CN110232691A - A kind of dividing method of multi-modal CT images - Google Patents
A kind of dividing method of multi-modal CT images Download PDFInfo
- Publication number
- CN110232691A CN110232691A CN201910312812.2A CN201910312812A CN110232691A CN 110232691 A CN110232691 A CN 110232691A CN 201910312812 A CN201910312812 A CN 201910312812A CN 110232691 A CN110232691 A CN 110232691A
- Authority
- CN
- China
- Prior art keywords
- images
- image
- modal
- benchmark
- dividing 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
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000011218 segmentation Effects 0.000 claims abstract description 18
- 230000003902 lesion Effects 0.000 claims abstract description 9
- 230000009466 transformation Effects 0.000 claims abstract description 7
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 6
- 210000000056 organ Anatomy 0.000 claims abstract description 5
- 230000000007 visual effect Effects 0.000 claims abstract description 3
- 210000004185 liver Anatomy 0.000 claims description 26
- 230000005489 elastic deformation Effects 0.000 claims description 6
- 238000002601 radiography Methods 0.000 claims description 6
- 230000006835 compression Effects 0.000 claims description 4
- 238000007906 compression Methods 0.000 claims description 4
- 230000006837 decompression Effects 0.000 claims description 4
- 238000007689 inspection Methods 0.000 claims description 4
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 2
- 241001269238 Data Species 0.000 claims description 2
- 238000003709 image segmentation Methods 0.000 claims description 2
- 230000002787 reinforcement Effects 0.000 claims description 2
- 239000002872 contrast media Substances 0.000 abstract description 16
- 238000004458 analytical method Methods 0.000 abstract description 3
- 230000009286 beneficial effect Effects 0.000 abstract 1
- 206010028980 Neoplasm Diseases 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000002440 hepatic effect Effects 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000002583 angiography Methods 0.000 description 2
- 238000002591 computed tomography Methods 0.000 description 2
- 238000013170 computed tomography imaging Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000029142 excretion Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 210000004072 lung Anatomy 0.000 description 2
- 230000004060 metabolic process Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 206010067484 Adverse reaction Diseases 0.000 description 1
- 208000037157 Azotemia Diseases 0.000 description 1
- 206010013082 Discomfort Diseases 0.000 description 1
- 206010020751 Hypersensitivity Diseases 0.000 description 1
- 206010047700 Vomiting Diseases 0.000 description 1
- 239000002253 acid Substances 0.000 description 1
- 230000006838 adverse reaction Effects 0.000 description 1
- 208000026935 allergic disease Diseases 0.000 description 1
- 230000007815 allergy Effects 0.000 description 1
- 210000001367 artery Anatomy 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 208000006673 asthma Diseases 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 210000000621 bronchi Anatomy 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 206010012601 diabetes mellitus Diseases 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 208000019622 heart disease Diseases 0.000 description 1
- 238000003706 image smoothing Methods 0.000 description 1
- 238000001727 in vivo Methods 0.000 description 1
- DGIAUNUPXILTJW-VRWDCWMNSA-N iodipamide dimeglumine Chemical compound CNC[C@H](O)[C@@H](O)[C@H](O)[C@H](O)CO.CNC[C@H](O)[C@@H](O)[C@H](O)[C@H](O)CO.OC(=O)C1=C(I)C=C(I)C(NC(=O)CCCCC(=O)NC=2C(=C(C(O)=O)C(I)=CC=2I)I)=C1I DGIAUNUPXILTJW-VRWDCWMNSA-N 0.000 description 1
- 230000003907 kidney function Effects 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 208000019423 liver disease Diseases 0.000 description 1
- 230000003908 liver function Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 229940100477 meglumine iodipamide Drugs 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 210000000496 pancreas Anatomy 0.000 description 1
- 210000003240 portal vein Anatomy 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000035939 shock Effects 0.000 description 1
- 210000002784 stomach Anatomy 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 210000001835 viscera Anatomy 0.000 description 1
- 230000008673 vomiting Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G06T3/153—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—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
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/30056—Liver; Hepatic
Abstract
The invention discloses a kind of dividing methods of multi-modal CT images, comprising: selection benchmark modality images arrange a series of control points wherein；Using similarity regression model, the control point of benchmark modality images is obtained in other modality images best match positions by recurrence with the coordinate transformation model based on depth convolutional neural networks by the principle of similarity identification；The image of other mode is matched by method for registering with benchmark modality images；Multi-modal image after matching is regarded to multiple visual channel data of same mode, organ and lesion segmentation are carried out with this, segmentation is realized by full convolutional network.The beneficial effect is that can be reduced the metering of contrast medium introducing, achieve the purpose that the harm reduced contrast medium to human body, reduction medical treatment cost.It is targetedly identified and is divided for the CT image for introducing a small amount of contrast medium simultaneously, full-automation analysis interpretation tag image mitigates doctor's operating pressure to improve interpretation result accuracy.
Description
Technical field
The present invention relates to a kind of dividing method of liver CT images, point of especially a kind of multi-modal liver CT images
Segmentation method belongs to medical image processing technical field.
Background technique
Liver is the major organs of body metabolism, participates in the metabolism such as synthesis, decomposition, the excretion of many kinds of substance in vivo, when
When obstacle occurs for liver function, entire body all can be impacted.
The image-forming principle of CR scanning (Computed Tomography, CT) is from all angles X-ray
Radiation exposure human body, since Different Organs in human body and the density of tissue are different with thickness, so that X-ray passes through human body
There can be different degrees of decaying after different tissues, the grayscale perspective view of different tissues or organ can be obtained according to attenuation degree
Picture.
When lesion or suspected lesion occur for liver, inspection often is scanned to liver using CT, however CT is examined
It looks into, between liver organization, especially intralesional, external structure display has different degrees of limitation, this due to isodensity
Caused lack of resolution causes difficulty to clinical qualitative, quantitative Diagnosis, therefore, the CT scan in hepatic disease at present
When operation, people are accustomed to introducing the contrast medium such as allusion quotation kind acid or the meglumine iodipamide through hepatic excretion, keep the difference of its ambient density most
It possibly separates, institutional framework, focus characteristic is shown sufficiently with its striking contrast.
With the application of the imaging techniques such as CT, medical image processing and analysis have become current medical technology it is with fastest developing speed,
One of most significant field of achievement.However for if any allusion quotation allergy medical history person；Azotemia patient；There are asthma, kidney function energy barrier
Hinder, heart disease, lung, bronchus patient；Diabetic；Some special populations such as baby and old man introduce liver contrast medium,
Adverse reaction may be caused, less serious case can cause vomiting, is uncomfortable in chest etc., and discomforts, severe one cause shock even dead.And it does not introduce completely
Contrast medium will cause negative effect to the interpretation effect of imaging, it is difficult to the accurate judgement state of an illness again.And draw due to reducing contrast medium
Enter the CT imaging shot after amount, in a kind of liver contrast medium Digital Subtraction proposed by designer of the present invention
After angiographic method processing, the artificial read tablet of image department doctor is needed, manual dividing mark lesions position interprets CT image, low efficiency
Lower and judgment accuracy can also be fluctuated with operating time, the psychology of doctor state, and general Hepatic CT image division method
Not for the Hepatic CT imaging after a small amount of contrast medium of introducing, the judgement of machine knowledge figure may be influenced because of contrast medium, the present invention
Designer proposes the matched present invention side while proposing a kind of liver contrast medium digital subtraction angiography method
Case.
Summary of the invention
In order to overcome the shortcomings of the dividing method in the prior art without multi-modal liver CT images, the purpose of the present invention
It is to provide for a kind of dividing method of multi-modal CT images, which can cooperate matches with the present invention program
A kind of liver contrast medium digital subtraction angiography method carries out artificial intelligence interpretation, while comparing mass data library data, accurately
The position of lesion in image is marked, and doctor is assisted to judge coincident with severity degree of condition.This angiographic method can reduce contrast medium introducing
Metering, thus achieve the purpose that reduce contrast medium to the harm of human body, reduce medical treatment cost.It is a small amount of right for introducing simultaneously
CT image than agent targetedly identified and divided, with " multi-modal 3D autoregistration and based on deep learning
Calculation machine vision " method, full-automation analysis interpretation tag image, improves interpretation result accuracy to reach, mitigates doctor's work
Make the purpose of pressure.
In order to reach the purpose of the present invention, the technical solution adopted by the present invention to solve the technical problems is:
It is including following using being based on the matched spatial alternation algorithm of object High redundancy and full convolutional network image segmentation
Step:
Step 1: being selected as benchmark modality images for any one mode in multi-modality images, other mode are subject to registration
Image.A series of control points, the generally fixed array (Fig. 4 a) of distance interval are positioned in benchmark modality images；
Step 2: spatial alternation is carried out to the image of other mode.As Fig. 2 passes through phase using similarity regression model
Benchmark modality images are obtained by returning with the coordinate transformation model based on depth convolutional neural networks like the principle of degree identification
Best match position (Fig. 4 b) of the control point on image subject to registration；
Step 3: according to control point by elastic deformation, each pixel of benchmark modality images is obtained on image subject to registration
Corresponding position, by sampling obtain with the matched deformed image (Fig. 4 c) subject to registration of benchmark modality images, with reference map
Merge as being stacked on channel dimension；
Step 4: the multi-modal image after merging is regarded to multiple visual channel data of same mode, device is carried out with this
Official and lesion segmentation, segmentation are realized by full convolutional network.
Further, in step 1, select 20 × 20 × 20 control lattice array, i.e., the length and width of reference images,
Coordinate of 20 equally distributed positions as control point in each dimension is respectively selected on high direction.
Further, in step 2, the measuring similarity of local image where control point and control point return and use two
The binary channels convolutional network of input is realized.
Further, the characteristics of image on the periphery at some control point of reference images is extracted in one channel of convolutional network, another
The characteristics of image of matching image same position is extracted in channel, and the same full articulamentum is accessed in two channels after convolutional layer, realizes
Information merges.
Further, in step 3, method for registering realizes that Control point mesh is in any deformation using elastic deformation
Registration under the conditions of (non-affine transforms).
Further, in step 4, the data of the different radiography phases in the homogeneous inspection after registration by stacking side by side
The mode in channel is superimposed, and the form that more image datas become single image data is split.
Further, parted pattern used in step 4 is full convolutional network, is divided into two ranks of compression and decompression
Section, the intensive connection reinforcement between convolution loop network layer and two stage each level that network passes through middle part is to data sky
Between on association modeling.Model does not need to post-process, and the output of network model is final segmentation result.
Further, the full convolutional neural networks model in step 2 and step 4 is all using the training of method end to end.
Present invention has the main advantage that (1) control point+spatial alternation+elastic deformation registration method is suitable for 2D, 3D
Or the data of other dimensions, in addition to CT images can also be used for multi-modal 2D medical imaging；(2) it is first registrated and is divided again
It cuts, so that different information of the same object on different modalities, which is able to be divided, utilizes (the lesion in existing liver CT images
Semantic segmentation is usually that each mode individually carries out), be conducive to little in the pixel intensity value and surrounding tissue difference of lesion
In the case of accurate segmentation, also therefore, the requirement to the concentration of contrast medium is lower；(3) based on same reason, training segmentation
Full convolutional network only need the image for benchmark mode a segmentation mark is provided rather than more parts, reduce mark work
It measures；(4) pseudo-crystalline lattice of matching based on spatial alternation is end-to-end working method, can further be subtracted by analogue data training
Data needed for few training.Other than liver radiography, the present invention also can be applied to other using contrast agent and carry out repeatedly at
The inspection of picture, it is relatively common especially on internal organs, such as liver, stomach, pancreas etc., lung can also on certain tumor examinations
To adopt this method.
Detailed description of the invention
Fig. 1 is implementation whole process figure of the invention.
Fig. 2 is method for registering schematic diagram.
Fig. 3 is the neural network model of liver ontology and tumor focus segmentation.
Fig. 4 a- Fig. 4 c is that Ground control point matching and image carry out elastically-deformable schematic diagram, wherein Fig. 4 a represents reference map
Picture and control point；Fig. 4 b represents the corresponding position of image subject to registration and control point on it；Fig. 4 c represent after elastic deformation to
It is registrated image.
Below by drawings and examples, the invention will be further described.
In conjunction with Fig. 1, the more radiography issues of the liver of input are according to here including same patient the same as the unenhanced phase once checked
(claiming NC afterwards), artery enhance the CT images of phase (claiming AT afterwards), portal vein enhancing phase (claiming PV afterwards) this 3 radiography phases.The size of image
For H × 512 × 512, i.e., length and width are 512, height it is indefinite (whether between different patients or the not same period of same patient it
Between), but it is not less than 64 layers.The thickness in the direction H is indefinite, but should be consistent with the 3 radiography issues evidences once checked.Image
It is pre-processed first, carries out image smoothing filtering technique, press (NC, AT) afterwards, two pairs of sequences of (NC, PV) carry out registration calculating, will
In AT, PV Image registration to NC image.
In conjunction with Fig. 2, registration process first by the image (by taking NC as an example) of a certain mode as benchmark mode image Ub, with
Contrast image U (by taking AT or PV as an example) while input space measuring similarity network.Measurement network is according to being determined in advance Ub's
3D grid G finds respective coordinates θ of each node of G on U.By θ input space transformation model T, transformation parameter T is obtainedθ
(G).Each space pixel of Ub is navigated to by the position on U using elastic deformation method according to the transformation parameter at control point:
Wherein: p represents control point, and x is coordinate position of a certain pixel of benchmark image on x dimension, xpIt is control point p
Coordinate position on the x dimension of benchmark image, △ p are displacement of the p after spatial alternation, and c is specific non-linear function, T
(x) it is coordinate position after this pixel transform on x dimension.
Then the numerical value that this pixel is determined by bilinear interpolation finally obtains the transformed image V of U.Transformation front and back
Effect refers to Fig. 4 a- Fig. 4 c.
The output of space similarity regression model is (corresponding to return using the coordinate of grid G node as three values of reference origin
3D coordinate), model is realized using convolutional neural networks and is returned, and network passes through people by the training of mode end to end, training sample
Work mark converts to obtain to image progress random deformation.
In conjunction with Fig. 3, after being registrated work, by NC, the image of AT, PV are successively arranged side by side on channel dimension.Former image
It is all 1 channel, the data after Juxtaposition and Superimposition are 3 × H × 512 × 512.Data carry out figure by the full convolutional network of 3D after superposition
As semantic segmentation, the categorical measure of segmentation object determines as needed, says by taking background, liver ontology, 3 class of tumour as an example here
It is bright.The full convolutional network of segmentation is divided into compression (left side) and decompression (right side) two parts, is respectively divided into the processing of 4 grades.Compression unit
Dividing each grade includes down-sampled, 2 layers of convolution connection+activation primitive layer, and image data is every to pass through level-one, and size halves, channel
Several times increase, and it is later 48,96,192 that first order E1 output, which is 24 channels,.The decompression each grade in part includes connection, rises and adopt
Sample, level 2 volume product connection+activation primitive layer, image data is every to pass through level-one, and size doubles, and output channel number halves, D4~D1
Respectively 96,48,24,3.Be convolution loop network layer in intermediate M grade, for realizing model on the direction height H each number of plies
According to association.In addition, there is intensive connection in model, between the two-part each level in left and right to increase height before and after data
Density connection, it is therefore intended that improve the precision of segmentation.
Claims (10)
1. a kind of dividing method of multi-modal CT images, it is characterised in that become using based on the matched space of object High redundancy
Scaling method and full convolutional network image segmentation, include the following steps:
Step 1: selection benchmark modality images arrange a series of control points wherein；
Step 2: similarity regression model is used, by the principle of similarity identification, with the seat based on depth convolutional neural networks
Transformation model is marked, by returning the control point for obtaining benchmark modality images in other modality images best match positions；
Step 3: other modality images are matched by method for registering with benchmark modality images；
Step 4: using the multi-modal image after matching as multiple visual channel data of same mode, pass through full convolutional network
Carry out organ and lesion segmentation.
2. the dividing method of the multi-modal liver CT images of one kind according to claim 1, which is characterized in that the step
The specific arrangement method at control point in rapid one are as follows: selection 20 × 20 × 20 control lattice array, i.e., the length of benchmark modality images,
Coordinate of 20 equally distributed positions as control point in three dimensions is respectively selected on wide, high direction.
3. the dividing method of the multi-modal liver CT images of one kind according to claim 1, which is characterized in that the step
The identification of the measuring similarity of local image where control point is realized using the binary channels convolutional network of two inputs in rapid two.
4. the dividing method of the multi-modal liver CT images of one kind according to claim 1, which is characterized in that the step
The recurrence at control point is realized using the binary channels convolutional network of two inputs in rapid two.
5. the dividing method of the multi-modal liver CT images of one kind according to claim 3 or 4, which is characterized in that described
In binary channels convolutional network, the characteristics of image on the periphery at some control point in benchmark modality images, Ling Yitong are extracted in a channel
Extract corresponding characteristics of image in matched modality images in road.
6. the dividing method of the multi-modal liver CT images of one kind according to claim 1, which is characterized in that the step
Method for registering in rapid three realizes matching under the conditions of Control point mesh is in any deformation (non-affine transforms) using elastic deformation
It is quasi-.
7. the dividing method of the multi-modal liver CT images of one kind according to claim 1, which is characterized in that the step
The data of the different radiography phases in homogeneous inspection after being registrated in rapid four are superimposed by way of channel arranged side by side, by more image datas
The form for becoming single image data is split.
8. the dividing method of the multi-modal liver CT images of one kind according to claim 7, which is characterized in that the segmentation
Model is full convolutional network, is divided into two stages of compression and decompression, convolution loop network layer and two ranks of the network by middle part
Intensive connection reinforcement between each level of section models the association of data spatially.
9. the dividing method of the multi-modal liver CT images of one kind according to claim 7 or 8, which is characterized in that described
The output of parted pattern is final segmentation result.
10. the dividing method of the multi-modal liver CT images of one kind according to claim 1, which is characterized in that described
Full convolutional neural networks model in step 2 and step 4 is all made of method training end to end.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910312812.2A CN110232691A (en) | 2019-04-18 | 2019-04-18 | A kind of dividing method of multi-modal CT images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910312812.2A CN110232691A (en) | 2019-04-18 | 2019-04-18 | A kind of dividing method of multi-modal CT images |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110232691A true CN110232691A (en) | 2019-09-13 |
Family
ID=67860771
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910312812.2A Pending CN110232691A (en) | 2019-04-18 | 2019-04-18 | A kind of dividing method of multi-modal CT images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110232691A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110853738A (en) * | 2019-10-12 | 2020-02-28 | 北京量健智能科技有限公司 | Imaging method and device under action of contrast agent |
| CN110874842A (en) * | 2019-10-10 | 2020-03-10 | 浙江大学 | Chest cavity multi-organ segmentation method based on cascade residual full convolution network |
| CN111080658A (en) * | 2019-12-16 | 2020-04-28 | 中南民族大学 | Cervical MRI image segmentation method based on deformable registration and DCNN |
| CN111079863A (en) * | 2019-12-31 | 2020-04-28 | 重庆青信科技有限公司 | System for utilizing energy spectrum curve statistics fingerprint identification focus tissue |
| CN111161258A (en) * | 2019-12-31 | 2020-05-15 | 重庆青信科技有限公司 | Energy spectrum CT image segmentation method based on energy spectrum curve statistical fingerprint |
| CN111598904A (en) * | 2020-05-21 | 2020-08-28 | 腾讯科技（深圳）有限公司 | Image segmentation method, device, equipment and storage medium |
| CN111784706A (en) * | 2020-06-28 | 2020-10-16 | 广州柏视医疗科技有限公司 | Automatic identification method and system for primary tumor image of nasopharyngeal carcinoma |
| CN111798410A (en) * | 2020-06-01 | 2020-10-20 | 深圳市第二人民医院(深圳市转化医学研究院) | Cancer cell pathological grading method, device, equipment and medium based on deep learning model |
| CN112330642A (en) * | 2020-11-09 | 2021-02-05 | 山东师范大学 | Pancreas image segmentation method and system based on double-input full convolution network |
| CN113139964A (en) * | 2020-01-20 | 2021-07-20 | 上海微创医疗器械（集团）有限公司 | Multi-modal image segmentation method and device, electronic equipment and storage medium |
| CN113140035A (en) * | 2021-04-27 | 2021-07-20 | 青岛百洋智能科技股份有限公司 | Full-automatic human cerebral vessel reconstruction method and device based on multi-modal image fusion technology |
| CN113192014A (en) * | 2021-04-16 | 2021-07-30 | 深圳市第二人民医院(深圳市转化医学研究院) | Training method, device, electronic equipment and medium for improving ventricle segmentation model |
| CN114782454A (en) * | 2022-06-23 | 2022-07-22 | 四川省肿瘤医院 | Image recognition system for preoperative navigation of pelvic tumor images |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130267841A1 (en) * | 2009-02-11 | 2013-10-10 | Siemens Medical Solutions Usa, Inc. | Extracting Application Dependent Extra Modal Information from an Anatomical Imaging Modality for use in Reconstruction of Functional Imaging Data |
| CN105103163A (en) * | 2013-03-07 | 2015-11-25 | 火山公司 | Multimodal segmentation in intravascular images |
| CN106204561A (en) * | 2016-07-04 | 2016-12-07 | 西安电子科技大学 | Prostate multi-modality images non-rigid registration method based on mixed model |
| CN108257134A (en) * | 2017-12-21 | 2018-07-06 | 深圳大学 | Nasopharyngeal Carcinoma Lesions automatic division method and system based on deep learning |
| WO2018129650A1 (en) * | 2017-01-10 | 2018-07-19 | 中国科学院深圳先进技术研究院 | Analysis method for multi-mode radiomics, apparatus and terminal |
-
2019
- 2019-04-18 CN CN201910312812.2A patent/CN110232691A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130267841A1 (en) * | 2009-02-11 | 2013-10-10 | Siemens Medical Solutions Usa, Inc. | Extracting Application Dependent Extra Modal Information from an Anatomical Imaging Modality for use in Reconstruction of Functional Imaging Data |
| CN105103163A (en) * | 2013-03-07 | 2015-11-25 | 火山公司 | Multimodal segmentation in intravascular images |
| CN106204561A (en) * | 2016-07-04 | 2016-12-07 | 西安电子科技大学 | Prostate multi-modality images non-rigid registration method based on mixed model |
| WO2018129650A1 (en) * | 2017-01-10 | 2018-07-19 | 中国科学院深圳先进技术研究院 | Analysis method for multi-mode radiomics, apparatus and terminal |
| CN108257134A (en) * | 2017-12-21 | 2018-07-06 | 深圳大学 | Nasopharyngeal Carcinoma Lesions automatic division method and system based on deep learning |
Non-Patent Citations (2)
| Title |
|---|
| XIAOHUAN CAO ET AL.: "Deformable Image Registration based on Similarity-Steered CNN Regression", 《MED IMAGE COMPUT COMPUT ASSIST INTERV》 * |
| 田亮: "基于数学减影技术的肾脏分割", 《中国优秀硕博士学位论文全文数据库（硕士） 医药卫生科技辑》 * |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110874842A (en) * | 2019-10-10 | 2020-03-10 | 浙江大学 | Chest cavity multi-organ segmentation method based on cascade residual full convolution network |
| CN110874842B (en) * | 2019-10-10 | 2022-04-29 | 浙江大学 | Chest cavity multi-organ segmentation method based on cascade residual full convolution network |
| CN110853738A (en) * | 2019-10-12 | 2020-02-28 | 北京量健智能科技有限公司 | Imaging method and device under action of contrast agent |
| CN110853738B (en) * | 2019-10-12 | 2023-08-18 | 拜耳股份有限公司 | Imaging method and device under action of contrast agent |
| CN111080658A (en) * | 2019-12-16 | 2020-04-28 | 中南民族大学 | Cervical MRI image segmentation method based on deformable registration and DCNN |
| CN111079863A (en) * | 2019-12-31 | 2020-04-28 | 重庆青信科技有限公司 | System for utilizing energy spectrum curve statistics fingerprint identification focus tissue |
| CN111161258A (en) * | 2019-12-31 | 2020-05-15 | 重庆青信科技有限公司 | Energy spectrum CT image segmentation method based on energy spectrum curve statistical fingerprint |
| CN111079863B (en) * | 2019-12-31 | 2023-05-19 | 重庆青信科技有限公司 | System for identifying focus tissue by utilizing spectral curve statistical fingerprint |
| CN111161258B (en) * | 2019-12-31 | 2023-05-09 | 重庆青信科技有限公司 | Energy spectrum CT image segmentation method based on energy spectrum curve statistical fingerprint |
| CN113139964A (en) * | 2020-01-20 | 2021-07-20 | 上海微创医疗器械（集团）有限公司 | Multi-modal image segmentation method and device, electronic equipment and storage medium |
| CN111598904A (en) * | 2020-05-21 | 2020-08-28 | 腾讯科技（深圳）有限公司 | Image segmentation method, device, equipment and storage medium |
| CN111798410A (en) * | 2020-06-01 | 2020-10-20 | 深圳市第二人民医院(深圳市转化医学研究院) | Cancer cell pathological grading method, device, equipment and medium based on deep learning model |
| CN111784706B (en) * | 2020-06-28 | 2021-06-04 | 广州柏视医疗科技有限公司 | Automatic identification method and system for primary tumor image of nasopharyngeal carcinoma |
| CN111784706A (en) * | 2020-06-28 | 2020-10-16 | 广州柏视医疗科技有限公司 | Automatic identification method and system for primary tumor image of nasopharyngeal carcinoma |
| CN112330642A (en) * | 2020-11-09 | 2021-02-05 | 山东师范大学 | Pancreas image segmentation method and system based on double-input full convolution network |
| CN113192014A (en) * | 2021-04-16 | 2021-07-30 | 深圳市第二人民医院(深圳市转化医学研究院) | Training method, device, electronic equipment and medium for improving ventricle segmentation model |
| CN113192014B (en) * | 2021-04-16 | 2024-01-30 | 深圳市第二人民医院(深圳市转化医学研究院) | Training method and device for improving ventricle segmentation model, electronic equipment and medium |
| CN113140035A (en) * | 2021-04-27 | 2021-07-20 | 青岛百洋智能科技股份有限公司 | Full-automatic human cerebral vessel reconstruction method and device based on multi-modal image fusion technology |
| CN114782454A (en) * | 2022-06-23 | 2022-07-22 | 四川省肿瘤医院 | Image recognition system for preoperative navigation of pelvic tumor images |
| CN114782454B (en) * | 2022-06-23 | 2022-09-09 | 四川省肿瘤医院 | Image recognition system for preoperative navigation of pelvic tumor images |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110232691A (en) | A kind of dividing method of multi-modal CT images | |
| CN110223352B (en) | Medical image scanning automatic positioning method based on deep learning | |
| Linguraru et al. | Automated segmentation and quantification of liver and spleen from CT images using normalized probabilistic atlases and enhancement estimation | |
| CN109035284B (en) | Heart CT image segmentation method, device, equipment and medium based on deep learning | |
| CN112529894B (en) | Thyroid nodule diagnosis method based on deep learning network | |
| CN109598722B (en) | Image analysis method based on recurrent neural network | |
| KR102060895B1 (en) | Method for producing medical image and device for producing medical image | |
| CN107072531A (en) | Method and system for the dynamic (dynamical) analysis of myocardial wall | |
| CN103295234B (en) | Based on the medical image segmentation system and method for deformation surface model | |
| WO2022105623A1 (en) | Intracranial vascular focus recognition method based on transfer learning | |
| US11830193B2 (en) | Recognition method of intracranial vascular lesions based on transfer learning | |
| Dai et al. | Deep learning‐based motion tracking using ultrasound images | |
| CN114943714A (en) | Medical image processing system, medical image processing apparatus, electronic device, and storage medium | |
| CN112562058B (en) | Method for quickly establishing intracranial vascular simulation three-dimensional model based on transfer learning | |
| Li et al. | Pulmonary CT image registration and warping for tracking tissue deformation during the respiratory cycle through 3D consistent image registration | |
| US20240078669A1 (en) | System and methods for inferring thickness of object classes of interest in two-dimensional medical images using deep neural networks | |
| Xiao et al. | PET and CT image fusion of lung cancer with siamese pyramid fusion network | |
| US20220164967A1 (en) | Method of establishing an enhanced three-dimensional model of intracranial angiography | |
| Galdames et al. | Registration of renal SPECT and 2.5 D US images | |
| US8675941B2 (en) | Method for visualization of airway anatomical changes | |
| CN112669256B (en) | Medical image segmentation and display method based on transfer learning | |
| Li et al. | 3D intersubject warping and registration of pulmonary CT images for a human lung model | |
| CN114511602A (en) | Medical image registration method based on graph convolution Transformer | |
| Mao et al. | Direct 3d ultrasound fusion for transesophageal echocardiography | |
| Hamarneh | Towards intelligent deformable models for medical image analysis |
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: 20190913 |