscle parts and the prediction of optic bundles and optic cross parts are realized.
It will be understood by those skilled in the art that all or part of the steps in the method of the above embodiments may be implemented by hardware instructions related to a program, the program may be stored in a computer-readable storage medium, and when executed, the program includes the steps of the method of the above embodiments, and the storage medium may be: ROM/RAM, magnetic disks, optical disks, memory cards, and the like. Therefore, in accordance with the method of the present invention, the present invention also includes a system for segmentation of rectus internus muscle and optic nerve of CT images, which is generally represented in the form of functional blocks corresponding to the steps of the method. The system comprises:
a statistical shape model construction module configured to train a shape correspondence of the data set, and construct a statistical shape model of the training shape using principal component analysis;
a MR/CT image fusion-based segmentation module configured to obtain a shape of a reference MR image by fitting a statistical shape model to a segmentation result of the MR image, the CT image being fused with the MR image by elastic registration to obtain an initial segmentation result of the anterior visual pathway and the medial rectus muscle;
a multi-feature constrained segmentation refinement module configured to obtain a multi-feature constrained surface from the target CT image, and, after fitting the initial segmentation results to the surface, segment structures not visible in the CT image, including the optic beam and the optic cross.
Preferably, the MR/CT image fusion based segmentation module performs: random selection of reference MR image I from training datarefAnd its corresponding segmentation image; by fitting a statistical shape model to the reference MR image IrefIs divided intoTTo obtain a reference shape; obtaining a reference MR image I by elastic registration of 3D imagesrefMapping to target CT image ItarFitting a parameterized deformation field by using a B spline; elastic registration of the two images can be achieved by solving the optimal transformation T.
Preferably, the multi-feature constrained segmentation refinement module performs:
the anterior visual pathway and the internal rectus muscle are soft tissues corresponding to a specific gray window in the CT image, and according to this feature, a good enhancement effect is obtained by setting appropriate upper and lower thresholds;
then, reducing noise in the enhanced image by utilizing bilateral filtering, and extracting boundary information of a foresight path and an internal rectus muscle by adopting a Sobel operator;
after fitting the optic nerve and internal rectus muscle models to the multi-feature constraint surface, driving the optic nerve and optic cross model; driving a statistical shape model through an optimization formula (3) to enable the space position I between the converted model and the multi-feature constraint surfaceSThe consistency is achieved; finally, the segmentation of optic nerve and internal rectus muscle parts and the prediction of optic bundles and optic cross parts are realized.
The present invention is described in more detail below.
The invention provides an anatomical shape model based on multi-modal image fusion, which is used for low-contrast anterior visual pathway and internal rectus muscle segmentation in a CT image, and the detailed flow is shown in figure 1. First, the statistical shape model is composed of a training dataset in which the shape of the anterior visual pathway and the internal rectus muscle are manually delineated. Second, the shape of the reference MR image is obtained by fitting a statistical shape model to the segmentation result of the MR image. The CT image is then fused with the MR image by elastic registration to obtain an initial segmentation result of the anterior visual pathway and the internal rectus muscle. Finally, a multi-feature constraining surface is obtained from the target CT image. After fitting the initial segmentation results to the surface, structures not visible in the CT image, including the bundle and cross-views, may also be segmented.
The contribution of the proposed method is two-fold: first, the MR data set is used to construct a prior shape model to assist in the segmentation of CT images of the structure. Due to the weakness of soft tissue CT imaging, it can effectively localize the optic chiasm and optic nerve bundle that are not imaged clearly in CT. Secondly, the multi-feature constrained surface can effectively make up the weakness of lack of local information in multi-modal fusion. It effectively improves the segmentation accuracy.
(1) Constructing statistical shape models
In order to build a statistical shape model, the shape correspondence of the data set needs to be trained. Shape correspondence may be represented as a dense mapping between a set of shape points in the MR data set. The correspondence of the two shapes can be obtained by a pair-wise non-rigid registration. For MR data sets
Unbiased point correspondences can be obtained by packet shape registration, and similarity metric shape registration can be obtained by packet-level means as:
where N is the number of training data, d (.) is the Euclidean distance, gijIs the connection between the ith and jth shapes in the dataset. The connection relationship between all shapes is represented by a graphical model, and then packet-level registration can be achieved through guidance of the graphical model. Finally, the corresponding relation of the shape is obtained
On the basis of whichThe alignment shape is analyzed using generalized equations. And constructing a statistical shape model of the training shape by adopting principal component analysis. Performing eigenvalue decomposition on the matrix, wherein
Are vectorized and then arranged together, and the feature vectors are arranged according to a descending order of feature values. The first few feature vectors are used to model the shape data. Thus, the statistical shape model can be expressed as:
where P represents a vectoring matrix vec (P), vectoring average shape
And the principal eigenmodes form a matrix Φ that is pre-computed from the training dataset, where b represents the parameters of the model.
(2) Segmentation based on MR/CT image fusion
Random selection of reference MR image I from training datarefAnd its corresponding segmented image. Can be generated by fitting a statistical shape model to the reference MR image IrefIs divided intoTTo obtain the reference shape. This process, called surface fitting, can be expressed as:
wherein DTIs ITThe distance of (a) is transformed,
coordinates representing points on the statistical shape model, diag (λ) representing the diagonal composed of eigenvalues λAnd (4) matrix. b is constrained in a hyper-rectangle defined by β and λ, where λiIs the i-th element of λ, biIs the ith parameter in b. The first term in equation (3) is from each point on the transformed shape model to the surface ITAnd is used to describe registration errors. The second term is a regularization term for statistical shape model deformation, which penalizes the degree of model deformation.
Obtaining a reference MR image I by elastic registration of 3D imagesrefMapping to target CT image ItarThe fusion of the MR/CT images is realized at the same time. The normalized mutual information is considered as a similarity measure between the two images. This patent uses a B-spline to fit the parametric deformation field. Elastic registration of the two images can be achieved by solving the optimal transformation T and is calculated as follows:
after obtaining the optimized transformation, a deformation field between the MR and CT images can be obtained. The reference shape is then transformed into the target image to achieve fusion of the MR and CT images. The corresponding result is considered as the initial segmentation result of the anterior visual pathway and the internal rectus muscle.
(3) Multi-feature constrained segmentation refinement
The anterior visual pathway and internal rectus muscle are soft tissue corresponding to a particular gray scale window in the CT image. According to this feature, a good enhancement effect can be obtained by setting appropriate upper and lower thresholds. Thus, the image is enhanced. The contrast of the anterior visual pathway and the internal rectus muscle in the enhanced image is improved compared to the contrast of the anterior visual pathway and the internal rectus muscle in the original CT image. Bilateral filtering is then used to reduce noise in the enhanced image, and the Sobel operator is employed to extract boundary information of the anterior visual pathway and the internal rectus muscle. Constraints on the size of the connected component domain can effectively eliminate the effects of noise, and constraints from the initial segmentation can ensure that most of the extracted surface belongs to the anterior visual pathway and the internal rectus muscle.
Modeling the optic nerve and the internal rectus muscleAfter fitting the profiles to the multi-feature constraint surface, the optic nerve and optic chiasm model can also be driven. Driving a statistical shape model through an optimization formula (3) to enable the space position I between the converted model and the multi-feature constraint surfaceSConsistency is achieved. Finally, the segmentation of optic nerve and internal rectus muscle parts and the prediction of optic bundles and optic cross parts are realized.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and all simple modifications, equivalent variations and modifications made to the above embodiment according to the technical spirit of the present invention still belong to the protection scope of the technical solution of the present invention.
Claims (10)
1. A method for segmenting internal rectus muscle and optic nerve of a CT image is characterized by comprising the following steps: which comprises the following steps:
(1) constructing a statistical shape model: the statistical shape model is composed of a training data set, wherein the shape of the anterior visual pathway and the internal rectus muscle are manually delineated;
(2) segmentation based on MR/CT image fusion: obtaining a shape of a reference MR image by fitting a statistical shape model to a segmentation result of the MR image, and fusing a CT image with the MR image through elastic registration to obtain an initial segmentation result of a anterior visual pathway and an internal rectus muscle;
(3) multi-feature constraint segmentation refinement: a multi-feature constraining surface is obtained from the target CT image, and structures not visible in the CT image, including the optic beam and the optic cross, are segmented after fitting the initial segmentation results to the surface.
2. The method for segmentation of rectus interna and optic nerve of a CT image as claimed in claim 1, wherein: in the step (1), in order to construct a statistical shape model, the shape correspondence of the data set is trained:
shape correspondence is expressed as a dense mapping between a set of shape points in the MR data set, the correspondence of the two shapes being obtained by pairwise non-rigid registration; for MR data sets
Unbiased point correspondences are obtained by grouping shape registration, and similarity measure shape registration obtained by grouping level is expressed as formula (1):
where N is the number of training data, d (.) is the Euclidean distance, gijIs the connection between the ith and jth shapes in the dataset;
the connection relation among all shapes is represented by a graphic model, and then the group level registration is realized through the guidance of the graphic model;
obtaining shape correspondence
On this basis, the alignment shape is analyzed using the generalized equation.
3. The method for segmentation of rectus interna and optic nerve of a CT image as claimed in claim 2, wherein: in the step (1), a statistical shape model of the training shape is constructed by adopting principal component analysis, and eigenvalue decomposition is carried out on the matrix, wherein
Are vectorized and then arranged together, and the eigenvectors are arranged according to a descending order of eigenvalues, the first few eigenvectors being used to model the shape data, the statistical shape model being formula (2):
where P represents a vectoring matrix vec (P), vectoring average shape
And the principal eigenmodes form a matrix Φ that is pre-computed from the training dataset, where b represents the parameters of the model.
4. The method of segmenting rectus internus and optic nerves of a CT image as set forth in claim 3, wherein: in the step (2), a reference MR image I is randomly selected from the training datarefAnd its corresponding segmentation image; by fitting a statistical shape model to the reference MR image IrefIs divided intoTTo obtain the reference shape, this process is called surface fitting and is expressed by equation (3):
wherein DTIs ITThe distance of (a) is transformed,
coordinates of points on the statistical shape model are represented, and diag (λ) represents a diagonal matrix composed of eigenvalues λ; b is constrained in a hyper-rectangle defined by β and λ, where λiIs the i-th element of λ, biIs the ith parameter in b; the first term in equation (3) is from each point on the transformed shape model to the surface ITAnd is used to describe registration errors, the second term being a regularization term for statistical shape model deformation, for penalizing the degree of model deformation.
5. The method of segmenting rectus internus and optic nerves of a CT image as set forth in claim 4, wherein: the reference MR image I is obtained by elastic registration of the 3D image in the step (2)refMapping to target CT image ItarFitting a parameterized deformation field by using a B spline; elastic registration of the two images can be achieved by solving the optimal transformation T and calculating according to equation (4):
after obtaining the optimized transformation, obtaining a deformation field between the MR and CT images; then, the reference shape is transformed into a target image to realize the fusion of the MR and CT images; the corresponding result is considered as the initial segmentation result of the anterior visual pathway and the internal rectus muscle.
6. The method of segmenting rectus internus and optic nerves of a CT image as set forth in claim 5, wherein: the anterior visual pathway and the internal rectus muscle in the step (3) are soft tissues corresponding to a specific gray window in the CT image, and according to the characteristics, a good enhancement effect is obtained by setting appropriate upper and lower thresholds; bilateral filtering is then used to reduce noise in the enhanced image, and the Sobel operator is employed to extract boundary information of the anterior visual pathway and the internal rectus muscle.
7. The method of segmenting rectus internus and optic nerves of a CT image as set forth in claim 6, wherein: in the step (3), after fitting the optic nerve and internal rectus muscle models to the multi-feature constraint surface, driving the optic nerve and optic cross model; driving a statistical shape model through an optimization formula (3) to enable the space position I between the converted model and the multi-feature constraint surfaceSThe consistency is achieved; finally, the segmentation of optic nerve and internal rectus muscle parts and the prediction of optic bundles and optic cross parts are realized.
8. A system for segmentation of rectus interna and optic nerve of CT images, comprising: it includes: a statistical shape model construction module configured to train a shape correspondence of the data set, and construct a statistical shape model of the training shape using principal component analysis;
a MR/CT image fusion-based segmentation module configured to obtain a shape of a reference MR image by fitting a statistical shape model to a segmentation result of the MR image, the CT image being fused with the MR image by elastic registration to obtain an initial segmentation result of the anterior visual pathway and the medial rectus muscle;
a multi-feature constrained segmentation refinement module configured to obtain a multi-feature constrained surface from the target CT image, and, after fitting the initial segmentation results to the surface, segment structures not visible in the CT image, including the optic beam and the optic cross.
9. The system for medial rectus muscle and optic nerve segmentation of CT images as set forth in claim 8, wherein: the MR/CT image fusion-based segmentation module performs: random selection of reference MR image I from training datarefAnd its corresponding segmentation image; by fitting a statistical shape model to the reference MR image IrefIs divided intoTTo obtain a reference shape; obtaining a reference MR image I by elastic registration of 3D imagesrefMapping to target CT image ItarFitting a parameterized deformation field by using a B spline; elastic registration of the two images can be achieved by solving the optimal transformation T.
10. The system for medial rectus muscle and optic nerve segmentation of CT images as set forth in claim 9, wherein: the multi-feature constraint segmentation refinement module performs:
the anterior visual pathway and the internal rectus muscle are soft tissues corresponding to a specific gray window in the CT image, and according to this feature, a good enhancement effect is obtained by setting appropriate upper and lower thresholds; then, reducing noise in the enhanced image by utilizing bilateral filtering, and extracting boundary information of a foresight path and an internal rectus muscle by adopting a Sobel operator;
after fitting the optic nerve and internal rectus muscle models to the multi-feature constraint surface, driving the optic nerve and optic cross model; driving a statistical shape model through an optimization formula (3) to enable the space position I between the converted model and the multi-feature constraint surfaceSThe consistency is achieved; final realization visionSegmentation of the nerve and internal rectus muscle portions and prediction of the optic tract and the optic chiasm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010891689.7A CN112184720B (en) | 2020-08-27 | 2020-08-27 | Method and system for segmenting internal rectus muscle and optic nerve of CT image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010891689.7A CN112184720B (en) | 2020-08-27 | 2020-08-27 | Method and system for segmenting internal rectus muscle and optic nerve of CT image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112184720A true CN112184720A (en) | 2021-01-05 |
| CN112184720B CN112184720B (en) | 2024-04-23 |
Family
ID=73925308
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010891689.7A Active CN112184720B (en) | 2020-08-27 | 2020-08-27 | Method and system for segmenting internal rectus muscle and optic nerve of CT image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112184720B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113744271A (en) * | 2021-11-08 | 2021-12-03 | 四川大学 | Neural network-based automatic optic nerve segmentation and compression degree measurement and calculation method |
| CN114974518A (en) * | 2022-04-15 | 2022-08-30 | 浙江大学 | Multi-mode data fusion lung nodule image recognition method and device |
| CN116258671A (en) * | 2022-12-26 | 2023-06-13 | 中山大学肿瘤防治中心（中山大学附属肿瘤医院、中山大学肿瘤研究所） | MR image-based intelligent sketching method, system, equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100266170A1 (en) * | 2009-04-20 | 2010-10-21 | Siemens Corporate Research, Inc. | Methods and Systems for Fully Automatic Segmentation of Medical Images |
| CN108898606A (en) * | 2018-06-20 | 2018-11-27 | 中南民族大学 | Automatic division method, system, equipment and the storage medium of medical image |
| CN108961256A (en) * | 2018-07-05 | 2018-12-07 | 艾瑞迈迪医疗科技（北京）有限公司 | Image partition method, operation navigation device, electronic equipment and storage medium |
| CN109961446A (en) * | 2019-03-27 | 2019-07-02 | 深圳视见医疗科技有限公司 | CT/MR three-dimensional image segmentation processing method, device, equipment and medium |
| CN110163847A (en) * | 2019-04-24 | 2019-08-23 | 艾瑞迈迪科技石家庄有限公司 | Liver neoplasm dividing method and device based on CT/MR image |
-
2020
- 2020-08-27 CN CN202010891689.7A patent/CN112184720B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100266170A1 (en) * | 2009-04-20 | 2010-10-21 | Siemens Corporate Research, Inc. | Methods and Systems for Fully Automatic Segmentation of Medical Images |
| CN108898606A (en) * | 2018-06-20 | 2018-11-27 | 中南民族大学 | Automatic division method, system, equipment and the storage medium of medical image |
| CN108961256A (en) * | 2018-07-05 | 2018-12-07 | 艾瑞迈迪医疗科技（北京）有限公司 | Image partition method, operation navigation device, electronic equipment and storage medium |
| CN109961446A (en) * | 2019-03-27 | 2019-07-02 | 深圳视见医疗科技有限公司 | CT/MR three-dimensional image segmentation processing method, device, equipment and medium |
| CN110163847A (en) * | 2019-04-24 | 2019-08-23 | 艾瑞迈迪科技石家庄有限公司 | Liver neoplasm dividing method and device based on CT/MR image |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113744271A (en) * | 2021-11-08 | 2021-12-03 | 四川大学 | Neural network-based automatic optic nerve segmentation and compression degree measurement and calculation method |
| CN113744271B (en) * | 2021-11-08 | 2022-02-11 | 四川大学 | Neural network-based automatic optic nerve segmentation and compression degree measurement and calculation method |
| CN114974518A (en) * | 2022-04-15 | 2022-08-30 | 浙江大学 | Multi-mode data fusion lung nodule image recognition method and device |
| CN116258671A (en) * | 2022-12-26 | 2023-06-13 | 中山大学肿瘤防治中心（中山大学附属肿瘤医院、中山大学肿瘤研究所） | MR image-based intelligent sketching method, system, equipment and storage medium |
| CN116258671B (en) * | 2022-12-26 | 2023-08-29 | 中山大学肿瘤防治中心（中山大学附属肿瘤医院、中山大学肿瘤研究所） | MR image-based intelligent sketching method, system, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112184720B (en) | 2024-04-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11455732B2 (en) | Knowledge-based automatic image segmentation | |
| WO2021088747A1 (en) | Deep-learning-based method for predicting morphological change of liver tumor after ablation | |
| CN112184720B (en) | Method and system for segmenting internal rectus muscle and optic nerve of CT image | |
| Mansoor et al. | Deep learning guided partitioned shape model for anterior visual pathway segmentation | |
| Zhu et al. | Automatic segmentation of the left atrium from MR images via variational region growing with a moments-based shape prior | |
| Wang et al. | Multi-atlas segmentation without registration: a supervoxel-based approach | |
| CN110993065B (en) | Brain tumor keyhole surgery path planning method based on image guidance | |
| US9727975B2 (en) | Knowledge-based automatic image segmentation | |
| CN113538533B (en) | Spine registration method, device and equipment and computer storage medium | |
| WO2023125828A1 (en) | Systems and methods for determining feature points | |
| Morra et al. | Automatic subcortical segmentation using a contextual model | |
| CN115082493A (en) | 3D (three-dimensional) atrial image segmentation method and system based on shape-guided dual consistency | |
| CN111080676B (en) | Method for tracking endoscope image sequence feature points through online classification | |
| EP4293618A1 (en) | Brain identifier positioning system and method | |
| CN113222979A (en) | Multi-map-based automatic skull base foramen ovale segmentation method | |
| CN115461790A (en) | Method and apparatus for classifying structure in image | |
| Yao et al. | Statistical location model for abdominal organ localization | |
| Sun et al. | Using cortical vessels for patient registration during image-guided neurosurgery: a phantom study | |
| Prasad et al. | Skull-stripping with machine learning deformable organisms | |
| Wodzinski et al. | Application of demons image registration algorithms in resected breast cancer lodge localization | |
| Sun | A Review of 3D-2D Registration Methods and Applications based on Medical Images | |
| Yao et al. | Non-rigid registration and correspondence finding in medical image analysis using multiple-layer flexible mesh template matching | |
| CN107730544A (en) | Cerebral vessels registration arrangement based on ICP | |
| Lankton | Localized statistical models in computer vision | |
| Kun | Dense correspondence and statistical shape reconstruction of fractured, incomplete skulls |
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 |