CN106997594B - Method and device for positioning eye tissue - Google Patents
Method and device for positioning eye tissue Download PDFInfo
- Publication number
- CN106997594B CN106997594B CN201610050712.3A CN201610050712A CN106997594B CN 106997594 B CN106997594 B CN 106997594B CN 201610050712 A CN201610050712 A CN 201610050712A CN 106997594 B CN106997594 B CN 106997594B
- Authority
- CN
- China
- Prior art keywords
- mask
- skull
- convex hull
- connected domain
- tissue
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—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
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30041—Eye; Retina; Ophthalmic
Abstract
The invention discloses a method for positioning eye tissues, which comprises the following steps: inputting a head image to be processed, and dividing the head image into an air mask and a skull mask according to the gray value of a pixel point; acquiring a convex hull outline of the skull mask based on morphological characteristics, and generating the skull convex hull mask; determining an in-vivo cavity mask in the air mask according to the skull convex hull mask; combining the skull mask and the in-vivo cavity mask to form a skull filling mask; and subtracting the skull filling mask from the skull convex hull mask to obtain the position of the eye tissue. The invention is based on morphological characteristics, carries out convex hull operation on the skull mask and combines a cone measurement algorithm, can accurately position the eye tissue position and has high operation speed. In addition, the invention also provides a positioning device for the eye tissue.
Description
Technical Field
The invention relates to the field of medical image processing, in particular to a method and a device for positioning eye tissues in a head medical image.
Background
With the development and improvement of medical level, Imaging devices such as Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) devices, which combine medical science and computer science, are increasingly introduced, and the diagnosis and understanding level of diseases is continuously improved. Medical image segmentation is an extremely important step in medical image processing, and can be used for three-dimensional reconstruction of medical images, registration and fusion of medical images, even for a plurality of subsequent works including lesion determination and the like. Provides a reliable and efficient automatic image segmentation tool, and can reduce the workload of medical workers.
The eye is one of the most important human sensory organs, and more than 70% of the external information acquired by human beings comes from the visual system, and soft tissues in the orbit include various tissues such as eyeball (vitreous body and crystalline lens), optic nerve, extraocular muscle and fat. In recent years, segmentation techniques for each soft tissue in the orbit have been increasingly applied to clinical diagnosis and treatment. In the disease diagnosis stage, the eye tissue segmentation can be used for quantitatively analyzing the volume and the edema degree of soft tissues in the orbit, so that doctors can conveniently observe the disease progress, determine the operation time and quantify the operation evaluation effect. However, due to the abundance of sensitive tissue around the eye and the extremely complex structure, both surgery and radiation therapy are prone to irreversible damage to the patient. The eye tissue segmentation can provide a three-dimensional model of the eye tissue, and determine the size, the number and the specific part of the lesion, so as to carry out targeted surgery or radiotherapy planning and optimize the treatment effect. Therefore, the accuracy of the eye image segmentation or positioning directly determines the accuracy of the three-dimensional reconstruction.
In the prior art, Bekes et al propose a method for three-dimensional shape modeling of eye tissue[1]For segmenting eye tissue, but this method requires an experienced physician to manually locate the eyeball center position. Other automated algorithms, such as the atlas-based image registration method proposed by Harrigan et al[2]Which guide the segmentation in the new image using the completed segmentation results in the atlas, typically using a method of image registration. Dobler et al[3]The two methods are combined and an ellipsoid model and atlas registration are used simultaneously for finer segmentation. However, due to poor soft tissue contrast in CT images, small volumes of eye tissue relative to the entire head, and the lack of distinct markers or feature points, registration methods sometimes produce significant deviations. Furthermore, differences between different individuals of ocular tissue increase the difficulty of registration. In view of the above, there is a need for an improved method for positioning eye tissue, which can help to improve the stability and automation of the segmentation algorithm.
[1].Bekes G,Mate E,Nyul L G,et al.Geometrical model-based segmentation of the organs of sight on CT images[J].Medical Physics,2007,35(2):735.
[2].Harrigan R L,Panda S,Asman A J,et al.Robust optic nerve segmentation on clinically acquired computed tomography[J].Journal of Medical Imaging,2014,1(3): 034006.
[3].Dobler B,Bendl R.Precise modelling of the eye for proton therapy of intra-ocular tumours[J].Physics in medicine and biology,2002,47(4):593.
Disclosure of Invention
The invention aims to provide an eye tissue positioning method, which can avoid the manual positioning work of a doctor, and has the advantages of accurate positioning, quick operation and good stability.
The technical scheme adopted by the invention for solving the technical problems is that the method for positioning the eye tissue comprises the following steps:
inputting a head image to be processed, and dividing the head image into an air mask and a skull mask according to the gray value of a pixel point;
acquiring a convex hull outline of the skull mask based on morphological characteristics, and generating the skull convex hull mask;
determining an in-vivo cavity mask in the air mask according to the skull convex hull mask;
combining the skull mask and the in-vivo cavity mask to form a skull filling mask;
and subtracting the skull filling mask from the skull convex hull mask to obtain the position of the eye tissue.
Further, extracting a three-dimensional connected domain of the skull convex hull mask minus the skull filling mask, and further determining the position of the eye tissue according to the morphology of the three-dimensional connected domain.
Further, the further determining the position of the eye tissue according to the form of the three-dimensional connected domain specifically includes: and performing cone measurement operation on the three-dimensional connected domain according to the morphological characteristics of the eye socket in the three-dimensional space, and screening the three-dimensional connected domain where the eye tissue is located.
Further, the concrete process of the cone measure is as follows:
acquiring a detection volume V of the three-dimensional connected domain;
obtaining a bottom area S of the three-dimensional connected domain and a height h of the connected domain, and then obtaining a calculated volume V' of the three-dimensional connected domain as S multiplied by h/3;
calculating the similarity between the detection volume V of the three-dimensional connected domain and the calculation volume V' of the three-dimensional connected domain, and if the similarity is within a set range, determining that the position of the three-dimensional connected domain is the position of the eye tissue; otherwise, judging that the position of the three-dimensional connected domain is not the position of the eye tissue.
Further, the skull convex hull mask is the smallest circumscribed convex hull that contains the contour of the skull mask.
Further, the in vitro air mask is located outside the skull convex hull mask, and the in vivo cavity mask is located inside the skull convex hull mask.
Further, the combined skull mask and the in-vivo cavity mask are subjected to morphological expansion and filling treatment.
Further, the method comprises the step of finely dividing the eye tissue position by adopting a multi-radius sphere detection method or a blood vessel detection method to obtain at least one of an eyeball mask, a lens mask, an extraocular muscle mask and an optic nerve mask.
The invention also provides a positioning device for eye tissue, comprising:
the image input module is used for acquiring a medical image to be processed, wherein the medical image is a head image, and the head image is divided into an air mask and a skull mask according to a pixel gray value;
the eye tissue positioning module is used for acquiring the convex hull outline of the skull mask based on morphological characteristics and generating the skull convex hull mask; determining an in-vivo cavity mask in the air mask according to the skull convex hull mask; combining the skull mask and the in-vivo cavity mask to form a skull filling mask; subtracting the skull filling mask from the skull convex hull mask to obtain the position of the eye tissue;
and the segmentation module is used for performing fine segmentation on the eye tissue position by adopting a multi-radius sphere detection or blood vessel detection method to obtain a soft tissue, wherein the soft tissue is at least one of an eyeball, a crystalline lens, extraocular muscles and an optic nerve.
Further, the head image acquisition module is further included and is used for acquiring CT head images or MR head images.
Compared with the prior art, the invention has the advantages that: during positioning, morphological characteristics that eye sockets are two deepest recesses on the surface of a human skull and prior information that all eye soft tissues are located in the eye sockets are fully utilized, a skull convex hull mask is generated by performing convex hull operation on the skull mask, and the skull mask filling an internal cavity is subtracted from the skull convex hull mask, so that the positions of the eye soft tissues are obtained, atlas registration is not needed, simplicity and effectiveness are realized, and the operation speed is high; according to the form of the soft tissue of the eye, the accuracy of the positioning position is further determined by a cone measurement method, and the stability of the detection method is improved.
Drawings
The invention will be further described with reference to the accompanying drawings in which:
FIG. 1 is a flow chart of a method for positioning ocular tissue according to the present invention;
FIG. 2 is a diagram of a CT head image structure processed by the present invention;
FIG. 3 is a cross-sectional view of the soft tissue of the eye positioned in FIG. 2 using the method of the present invention;
FIG. 4 is a structural diagram of the positioning device for eye tissue according to the present invention.
Detailed Description
The above objects, features and advantages of the present invention will become more apparent from the following detailed description of the preferred embodiments of the present invention when taken in conjunction with the accompanying drawings and examples.
The four walls of the orbit form a deep cavity in the shape of a cone bone, and soft tissues or contents such as eyeballs (vitreous bodies and crystalline lenses), optic nerves, extraocular muscles, fat and the like are arranged inside the deep cavity. Imaging examinations such as CT and MRI are often used for segmentation of eye tissue. A typical image segmentation typically includes two parts: coarse segmentation and fine segmentation. The former also generally includes initial positioning of the organ to be segmented, which is the basis for accurate segmentation. Due to the difference in voxel density of the human body, images of different body parts will differ slightly. Taking a CT image as an example, the CT value of a pixel in the image depends on how much the tissue structure contained in the voxel (rectangular tissue structure corresponding to each pixel in the CT image scanning field of view) absorbs X-rays, and the larger the average voxel density is, the higher the attenuation coefficient is; the smaller the density, the smaller the attenuation coefficient. For a specific tissue structure, the density of the tissue structure is not greatly different in different individuals, and the corresponding CT value is relatively fixed in a certain range. A typical brain CT image can be divided into two large regions: i.e. the circular scan field of view in the center of the image and the area outside the scan field of view, the CT values for the area outside the scan field of view are typically: 2048HU (Hounsfield unit), the influence of the examined region is only the image in the scanning field, and in the case of a craniocerebral CT image, the scanning field includes the air around the skull in addition to the image of the skull of the examined person, and the CT values of each main component region in the scanning field are approximately distributed in the following ranges: the CT value of the air around the skull is-1000 HU, the skull is + 200- +1000HU, the cerebrospinal fluid is-20- +20HU, and the brain tissue is + 20- +40 HU.
The invention provides a method for positioning eye tissues, which is based on three-dimensional morphological characteristics, extracts a head skull and air to position the eye tissue position together, and further screens and finely segments the obtained eye tissue position based on anatomical priori knowledge of the periphery of the eye tissue, and comprises the following specific steps:
and S10, inputting the head image to be processed, and dividing the head medical image into an air mask and a skull mask according to the gray value of the pixel point. The head image may be a CT image or an MRI image, in this embodiment, the CT head image is used as a processing object, a gray value of each pixel point in the CT image is usually represented by a CT value, the CT value represents an attenuation value of the X-ray after passing through the tissue and being absorbed, the CT value of a certain substance is equal to a difference between an attenuation coefficient of the substance and an attenuation coefficient of water, and the difference is multiplied by 1000 after being compared with the attenuation coefficient of water. The CT values of different tissues are different, wherein the CT value of the skull is the highest and is 1000HU, the CT value range of soft tissues is 20-70HU, the CT value of water is 0HU, the CT value of fat is below-50 to-100 HU, and the CT value of air is-1000 HU. In the CT scanning of the human body, the CT value is greatly different from the CT value of the peripheral air, and the bone and air parts in the head image can be extracted by setting the threshold, in this embodiment, the first threshold K is set1 (K1300HU) and a second threshold K2(K2-500HU) for all grey values (CT values) greater than a first threshold K1And extracting the largest three-dimensional connected domain to obtain the skull mask (such as bone regions 11, 12 and 13 in fig. 2), and searching all gray values smaller than the second threshold value K2The pixel points are provided with air masks (e.g. the nasal cavity air 21 and the external air part in fig. 2), andthe pixel points between the first threshold and the second threshold are fat, muscle and various soft tissues.
Before the original image is processed, in order to enhance useful information of the image, the original image may be subjected to smoothing and denoising preprocessing before the processing. The commonly used image smoothing and denoising methods mainly include median filtering, gaussian filtering, bilateral filtering, mean filtering and the like. The invention adopts median filtering to carry out smooth denoising of an image, specifically to sort the gray values of pixels in a sliding window, and replace the gray value of the central pixel of the window with the median, namely i (x, y) ═ med [ x ] x1,x2,…,xn]Wherein x is1，x2，…， xnI (x, y) represents the gray value of the pixel at the center of the sliding window with the coordinate (x, y).
And S20, acquiring the convex hull contour of the skull mask contour based on the morphological characteristics, and generating the skull convex hull mask. A convex hull is a smallest simple polygon enclosing a given convex set, where the points on the polygon are salient points, i.e. only the outermost points of the convex set may be called salient points, while the center points of the convex set may not be called salient points. The convex hull (ConvexHull) can be obtained by Graham scanning (refer to Graham R L. an effective algorithm for determining the resultant convex hull of a fine planar set [ J ]. Information Processing Letters,1972,1(4): 132) or by Jarvis stepping (refer to Jarvis R A. on the identification of the resultant convex hull of a fine set of points in the planar set [ J ]. Information Processing Letters,1973,2(1):18-21), or by fast convex hull (refer to Eddy W F.A new algorithm for displaying sets [ J ]. Information Processing Letters,1973,2(1):18-21) or by Soddy W F.A new algorithm for displaying sets [ J ]. Information Processing Letters, ACreact [ 403 ], TOMS (1974): 7). In an embodiment of the invention, the convex hull obtaining method adopts a Harris corner-based algorithm, sorts the diagonal points from high to low according to scores, reserves partial front-scoring corner points, and finds the minimum circumscribed convex hull of the front-scoring corner points, wherein the minimum circumscribed convex hull is the convex hull of the skull mask outline, namely the skull convex hull mask. The convex hull operation is performed on the bone region, and a polygonal region (the minimum circumscribed convex hull of the contour of the skull mask) surrounded by a plurality of boundary lines 31-34 as shown in fig. 2 can be obtained, namely the skull convex hull mask, wherein the polygonal region contains the skull mask consisting of the regions 11-13 and comprises soft tissues such as a body cavity (region 21), a cheek (region 22), an eye socket (region 41) and a brain (region 42).
And S30, determining an in-vivo cavity mask in the air mask according to the skull convex hull mask, wherein the area covered by the skull convex hull mask and the air mask simultaneously is the in-vivo cavity mask. In actual operation, the internal cavity mask can be determined by performing intersection operation on the skull convex hull mask and the air mask (intersection of the two masks). The body cavity mask (area 21) mainly includes the oral cavity, nasal cavity, and the like.
And S40, combining the skull mask and the in-vivo cavity mask to form the skull filling mask. It should be noted that, in order to fully combine the skull mask and the internal cavity mask, the invention also performs morphological expansion and filling treatment on the combined skull mask and internal cavity mask, and performs morphological cavity filling operation by using the internal cavity mask, so as to obtain the skull filling mask with a sealed interior. Through the operation, the skull filling mask comprises the skull mask and also comprises the body cavity, the brain and other areas.
S50, eye sockets are the anatomy priori knowledge of two deepest depressions on the surface of the human skull, and the position of the eye sockets can be accurately positioned by subtracting the hole filling body of the skull from the convex hull generated by the skull. The skull fill mask is therefore subtracted from the skull convex hull mask to obtain the position of the ocular tissue. As shown in FIG. 2, the skull convex hull mask comprises the skull, the internal cavity, the brain, the cheek and the soft eye tissue, wherein the cheek and the soft eye tissue belong to the soft tissue outside the skull. The skull filling mask fills the entire area within the skull, including the skull, the body cavity, and the brain. Thus, the two masks are subtracted to obtain a location that includes soft tissue of the eye. It should be noted that the position of the eye tissue (soft tissue) obtained at this time may also include the extracranial soft tissue regions (22) of the cheeks on both sides, such as the facial muscles. Therefore, the invention further extracts the residual three-dimensional connected domain of the skull convex hull mask minus the skull filling mask, and determines whether the position of the connected domain is the position of the eye tissue according to the form of the three-dimensional connected domain. As the eye socket of a normal person is in a three-dimensional cone shape, the form of the three-dimensional connected domain can be detected by adopting a cone measurement method, and whether the position of the three-dimensional connected domain is the position of a target organ (eye tissue) or not can be determined. The method specifically comprises the following steps:
counting the number of all pixel points in the three-dimensional connected domain obtained by subtracting the skull filling mask from the skull convex hull mask, thereby obtaining the detection volume V of the connected domain;
and obtaining the bottom area S of the three-dimensional connected domain and the height h of the three-dimensional connected domain, and obtaining the calculated volume V' of the connected domain as S multiplied by h/3 according to the bottom area S and the height h of the three-dimensional connected domain, wherein the bottom area of the three-dimensional connected domain is a set of pixel points with the distance from all connected domain points to an external air mask being 1 in a set, and the height of the three-dimensional connected domain is the distance from the pixel point with the maximum distance from the external air mask to the external air mask in the set.
Calculating the similarity or difference between the detection volume V of the three-dimensional connected domain and the calculation volume V' of the three-dimensional connected domain, wherein the similarity can be obtained by calculating the ratio of the two, comparing whether the calculated similarity or difference is within a set range (the ratio is close to 1), and if the similarity or difference is within the set range, judging that the position of the connected domain is the position of the eye tissue; otherwise, the connected domain may be soft tissue regions on both sides of the cheek, and the position of the connected domain is determined not to be the eye tissue position but to belong to the cheek soft tissue position. By the above operation, the interference of the region 22 on the positioning result can be eliminated, and two conical regions, namely, the positions of the eye soft tissue, as shown in fig. 3, can be obtained.
The above operations can be used for accurately positioning the soft tissue in the eye socket in the CT image, all 24 groups of head images are randomly selected to be accurately positioned, and for 100 groups of lower-half brain data (the eye socket part in the image of the lower-half brain has defects, only contains part of the eye socket, and the shape of the eye socket is not conical), the correct positioning is 97 groups, the correct rate can still reach 97%, and the better stability is achieved. In addition, multiple organ tissues including eyeballs, crystalline lenses, extraocular muscles, optic nerves and the like in the eye sockets can be further accurately segmented by adopting methods such as multi-radius sphere detection and blood vessel detection. The head image processing method can efficiently position and automatically segment the eye soft tissue of the medical image without manually selecting seed points, avoids using atlas registration, improves the positioning and segmentation speed, and has reliable performance.
In addition, the present invention also provides a positioning device for ocular tissue, as shown in fig. 4, comprising:
the image input module 100 is configured to acquire a medical image to be processed, where the medical image is a CT head image or an MR head image, and divide the head image into an air mask and a skull mask according to a gray value of a pixel point.
The eye tissue positioning module 200 is connected with the image input module 100 and is used for acquiring a convex hull outline of the skull mask based on morphological characteristics and generating the skull convex hull mask; determining a mask of an internal cavity in an air mask according to the skull convex hull mask; combining the skull mask and the in-vivo cavity mask to form a skull filling mask; the skull convex hull mask is used to subtract the skull filling mask to obtain the position of the eye tissue.
The ocular tissue localization module 200 may also implement the following screening functions: extracting the three-dimensional connected domain of the skull convex hull mask minus the skull filling mask, respectively obtaining a detection volume V and a calculation volume V 'of the three-dimensional connected domain, calculating the similarity of the detection volume V and the calculation volume V', and determining the position of the eye tissue in the three-dimensional connected domain according to the similarity.
The segmentation module 300 is connected to the eye tissue positioning module 200, and is configured to perform fine segmentation on the eye tissue position by using a multi-radius sphere detection method to obtain a soft tissue, where the soft tissue may be an eyeball, a crystalline lens, an extraocular muscle, an optic nerve, and the like, and the CT head image or the MRI head image may be processed by the above-mentioned device to display a precise positioning result of the eye tissue organ on a display device. It should be noted that the positioning device for eye tissue of the present invention further includes a head image acquiring module, connected to the image input module, for acquiring a CT head image or an MR head image, and taking the acquired image as an input of the image input module 100.
It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (10)
1. A method of positioning ocular tissue, comprising the steps of:
inputting a head image to be processed, and dividing the head image into an air mask and a skull mask according to the gray value of a pixel point;
acquiring a convex hull outline of the skull mask based on morphological characteristics, and generating the skull convex hull mask;
determining an in-vivo cavity mask in the air mask according to the skull convex hull mask;
combining the skull mask and the in-vivo cavity mask to form a skull filling mask;
and subtracting the skull filling mask from the skull convex hull mask to obtain the position of the eye tissue.
2. The method of claim 1, further comprising extracting a three-dimensional connected domain of the skull convex hull mask minus the skull fill mask, and further determining the location of the ocular tissue based on the morphology of the three-dimensional connected domain.
3. The method according to claim 2, wherein the further determining the position of the ocular tissue according to the morphology of the three-dimensional connected domain specifically comprises: and performing cone measurement operation on the three-dimensional connected domain according to the morphological characteristics of the eye socket in the three-dimensional space, and screening the three-dimensional connected domain where the eye tissue is located.
4. The method for locating ocular tissue according to claim 3, wherein the specific process of the conic measure is as follows:
acquiring a detection volume V of the three-dimensional connected domain;
obtaining a bottom area S of the three-dimensional connected domain and a height h of the three-dimensional connected domain, and then obtaining a calculated volume V' of the three-dimensional connected domain as S multiplied by h/3;
calculating the similarity between the detection volume V of the three-dimensional connected domain and the calculation volume V' of the three-dimensional connected domain, and if the similarity is within a set range, determining that the position of the three-dimensional connected domain is the position of the eye tissue; otherwise, judging that the position of the three-dimensional connected domain is not the position of the eye tissue.
5. The method of claim 1, wherein the skull convex hull mask is a minimal circumscribed convex hull containing the contour of the skull mask.
6. The method for positioning ocular tissue according to claim 4, wherein the air mask further comprises an external air mask, the external air mask is located outside the skull convex hull mask, the internal cavity mask is located inside the skull convex hull mask, a bottom area S of the three-dimensional connected domain is a set of pixel points of which the distances from the three-dimensional connected domain points to the external air mask are all 1, and a height h of the three-dimensional connected domain is a distance between a pixel point of which the distance from the three-dimensional connected domain points to the external air mask is the maximum and the external air mask.
7. The method of claim 1, further comprising morphologically expanding and filling the combined skull mask and body cavity mask.
8. The method for positioning ocular tissue according to any one of claims 1 to 7, further comprising obtaining at least one of an eye mask, a lens mask, an extraocular muscle mask, and an optic nerve mask by performing a fine segmentation on the ocular tissue location using multi-radius sphere detection or blood vessel detection.
9. A positioning device for ocular tissue, comprising:
the image input module is used for acquiring a medical image to be processed, wherein the medical image is a head image, and the head image is divided into an air mask and a skull mask according to a pixel gray value;
the eye tissue positioning module is used for acquiring the convex hull outline of the skull mask based on morphological characteristics and generating the skull convex hull mask; determining an in-vivo cavity mask in the air mask according to the skull convex hull mask; combining the skull mask and the in-vivo cavity mask to form a skull filling mask; subtracting the skull filling mask from the skull convex hull mask to obtain the position of the eye tissue;
and the segmentation module is used for performing fine segmentation on the eye tissue position by adopting a multi-radius sphere detection or blood vessel detection method to obtain a soft tissue, wherein the soft tissue is at least one of an eyeball, a crystalline lens, extraocular muscles and an optic nerve.
10. The device of claim 9, further comprising a head image acquisition module for acquiring a CT head image or an MR head image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610050712.3A CN106997594B (en) | 2016-01-26 | 2016-01-26 | Method and device for positioning eye tissue |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610050712.3A CN106997594B (en) | 2016-01-26 | 2016-01-26 | Method and device for positioning eye tissue |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106997594A CN106997594A (en) | 2017-08-01 |
| CN106997594B true CN106997594B (en) | 2020-12-01 |
Family
ID=59428599
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610050712.3A Active CN106997594B (en) | 2016-01-26 | 2016-01-26 | Method and device for positioning eye tissue |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106997594B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108447044B (en) * | 2017-11-21 | 2022-01-28 | 四川大学 | Osteomyelitis lesion analysis method based on medical image registration |
| CN108520512B (en) * | 2018-03-26 | 2020-11-20 | 北京医拍智能科技有限公司 | Method and device for measuring eye parameters |
| CN108648172B (en) * | 2018-03-30 | 2021-08-03 | 四川元匠科技有限公司 | CT (computed tomography) map pulmonary nodule detection system based on 3D-Unet |
| CN109509179B (en) * | 2018-10-24 | 2023-03-28 | 深圳市旭东数字医学影像技术有限公司 | Automatic segmentation method and system for eyeballs and crystalline lenses based on medical images |
| CN109872333B (en) | 2019-02-20 | 2021-07-06 | 腾讯科技（深圳）有限公司 | Medical image segmentation method, medical image segmentation device, computer equipment and storage medium |
| WO2020186514A1 (en) * | 2019-03-21 | 2020-09-24 | Covidien Lp | Organ segmentation method and system |
| WO2023186133A1 (en) * | 2022-04-02 | 2023-10-05 | 武汉联影智融医疗科技有限公司 | System and method for puncture path planning |
| CN116152253B (en) * | 2023-04-20 | 2023-06-16 | 四川大学华西医院 | Cardiac magnetic resonance mapping quantification method, system and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103854284A (en) * | 2014-02-26 | 2014-06-11 | 苏州大学 | Cutting method for serous pigment epithelium layer disengagement retina based on three-dimensional diagram searching |
| CN104173071A (en) * | 2013-05-28 | 2014-12-03 | Ge医疗系统环球技术有限公司 | Method and device for automatically determining scanning range of petrous bone |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7333644B2 (en) * | 2003-03-11 | 2008-02-19 | Siemens Medical Solutions Usa, Inc. | Systems and methods for providing automatic 3D lesion segmentation and measurements |
| CN104933729B (en) * | 2014-03-18 | 2017-10-03 | 上海联影医疗科技有限公司 | The extracting method and device of a kind of intracranial tissues |
-
2016
- 2016-01-26 CN CN201610050712.3A patent/CN106997594B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104173071A (en) * | 2013-05-28 | 2014-12-03 | Ge医疗系统环球技术有限公司 | Method and device for automatically determining scanning range of petrous bone |
| CN103854284A (en) * | 2014-02-26 | 2014-06-11 | 苏州大学 | Cutting method for serous pigment epithelium layer disengagement retina based on three-dimensional diagram searching |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106997594A (en) | 2017-08-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106997594B (en) | Method and device for positioning eye tissue | |
| Fernandez | Delineating fluid-filled region boundaries in optical coherence tomography images of the retina | |
| US7817836B2 (en) | Methods for volumetric contouring with expert guidance | |
| Lu et al. | Automatic re-contouring in 4D radiotherapy | |
| US8275182B2 (en) | Method for automated delineation of contours of tissue in medical images | |
| CN107563998B (en) | Method for processing heart image in medical image | |
| Datta et al. | Gray matter segmentation of the spinal cord with active contours in MR images | |
| Tingelhoff et al. | Comparison between manual and semi-automatic segmentation of nasal cavity and paranasal sinuses from CT images | |
| WO2013131420A1 (en) | Device and method for determining boundary of target region of medical image | |
| Linguraru et al. | Liver and tumor segmentation and analysis from CT of diseased patients via a generic affine invariant shape parameterization and graph cuts | |
| EP2689344B1 (en) | Knowledge-based automatic image segmentation | |
| CN108961278B (en) | Method and system for abdominal wall muscle segmentation based on image data | |
| US20150206300A1 (en) | Method and apparatus for extraction and quantification of hematoma from a brain scan such as computed tomography data | |
| Ge et al. | Accurate localization of cortical convolutions in MR brain images | |
| Zhu et al. | 3D automatic MRI level set segmentation of inner ear based on statistical shape models prior | |
| Memiş et al. | Computerized 2D detection of the multiform femoral heads in magnetic resonance imaging (MRI) sections with the integro-differential operator | |
| Fabijańska et al. | Assessment of hydrocephalus in children based on digital image processing and analysis | |
| Memiş et al. | A new scheme for automatic 2D detection of spheric and aspheric femoral heads: A case study on coronal MR images of bilateral hip joints of patients with Legg-Calve-Perthes disease | |
| US11769253B2 (en) | Method and system for selecting a region of interest in an image | |
| Eskildsen et al. | Quantitative comparison of two cortical surface extraction methods using MRI phantoms | |
| Ed-dhahraouy et al. | Proposition of local automatic algorithm for landmark detection in 3D cephalometry | |
| EP4345747A1 (en) | Medical image data processing technique | |
| AU2019204372B1 (en) | Method and System for Selecting a Region of Interest in an Image | |
| Heimann et al. | Prostate segmentation from 3D transrectal ultrasound using statistical shape models and various appearance models | |
| Kuhn | Aim project a2003: Computer vision in radiology (covira) |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 201807 Shanghai city Jiading District Industrial Zone Jiading Road No. 2258 Applicant after: Shanghai Lianying Medical Technology Co., Ltd Address before: 201807 Shanghai city Jiading District Industrial Zone Jiading Road No. 2258 Applicant before: SHANGHAI UNITED IMAGING HEALTHCARE Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |