CN106683078A - Knowledge based spinal cord automatic extracting method - Google Patents
Knowledge based spinal cord automatic extracting method Download PDFInfo
- Publication number
- CN106683078A CN106683078A CN201611130965.8A CN201611130965A CN106683078A CN 106683078 A CN106683078 A CN 106683078A CN 201611130965 A CN201611130965 A CN 201611130965A CN 106683078 A CN106683078 A CN 106683078A
- Authority
- CN
- China
- Prior art keywords
- spinal cord
- region
- circle
- point
- extraction method
- 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.)
- Pending
Links
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
- G06T2207/20—Special algorithmic details
- G06T2207/20076—Probabilistic image processing
-
- 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
- G06T2207/30008—Bone
- G06T2207/30012—Spine; Backbone
Landscapes
- Engineering & Computer Science (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)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a knowledge based spinal cord automatic extracting method. The method includes three steps: 1, determining a trunk outline; determining a spinal cord probability area based on an area increase point which is found in spinal cord based on a spine structure model; and detecting spinal cord in the spinal cord probability area. According to the scheme provided by the invention, the area increase point of the spinal cord is found directly based on the spine structure model and then the spinal cord probability area is found. The whole spinal cord automatic extraction process is small in step number and high in detection rate. Besides, since the spinal cord probability area is obtained based on area growing of spinal cord attributive characters, said area based extraction of spinal cord is more accurate.
Description
Technical field
The present invention relates to image processing techniquess, and in particular to CT image processing techniques.
Background technology
Radiotherapy is as operation and Drug therapy, it has also become an important means of oncotherapy.Three-dimensional radiation is controlled
It is important core in radiotherapy product to treat planning system.Planning system need to provide it is good enough delineate instrument, it is accurate quick
Sketch out target area and its important sensitive organ of surrounding.At present, this part work also relies primarily on doctor and is delineated manually.
But often due to plan needs the target area delineated and to jeopardize organ more so that this part work becomes especially loaded down with trivial details, needs
Wanting the careful work of some time just can complete.Therefore, in the last few years, the either plan system of academia or some international mainstreams
System software, for the automatic segmentation for jeopardizing organ is all paid special attention to.In various organ-tissues, spinal cord is one very important
Sensitive organ, will as far as possible receive few exposure dose during roentgenotherapia.Otherwise be easy to cause it is unpredictable as
Nerves complication in terms of the spinal cords such as paralysis.For spinal cord, main at present still in layer image is delineated manually.To the greatest extent
Pipe also has some semi-automatic or full-automatic method appearance, but often due to automatically extracts whole identification process excessively complexity and make
Discrimination is relatively low, and the time is longer, it is impossible to is clinically applied well and is approved.Therefore the spinal cord based on CT images is automatic
Delineate and needed to be further improved in clinical practice.
Referring to Fig. 1, current spinal cord automatic detection process is main in four steps：(1) body profile is found, (2) find epimere,
(3) canalis spinalis is found, (4) find spinal cord.This four steps are orders, and latter step relies on back, so accuracy rate is to take advantage of
Product relation.Because the 2nd step verification and measurement ratio is typically relatively low, and total verification and measurement ratio is four step products, so as to cause last verification and measurement ratio
It is low；And because the 2nd step and the 3rd step are more complicated, cause whole automatic detection process time-consuming longer.
Because the common method that automatically extracts of spinal cord is first to extract the bone structures such as the peripheral spinal canal of spinal cord, it is then based on
The similar approach that the profile and M.Held and D.T.Lee et al. of the bone structures such as the spinal canal of periphery are provided carries out spinal cord detection.But
The bone structure automatic detection processes such as the spinal canal due to spinal cord periphery are more complicated, and step is more, and often verification and measurement ratio is relatively low, spinal canal
Verification and measurement ratio there was only 85% or so, the automatic detection of spinal cord is dependence with it, so as to the verification and measurement ratio for causing spinal cord is reduced.
The content of the invention
For existing spinal cord automatic detection scheme there are problems that verification and measurement ratio it is low, need a kind of efficient spinal cord
Automatic detection scheme.
For this purpose, the technical problem to be solved is to provide a kind of Knowledge based engineering spinal cord extraction method, carry
The verification and measurement ratio of high spinal cord automatic detection.
In order to achieve the above object, the Knowledge based engineering spinal cord extraction method that the present invention is provided, including：
Determine trunk profile；
Determine spinal cord Probability Region, be primarily based on the region growth point that ridge structure model is found in spinal cord, be then based on region increasing
Long point obtains spinal cord Probability Region；
Spinal cord is detected in spinal cord Probability Region.
In the method, trunk profile is determined as follows：
(1) the corresponding gradient of picture element in CT images is calculated；
(2) search for by the picture centre row the first row pixel, along column direction find first gradient more than threshold epsilon as
Vegetarian refreshments, threshold epsilon is to determine the boundary point pixel of trunk in image；
(3) from the beginning of the picture element found from step (2), picture element of the gradient more than threshold epsilon is searched for along clockwise direction, directly
To first point.
In this programme, spinal cord Probability Region is determined as follows：
(1) some detection zones in spinal cord are determined；
(2) based in ridge structure model inspection spinal cord a bit, as region growth point；
(3) growth that a little carries out detected based on step (2) obtains spinal cord Probability Region.
In this programme, first trunk bounding box is determined according to trunk profile in the step (1)；Then further according to encirclement
Box and axis determine some detection zones in spinal cord.
In this programme, the ridge structure model is based on spinal cord and its vertebrae architectural feature knowledge composition of surrounding wrapping.
Further, the ridge structure model includes 4 basic configuration units：E0, E1, E2 and E3, basic configuration unit E0, E1, E2 and
E3 is used to characterize the feature of ridge structure model.
In this programme, some detection zones move E0 unit centers in ridge structure model in spinal cord in the step (2),
Until E0 unit centers meet ridge structure model attributes feature request, now the corresponding point of E0 unit centers is in spinal cord a bit.
In this programme, in the step (3) based in the spinal cord that detects a little and spinal cord CT value scopes in spinal cord
Region growth is carried out in the limited area of Probability Region, the region for obtaining is spinal cord Probability Region.
In this programme, by connecing greatest circle in search spinal cord Probability Region border spinal cord is detected.
In this programme, as follows detecting spinal cord：
In the range of human body spinal cord radius actual size, different radii is defined successively by radius actual size is descending
Search circle template；
Then the descending mobile search circle template center in spinal cord Probability Region of radius size is pressed, search circle is met first
The point for inside including belongs to the circle of spinal cord Probability Region pixel and is the target circle for detecting；If being not detected by target circle, move
The minimum template circle center of dynamic radius, the point included in circle belongs to the most circle of spinal cord Probability Region pixel as target circle.
The scheme that the present invention is provided is primarily based on the region growth point that ridge structure model directly finds spinal cord, then obtains spinal cord
Probability Region, it is few that whole spinal cord automatically extracts process steps, and verification and measurement ratio is high, and because spinal cord Probability Region is based on spinal cord attribute character
Carry out region and increase to obtain, so the spinal cord obtained based on this region is more accurate.
Meanwhile, this programme realizes automatically extracting for spinal cord, and clinical radiation therapy doctor can be avoided to carry out manually in practical application
The tedious work delineated, while the dependence for delineating result to radiotherapy doctor's technical merit is also reduced, in clinical radiotherapy neck
Domain is significant.
Description of the drawings
The present invention is further illustrated below in conjunction with the drawings and specific embodiments.
Fig. 1 is existing spinal cord automatic detection schematic diagram；
Fig. 2 is the flow chart of spinal cord detection in present example；
Fig. 3 is the schematic diagram of spinal cord detection in present example；
Fig. 4 is spinal cord Probability Region detects schematic diagram in present example；
Fig. 5 is structure model schematic in present example；
Fig. 6 is present example median ridge structure model adaptation figure；
Fig. 7 is to increase the spinal cord Probability Region design sketch for obtaining in present example；
Fig. 8 is the design sketch of the spinal cord detected in present example.
Specific embodiment
In order that technological means, creation characteristic, reached purpose and effect that the present invention is realized are easy to understand, tie below
Conjunction is specifically illustrating, and the present invention is expanded on further.
The spinal cord that this example is provided automatically extracts scheme, based on the feature knowledge of CT image organizations of human body, in detection spinal cord
In Probability Region, based on spinal cord and its feature knowledge of surrounding structure, setting up corresponding characteristic model (i.e. ridge structure model) carries out ridge
The detection of any in marrow, and carry out region based on the point and increase to obtain spinal cord Probability Region, and then realize ridge in spinal cord Probability Region
The detection of marrow.
Under this principle, this example can be automatic for spinal cord to realize by setting up structure and two kinds of knowledge models of program
Extract.
Wherein, structure knowledge model：The definition of the histoorgan being related to mainly is automatically extracted to spinal cord and spinal cord, is wrapped
Include the contact between attribute and each histoorgan such as shape, position, CT value scopes.
And procedural knowledge model：It is mainly used in decomposition and Row control that spinal cord automatically extracts subtask.
Knowledge based model, procedure decomposition is automatically extracted into three below committed step in this example approach by spinal cord
(referring to Fig. 2)：
1. trunk profile is determined.
2. spinal cord Probability Region is determined, it is based on the feature knowledge of spinal cord and its surrounding structure, sets up corresponding characteristic model
(i.e. ridge structure model) carries out the detection of any in spinal cord, and carries out region based on the point and increase to obtain spinal cord Probability Region.
3. spinal cord is detected in spinal cord Probability Region.
Referring to Fig. 3, it show the schematic diagram of spinal cord detection, the trunk wheel wherein in figure in region (1) correspondence step 1
Exterior feature, the spinal cord Probability Region in region (2) correspondence step 2, the spinal cord in region (3) correspondence step 3.
It follows that in three committed steps in this spinal cord Detection and Extraction scheme there is dependence in corresponding region,
Wherein trunk profile is the references object of spinal cord Probability Region, and spinal cord Probability Region is the references object of spinal cord.Accordingly, it would be desirable to three
Individual step has time sequencing.Specifically, trunk profile is first detected, then, based between spinal cord and spinal cord and trunk
Contact determines spinal cord Probability Region, finally in spinal cord Probability Region, based on the knowledge model of spinal cord spinal cord is searched.
For such scheme, detailed description below once its realize process (referring to Fig. 2).
Step (one) trunk contour detecting, that is, determine trunk profile.
Because the bed board of CT equipment in scanning process is flat Carbon fibe plate, so for radiocurable CT images
In, body skin part is with periphery CT value difference not clearly.Therefore, in image the boundary point pixel of trunk should be greater than it is a certain
Particular value ε, therefore this example determines concrete ε values by the histogram analysis to trunk and air CT values, is determined with the simultaneously ε values
The boundary point pixel of trunk in image.
Accordingly, this example realizes that process is as follows for what CT images carried out trunk contour detecting：
(1) the corresponding gradient of picture element in image is calculated.
(2) search for by picture centre row the first row pixel, along column direction pixel of first gradient more than ε is found
Point.
(3) from the beginning of picture element of first gradient searched from step (2) more than ε, gradient is searched for along clockwise direction
Picture element more than ε, until first point.
Trunk profile (as shown in Figure 3) can determine that based on the pixel for searching.Meanwhile, the trunk profile for thus extracting
Verification and measurement ratio to improve trunk profile provides corresponding space.
Step (two) determines spinal cord Probability Region.
Automatically extract in scheme in spinal cord, determine that spinal cord Probability Region is a most key step, this programme is specifically based on ridge
Marrow and its vertebrae architectural feature knowledge of surrounding wrapping, building corresponding structural model is used to detect in spinal cord a bit, and base
Growth is carried out in the point obtain spinal cord Probability Region.
Accordingly, detection concrete point following three step in spinal cord Probability Region is carried out in this example approach (referring to Fig. 4)：
(1) some detection zones in spinal cord are determined.
(2) detect in spinal cord a bit.
(3) spinal cord Probability Region is determined.
Wherein, some detection zones are mainly determined according to trunk profile in spinal cord.First, determined according to trunk profile
Trunk bounding box (see Fig. 3)；Then some detection zones in spinal cord, the side seen in Fig. 4 are determined further according to bounding box and axis
Shape region, the distance demarcated in figure is, it is ensured that the certain point in spinal cord is in the region.
Some detection is and finds a pixel in spinal cord, makes this pixel belong to the certain point in spinal cord, to be based on
The point carries out the growth of spinal cord Probability Region.
In order to carry out the certain point in detection spinal cord, this example is based on spinal cord and its vertebrae architectural feature of surrounding wrapping
Knowledge architecture is corresponding " ridge structure model ", and carries out the detection of any in spinal cord in some detection zones in spinal cord with this.
Referring to Fig. 5, the ridge structure model includes 4 basic configuration units, E0, E1, E2 and E3.This four basic configuration units
E0, E1, E2 and E3 are used to characterize the feature of ridge structure model, such as average, variance etc..
Knowledge based, this example selects the combination of triangle and circular or two kinds of shapes as the basic configuration list of model
Unit.Thus the versatility of model and the operability of realization can be ensured.
Furthermore, the CT values of E0 units have three attributes, standard deviation, minima and maximum in the ridge structure model.E1、E2
And the CT values of E3 units have two attributes, minima and maximum.
In addition, the property value of E0, E1, E2 and E3 unit can be organized based on clinically each unit region in ridge structure model
CT value scopes obtain, also can be worth to by the manual mobility model on CT images, statistical correlation attribute.
Referring to Fig. 6, it show the typical spine node composition of CT faultage images of the breast comprising spinal cord, abdomen, hip position.From
As can be seen that this ridge structure model can be good at being adapted to structure of the spinal cord in various CT faultage images in figure.
When carrying out the detection of any in spinal cord based on above-mentioned ridge structure model, some detection zones move ridge structure in spinal cord
E0 unit centers in model, until E0 unit centers meet ridge structure model attributes feature request, now E0 unit centers are corresponding
Point is in spinal cord a bit.
Thus a little can ensure that the point is the certain point in spinal cord in the spinal cord for detecting.On this basis, in order to determine
The particular location and size in the spinal cord center of circle, in addition it is also necessary to which spinal cord Probability Region is determined based on the point.
It is determined that during spinal cord Probability Region, it is necessary first to determine the scope of spinal cord Probability Region.In view of spinal cord section own dimensions
Size, in the spinal cord that spinal cord center one is scheduled on to detect a little centered on the length of side be in 4 centimetres of square area.
For this purpose, this example be the square region as spinal cord Probability Region limited area, and based in the spinal cord for detecting one
The CT value scopes (i.e. with the attribute of the unit of ridge structure model E 0) of point and spinal cord carry out region growth in the square area, obtain
Region be spinal cord Probability Region.As an example, as shown in fig. 7, square region is spinal cord Probability Region limited area in figure, cross
Line is that spinal cord is a little interior, and the irregular area in square region is then spinal cord Probability Region.
Step (three) spinal cord is detected.
After spinal cord Probability Region determines, obtain because the region is increased based on spinal cord knowledge model attribute, then spinal cord
It is the subset in the region.Principle is delineated based on spinal cord on clinical radiotherapy, automatic detection spinal cord is defined as the osteoplastic chamber of ridge
Greatest circle is connect in body, thus, spinal cord detection is completed by connecing greatest circle in search spinal cord Probability Region border in this example.
Because spinal cord Probability Region region of search is less, spinal cord radius also has certain limit, and this example specifically employs as follows
Method：
First, in the range of human body spinal cord radius actual size, difference is defined successively by radius actual size is descending
The search circle template of radius；
Then the descending mobile search circle template center in spinal cord Probability Region of radius size is pressed, search circle is met first
The point for inside including belongs to the circle of spinal cord Probability Region pixel and is the target circle for detecting；If being not detected by target circle, move
The minimum template circle center of dynamic radius, the point included in circle belongs to the most circle of spinal cord Probability Region pixel as target circle.
When implementing, in order to ensure detecting target, preferred spinal cord radius actual size scope be 0.5cm~
2.0cm；And search for circle template radius and be based on pixel actual size psCarry out discrete.Specifically follow the example of as maximum template radius
2.0cm, then successively decreases successively by formula (1)：
rn+1=rn-ps-0.000001 (1)
Wherein r maximums are 2cm, and minima is the maximum in the descending series obtained by formula (1) less than 0.5cm.
In addition, the greatest circle detection scheme that this example is provided is in practical application, will not limited by boundary shape, can
Accurately find target.As an example, as shown in figure 8, the circle in its figure is to correspond to Fig. 7 to detect in spinal cord Probability Region
Greatest circle, that is, spinal cord.
From the foregoing, it will be observed that automatically extracting scheme relative to existing spinal cord generally first extracts the bone knot such as spinal canal of spinal cord periphery
Structure, being then based on the profile of the bone structures such as the spinal canal of periphery carries out spinal cord detection；This example approach adopts contrary strategy, i.e.,
The region growth point that ridge structure model directly finds spinal cord is primarily based on, spinal cord Probability Region is then obtained, step is thus not only reduced,
Also improve verification and measurement ratio；And obtain because spinal cord Probability Region carries out region growth based on spinal cord attribute character, so being based on
The spinal cord that this region obtains is more accurate.
Ultimate principle, principal character and the advantages of the present invention of the present invention has been shown and described above.The technology of the industry
Personnel it should be appreciated that the present invention is not restricted to the described embodiments, the simply explanation described in above-described embodiment and description this
The principle of invention, without departing from the spirit and scope of the present invention, the present invention also has various changes and modifications, these changes
Change and improvement is both fallen within scope of the claimed invention.The claimed scope of the invention by appending claims and its
Equivalent thereof.
Claims (10)
1. a kind of Knowledge based engineering spinal cord extraction method, it is characterised in that methods described includes：
Determine trunk profile；
Determine spinal cord Probability Region, be primarily based on the region growth point that ridge structure model is found in spinal cord, be then based on region growth point
Obtain spinal cord Probability Region；
Spinal cord is detected in spinal cord Probability Region.
2. Knowledge based engineering spinal cord extraction method according to claim 1, it is characterised in that true as follows
Determine trunk profile：
(1) the corresponding gradient of picture element in CT images is calculated；
(2) search for by picture centre row the first row pixel, along column direction pixel of first gradient more than threshold epsilon is found
Point, threshold epsilon is to determine the boundary point pixel of trunk in image；
(3) from the beginning of the picture element found from step (2), picture element of the gradient more than threshold epsilon, Zhi Dao are searched for along clockwise direction
One point.
3. Knowledge based engineering spinal cord extraction method according to claim 1, it is characterised in that true as follows
Determine spinal cord Probability Region：
(1) some detection zones in spinal cord are determined；
(2) based in ridge structure model inspection spinal cord a bit, as region growth point；
(3) growth that a little carries out detected based on step (2) obtains spinal cord Probability Region.
4. Knowledge based engineering spinal cord extraction method according to claim 3, it is characterised in that in the step (1)
First trunk bounding box is determined according to trunk profile；Then some detection zones in spinal cord are determined further according to bounding box and axis
Domain.
5. Knowledge based engineering spinal cord extraction method according to claim 3, it is characterised in that the ridge structure model base
In spinal cord and its vertebrae architectural feature knowledge composition of surrounding wrapping.
6. Knowledge based engineering spinal cord extraction method according to claim 5, it is characterised in that the ridge structure model bag
Include 4 basic configuration units：E0, E1, E2 and E3, basic configuration unit E0, E1, E2 and E3 are used to characterize the spy of ridge structure model
Levy.
7. Knowledge based engineering spinal cord extraction method according to claim 6, it is characterised in that in the step (2)
The E0 unit centers in some detection zone movement ridge structure models in spinal cord, until E0 unit centers meet ridge structure model attributes spy
Requirement is levied, now the corresponding point of E0 unit centers is in spinal cord a bit.
8. Knowledge based engineering spinal cord extraction method according to claim 3, it is characterised in that in the step (3)
Based on a little and the CT values scope of spinal cord carries out region growth in the limited area of spinal cord Probability Region, obtaining in the spinal cord for detecting
Region be spinal cord Probability Region.
9. Knowledge based engineering spinal cord extraction method according to claim 1, it is characterised in that general by searching for spinal cord
Connect greatest circle to detect spinal cord in rate area border.
10. Knowledge based engineering spinal cord extraction method according to claim 9, it is characterised in that as follows
To detect spinal cord：
In the range of human body spinal cord radius actual size, by the descending search for defining different radii successively of radius actual size
Circle template；
Then the descending mobile search circle template center in spinal cord Probability Region of radius size is pressed, bag in search circle is met first
The point for containing belongs to the circle of spinal cord Probability Region pixel and is the target circle for detecting；If being not detected by target circle, mobile half
The minimum template circle center in footpath, the point included in circle belongs to the most circle of spinal cord Probability Region pixel as target circle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611130965.8A CN106683078A (en) | 2016-12-09 | 2016-12-09 | Knowledge based spinal cord automatic extracting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611130965.8A CN106683078A (en) | 2016-12-09 | 2016-12-09 | Knowledge based spinal cord automatic extracting method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106683078A true CN106683078A (en) | 2017-05-17 |
Family
ID=58868624
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611130965.8A Pending CN106683078A (en) | 2016-12-09 | 2016-12-09 | Knowledge based spinal cord automatic extracting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106683078A (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104346799A (en) * | 2013-08-01 | 2015-02-11 | 上海联影医疗科技有限公司 | Method for extracting spinal marrow in CT (Computed Tomography) image |
-
2016
- 2016-12-09 CN CN201611130965.8A patent/CN106683078A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104346799A (en) * | 2013-08-01 | 2015-02-11 | 上海联影医疗科技有限公司 | Method for extracting spinal marrow in CT (Computed Tomography) image |
Non-Patent Citations (1)
| Title |
|---|
| 王玉等: "CT影像中一种基于知识的脊髓自动提取方法", 《仪器仪表学报》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109074500B (en) | System and method for segmenting medical images of the same patient | |
| CN106778005B (en) | Prostate cancer computer-aided detection system based on multi-parameter MRI | |
| CN107596578B (en) | Alignment mark recognition method, alignment mark position determination method, image forming apparatus, and storage medium | |
| US10269122B2 (en) | Knowledge-based automatic image segmentation | |
| EP1399888A2 (en) | Image segmentation | |
| JP2017532092A (en) | System and method for segmenting medical images based on anatomical landmark-based features | |
| CN103810363B (en) | Blood vessel seed point selecting method and blood vessel extracting method in angiography | |
| CN104036109A (en) | Image based system and method for case retrieving, sketching and treatment planning | |
| US10275895B2 (en) | Mechanism for advanced structure generation and editing | |
| CN109615642A (en) | Jeopardize the automatic delineation method of organ and device in a kind of radiotherapy planning | |
| GB2528249A (en) | Automatic background region selection for lesion delineation in medical images | |
| CN109753997B (en) | Automatic accurate robust segmentation method for liver tumor in CT image | |
| CN102369555B (en) | For the method and system that the auto contrast carrying out profile description strengthens | |
| CN103020969A (en) | Processing method and processing system for dividing liver graphs of CT (computed tomography) image | |
| CN104094315B (en) | Object images indication device, method and apparatus | |
| CN104867104A (en) | Method for obtaining anatomical structural atlas for target mouse based on XCT image non-rigid registration | |
| CN107507189A (en) | Mouse CT image kidney dividing methods based on random forest and statistical model | |
| CN106846346A (en) | Sequence C T image pelvis profile rapid extracting methods based on key frame marker | |
| CN109378068A (en) | A kind of method for automatically evaluating and system of Therapeutic Effects of Nasopharyngeal | |
| EP3075415B1 (en) | Method, computer program and system for dose calculation in radiotherapy | |
| CN115222937A (en) | Method and device for detecting scoliosis | |
| CN106780491A (en) | The initial profile generation method used in GVF methods segmentation CT pelvis images | |
| CN105556567A (en) | Method and system for spine position detection | |
| WO2011163414A2 (en) | Mechanism for advanced structure generation and editing | |
| CN105869169B (en) | A kind of automatic division method of the micro- arrangement image of tumor tissues |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170517 |