CN110279433A - A kind of fetus head circumference automatic and accurate measurement method based on convolutional neural networks - Google Patents
A kind of fetus head circumference automatic and accurate measurement method based on convolutional neural networks Download PDFInfo
- Publication number
- CN110279433A CN110279433A CN201910384672.XA CN201910384672A CN110279433A CN 110279433 A CN110279433 A CN 110279433A CN 201910384672 A CN201910384672 A CN 201910384672A CN 110279433 A CN110279433 A CN 110279433A
- Authority
- CN
- China
- Prior art keywords
- image
- training
- deeplabv2
- head circumference
- neural networks
- Prior art date
Links
- 210000003754 Fetus Anatomy 0.000 title claims abstract description 33
- 230000001537 neural Effects 0.000 title claims abstract description 22
- 238000000691 measurement method Methods 0.000 title claims abstract description 20
- 238000002604 ultrasonography Methods 0.000 claims abstract description 39
- 230000001605 fetal Effects 0.000 claims abstract description 18
- 238000003709 image segmentation Methods 0.000 claims abstract description 10
- 230000004927 fusion Effects 0.000 claims abstract description 4
- 238000000638 solvent extraction Methods 0.000 claims abstract description 4
- 238000000034 methods Methods 0.000 claims description 18
- 210000004556 Brain Anatomy 0.000 claims description 4
- 239000000284 extracts Substances 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims 1
- 238000003745 diagnosis Methods 0.000 abstract description 5
- 230000000694 effects Effects 0.000 abstract description 4
- 210000003625 Skull Anatomy 0.000 abstract description 3
- 230000001154 acute Effects 0.000 abstract description 2
- 230000011218 segmentation Effects 0.000 description 6
- 230000018109 developmental process Effects 0.000 description 4
- 210000001103 Thalamus Anatomy 0.000 description 3
- 230000035935 pregnancy Effects 0.000 description 2
- 230000003252 repetitive Effects 0.000 description 2
- 239000007787 solids Substances 0.000 description 2
- 210000000689 upper leg Anatomy 0.000 description 2
- 210000001015 Abdomen Anatomy 0.000 description 1
- 210000001161 Embryo, Mammalian Anatomy 0.000 description 1
- 281000019761 Intel, Corp. companies 0.000 description 1
- 241000272168 Laridae Species 0.000 description 1
- 208000001254 Occupational Diseases Diseases 0.000 description 1
- 280000545949 Relevant Technologies companies 0.000 description 1
- 230000003187 abdominal Effects 0.000 description 1
- 238000004458 analytical methods Methods 0.000 description 1
- 238000004364 calculation methods Methods 0.000 description 1
- 230000002490 cerebral Effects 0.000 description 1
- 201000000522 chronic kidney diseases Diseases 0.000 description 1
- 230000000875 corresponding Effects 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000010586 diagrams Methods 0.000 description 1
- 201000010099 diseases Diseases 0.000 description 1
- 238000002224 dissection Methods 0.000 description 1
- 238000005516 engineering processes Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000002045 lasting Effects 0.000 description 1
- 239000010410 layers Substances 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006011 modification reactions Methods 0.000 description 1
- 238000004393 prognosis Methods 0.000 description 1
- 230000000306 recurrent Effects 0.000 description 1
- 238000006467 substitution reactions Methods 0.000 description 1
- 210000001519 tissues Anatomy 0.000 description 1
- 230000001131 transforming Effects 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/08—Detecting organic movements or changes, e.g. tumours, cysts, swellings
- A61B8/0866—Detecting organic movements or changes, e.g. tumours, cysts, swellings involving foetal diagnosis; pre-natal or peri-natal diagnosis of the baby
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/52—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/5215—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves involving processing of medical diagnostic data
Abstract
Description
Technical field
The present invention relates to medical ultrasonic diagnostic technical fields, and in particular to a kind of fetus head circumference based on convolutional neural networks Automatic and accurate measurement method.
Background technique
Accurate evaluation development of fetus situation is for ensuring the lasting health of mother and newborn during pregnancy and after pregnancy It is most important.Two-dimensional ultrasound since with strong real-time, at low cost, availability is wide, be not present harmful radiation the advantages that, so by It is widely used in antenatal exaination.Prenatal ultrasonography for diagnosing is the observation most common method of embryo growth and development situation, and prevents defect Fetal birth evaluates the important evidence of development of fetus degree.
The main flow of Prenatal ultrasonography for diagnosing includes: first, the acquisition of fetus key position ultrasound standard section；Second, Biological parameter measurement based on fetal ultrasound standard section；Third judges the life of fetus using the gestational age physique and health of fetus Long development condition.By the process of Prenatal ultrasonography for diagnosing it is found that doctor obtains the ultrasound mark at each position of fetus using two-dimensional ultrasound Quasi- section, and biological parameter measurement is carried out on this basis, such as document Chen Hao, Dou Qi, Ni Dong, et al.Automatic fetal ultrasound standard plane detection using knowledge transferred recurrent neural networks[C]//Nassir Navab,Joachim Hornegger, Medical Image Computing and Computer-Assisted Intervention.Munich:Springer In International Publishing, 2015:507-514., abdominal circumference is measured on upper abdomen section；On sectiones thalamencephali Measure head circumference；It is long that femur is measured on femur section；Measuring head stern length and nuchal translucency thickness etc. on median sagittal section.
Currently, the head circumference measurement on sectiones thalamencephali is by the manually implemented Fig. 2 institute of ultrasonic doctor in clinical practice diagnosis Show, it is insufficient that this results in following five aspects:
1, ultrasonic image, which is checked on the quality, is difficult to effectively control；Get parms measurement depend on ultrasonic doctor clinical experience and Professional standards, if experience level is insufficient, ultrasonic image quality is difficult to ensure；
2, result of ultrasonography does not have uniformity；Since different ultrasonic doctor abilities and experience are different, to fetus Head circumference measurement result is also not quite similar；
3, it is difficult to effectively improve ultrasonic examination working efficiency；Since complete Prenatal ultrasonography for diagnosing needs to obtain fetus head circumference The mode of parameter, manual measurement leads to inefficiency；
4, sonographer is susceptible to suffer from occupational disease；On the one hand sonographer quantity lacks at present, and working frequency is very high, while ultrasound Doctor need to largely carry out such as mobile probe, freeze frame repetitive operation, be susceptible to suffer from repetitive pressure damage；
5, basic hospital ultrasonic imaging diagnosis technology is to be improved；Basic hospital sonographer diagnostic techniques needs further It improves, is diagnosed particular for elderly parturient women, Grade A hospital is needed to diagnose, cause patient's medical treatment interrogation inconvenient, Grade A hospital Working strength increases, and medical resource does not sink to base really.
In recent years, there are many be based on Hough transformation (Hough transform), Haar feature (Haar-Like Features), multi-threshold (Multilevel thresholding), round shortest path (Circular shortest Paths), active contour model, the automatic head circumference measuring system such as (Active contouring) are suggested.But these methods, all It is the approximate measure mode based on ellipse fitting, there has been no the schemes automatically accurately measured based on skull contour line, this is also The problem that " Intelligence Ultrasound " field value must be broken through.
Computer vision technique (Computer Vision, CV) is fast-developing, wherein depth convolutional neural networks (Deep Convolutional Neural Networks, DCNN) it is with the fastest developing speed.DCNN is on the basis of conventional multilayer perceptron It is inspired and is made a variation by cerebral nerve meta structure, it passes through the combined application of convolution kernel and Chi Huahe, from the high dimension of large amount of complex According to the abstract characteristics of middle extraction, spatial correlation therein is excavated.DCNN has been successfully applied to the field of medical imaging, such as in image Registration, positioning, dissection/eucaryotic cell structure detection, tissue segmentation, the assessment of area of computer aided disease diagnosis and prognosis etc..But it is directed to tire Head circumference measurement on youngster's ultrasound thalamus standard section, occurs still without the full-automatic measuring method based on DCNN.
Summary of the invention
The technical problems to be solved by the present invention are: being deposited during the existing clinical practice diagnosis to the measurement of fetus head circumference Efficiency is lower, doctor's subjective factor is affected and is unfavorable for the problems such as accurately detecting, the present invention provides solve the above problems A kind of fetus head circumference automatic and accurate measurement method based on convolutional neural networks, so the present invention propose based on depth convolution mind Through the full-automatic accurate measurement method of network-DeepLabv2 fetus head circumference.
The present invention is achieved through the following technical solutions:
A kind of fetus head circumference automatic and accurate measurement method based on convolutional neural networks, comprising the following steps:
Setp1 inputs primary fetal ultrasound sectiones thalamencephali ultrasonic scan image；
Setp2, multi-angle cut ultrasound image；
Setp3 inputs to trained DeepLabv2, obtains multi-angle image segmentation result；
Setp4 is overlapped fusion to multi-angle image segmentation result；
Setp5 extracts image segmentation boundary, completes measurement after obtaining image boundarg pixel length.
Further, the measurement method is to carry out head circumference automatic measurement system at deep learning frame TensorFlow System design and test.
It further, further include the sensitive information for removing input picture, the sensitive information between Setp1 and Setp2 Including about patients' privacy information.
Further, the Setp2 multi-angle cutting refers to choosing multiple angle directions above image and is cut.I.e. Side chooses any position and is cut on the image, in the forward projection point and image of the cutting position of selection between object Distance >=0.
Further, in the Setp3, DeepLabv2 model training method the following steps are included:
Setp31 builds DeepLabv2 network model；
Setp32 is finely adjusted using pre-training collection, to complete the DeepLabv2 network model of pre-training；
Setp33 measures the difference between predicted value and actual value, to correct DeepLabv2 network model；
Setp34, in training process, original image and the pairs of input training of tag image, each pixel pair in image Loss function cross entropy is equivalent weight；
Setp35 optimizes DeepLabv2 network model parameter using stochastic gradient descent algorithm.
Further, the DeepLabv2 model training method concrete operations are as follows: Setp31 uses ImageNet data Collection carries out pre-training to DeepLabv2 network model, completes initialization, achievees the purpose that transfer learning；Setp32 completes pre- instruction Experienced DeepLabv2 network model continues the training on fetal ultrasound image, is finely adjusted, is cut with adapting to fetal ultrasound thalamus The cranium brain of face image divides task；Setp33, in training process, we use cross entropy cost function as loss function, come Measure the difference between predicted value and actual value；Setp34, in training process, original image and the pairs of input of tag image are instructed Practice, each pixel in image is equivalent weight to loss function cross entropy；Setp35, it is excellent using stochastic gradient descent algorithm Change DeepLabv2 network model parameter, to be optimal, entire training process amounts to iteration 2 × 105Secondary, initial learning rate is 2.5×10-4, every 1 × 104Learning rate decline 5 × 10-4Times.
Further, in the Setp5, concrete operation method are as follows: in such a way that multi-angle is cut, inversely by binaryzation Segmented image splicing and recovery at original image, the input picture cut with multi-angle corresponds in original image.
Further, in the Setp6, the concrete operations of partitioning boundary are extracted are as follows: the segmented image obtained for Setp5 It is first expanded, then subtracts the segmented image that Setp5 is obtained with the image after expansion.
The present invention has the advantage that and the utility model has the advantages that
The present invention is realized to fetal ultrasound sectiones thalamencephali head circumference precise measurement, uses DeepLabV2 convolutional neural networks mould Type effectively divides the region of the cranium of identification fetus sectiones thalamencephali, accurately finds out the edge in ultrasound image in hyperechoic skull, And then the precise measurement of fetus head circumference is realized, it is suitable for Prenatal ultrasonography for diagnosing, to alleviation doctor's operating pressure, improves antenatal surpass Audio clinic working efficiency has positive effect, while can also alleviate the conflict between doctors and patients being becoming increasingly acute and improve having for medical resource Effect utilizes, and has biggish social value and actual use value.
Detailed description of the invention
Attached drawing described herein is used to provide to further understand the embodiment of the present invention, constitutes one of the application Point, do not constitute the restriction to the embodiment of the present invention.In the accompanying drawings:
Fig. 1 is sectiones thalamencephali ideograph；
Fig. 2 is that doctor manually scheme by oval measurement in the prior art；
Fig. 3 is that multi-angle cuts original image；
Fig. 4 is image after multi-angle is cut；
Fig. 5 is head circumference measuring method flow chart of the invention；
Fig. 6 is that fetal ultrasound head circumference marks figure comparison diagram.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below with reference to embodiment and attached drawing, to this Invention is described in further detail, and exemplary embodiment of the invention and its explanation for explaining only the invention, are not made For limitation of the invention.
Embodiment 1
Present embodiments provide a kind of fetus head circumference automatic and accurate measurement method based on convolutional neural networks, operating procedure It is successively as follows:
Setp1 inputs primary fetal ultrasound sectiones thalamencephali ultrasonic scan image；It is described a kind of based on convolutional neural networks Fetus head circumference automatic and accurate measurement method is to carry out the design of head circumference automatic measurement system at deep learning frame TensorFlow With test；
Setp2, removes the sensitive information of input picture, and the sensitive information includes the patient's name of image border, age Equal privacy informations；
Quick ultrasound image is removed in Setp3, multi-angle cutting；It as shown in Figure 3 and Figure 4, is the original image of H to an a length of W, width, It is cut from upper left, upper right, bottom right, the angle of lower-left four with ratio Q (0 < Q < 1, Q=0.9 of the present invention), then cutting is obtained Image, be input in model with the size of 321*321；
Setp4 inputs to trained DeepLabv2, obtains multi-angle image segmentation result, to the target area of image The boundary for obscuring region of domain and background area is split；DeepLabv2 the relevant technologies can be found in document Chen L C, Papandreou G,Kokkinos I,et al.DeepLab:Semantic Image Segmentation with Deep Convolutional Nets,Atrous Convolution,and Fully Connected CRFs[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2018,40(4):834-848.
It is 32 core 2.9GHz Intel Xeon E5-2670CPU, NVIDIA that the system experimentation, which runs hardware environment, The computer of 1080TI GPU and 64GB memory.In order to avoid there is overfitting problem, the DeepLabv2 model that the present invention uses It does not take Gaussian Profile to initialize, but directly uses the mode of transfer learning, it is each by what is learnt on natural image Layer parameter migration initialization, and be finely adjusted in fetal ultrasound image training process, it is specific as follows.
DeepLabv2 model training process: 1. first carry out in advance DeepLabv2 network model using ImageNet data set Training completes initialization, achievees the purpose that transfer learning；2. completing the DeepLabv2 network model of pre-training, continue in fetus Training, is finely adjusted on ultrasound image, divides task to adapt to the cranium brain of fetal ultrasound sectiones thalamencephali image；3. training process In, use cross entropy cost function as loss function, the difference between Lai Hengliang predicted value and actual value；4. training process In, original image is trained with the pairs of input of tag image, and each pixel in image is equivalent power to loss function cross entropy Weight；5. optimizing DeepLabv2 network using stochastic gradient descent algorithm (Stochastic gradient descent, SGD) Model parameter, to be optimal；6. entire training process amounts to iteration 2 × 105Secondary, initial learning rate is 2.5 × 10-4, every 1 ×104Learning rate decline 5 × 10-4Times.
Setp5 is overlapped fusion to multi-angle segmentation result；I.e. in such a way that multi-angle is cut, inversely by two-value At original image, the input picture cut with multi-angle corresponds the segmented image splicing and recovery of change in original image；For overlapping Part, the segmentation figure each point pixel value corresponding with original image made using divide operations are maintained within the scope of 0-255.
Setp6 completes measurement after extracting partitioning boundary；For the segmentation figure A that step 6 obtains, first A is expanded, so Original image A is subtracted with the image B after expansion afterwards；Boundary pixel length is finally obtained, measurement is completed.
Embodiment 2
Based on a kind of fetus head circumference automatic and accurate measurement method based on convolutional neural networks that embodiment 1 provides, carry out Outcome evaluation:
The fetal ultrasound thalamus data set that the present invention uses amounts to 1261, wherein 1145 are training set, 116 are Test set.
1, qualitative evaluation
Fig. 6 is fetal ultrasound head circumference mark figure.It successively arranges from left to right, first to be classified as clinically used ellipse semi-automatic It marks (dotted line), second is classified as manual contours mark (solid line), and third is classified as the automatic marking result (solid line) of algorithm.It can be very Apparent discovery, our algorithm successfully split the identification of fetus cranium brain, and accurately find out in ultrasound image in Qiang Huisheng Skull edge.At the same time, the segmentation result of algorithm is more accurate than clinically used oval mark, with manual mark base This is consistent, illustrates feasibility of this algorithm in this problem.
2, it is quantitatively evaluated
The present invention uses the evaluation index based on region: coincidence factor (Dice), similarity (Jaccard), to assess segmentation As a result；Carry out assessment of the measurement result using mean value (d), standard deviation (s).
Assuming that A is the target area of hand labeled, B is the target area that algorithm is divided automatically.Dice and Jaccard is calculated Formula is as follows:
Wherein, Area (*) indicates areal calculation symbol.Assessment result is as shown in table 1:
Table 1 is quantitatively evaluated
As shown in Table 1, result of the invention with manually mark coincidence factor, similarity more than 0.95, measurement Mean value error is -0.3517cm (measurement result that-expression marks manually is less than measurement result of the invention), and standard deviation is 0.2816, indicate that the present invention has preferable consistency with manually mark；At the same time, clinically used oval mark is surveyed It is generally bigger than normal than measurement result of the invention to measure result, mean value error is+1.5621cm, the mean value error with manual mark measurement Reach+1.9128cm, illustrates the validity of precise measurement head circumference of the present invention.
Above-described specific embodiment has carried out further the purpose of the present invention, technical scheme and beneficial effects It is described in detail, it should be understood that being not intended to limit the present invention the foregoing is merely a specific embodiment of the invention Protection scope, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should all include Within protection scope of the present invention.
Claims (8)
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811108648 | 2018-09-21 | ||
| CN2018111086485 | 2018-09-21 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110279433A true CN110279433A (en) | 2019-09-27 |
| CN110279433B CN110279433B (en) | 2020-03-27 |
Family
ID=68002206
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910384672.XA CN110279433B (en) | 2018-09-21 | 2019-05-09 | Automatic and accurate fetal head circumference measuring method based on convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110279433B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106600595A (en) * | 2016-12-21 | 2017-04-26 | 厦门可睿特信息科技有限公司 | Human body characteristic dimension automatic measuring method based on artificial intelligence algorithm |
| CN106897573A (en) * | 2016-08-01 | 2017-06-27 | 12西格玛控股有限公司 | Use the computer-aided diagnosis system for medical image of depth convolutional neural networks |
| WO2018009405A1 (en) * | 2016-07-08 | 2018-01-11 | Avent, Inc. | System and method for automatic detection, localization, and semantic segmentation of anatomical objects |
| CN108053410A (en) * | 2017-12-11 | 2018-05-18 | 厦门美图之家科技有限公司 | Moving Object Segmentation method and device |
| CN108288271A (en) * | 2018-02-06 | 2018-07-17 | 上海交通大学 | Image detecting system and method based on three-dimensional residual error network |
-
2019
- 2019-05-09 CN CN201910384672.XA patent/CN110279433B/en active IP Right Grant
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018009405A1 (en) * | 2016-07-08 | 2018-01-11 | Avent, Inc. | System and method for automatic detection, localization, and semantic segmentation of anatomical objects |
| CN106897573A (en) * | 2016-08-01 | 2017-06-27 | 12西格玛控股有限公司 | Use the computer-aided diagnosis system for medical image of depth convolutional neural networks |
| CN106600595A (en) * | 2016-12-21 | 2017-04-26 | 厦门可睿特信息科技有限公司 | Human body characteristic dimension automatic measuring method based on artificial intelligence algorithm |
| CN108053410A (en) * | 2017-12-11 | 2018-05-18 | 厦门美图之家科技有限公司 | Moving Object Segmentation method and device |
| CN108288271A (en) * | 2018-02-06 | 2018-07-17 | 上海交通大学 | Image detecting system and method based on three-dimensional residual error network |
Non-Patent Citations (2)
| Title |
|---|
| CHEN L C ET AL.: "DeepLab:Semantic Image Segmentation with Deep Convolutional Nets,Atrous Convolution, and Fully Connected CRFs", 《IEEE TRANSACTIONS ON PATTERN ANALYSIS & MACHINE INTELLIGENCE》 * |
| 吴凌云: "胎儿腹部超声图像自动质量控制与参数测量", 《中国优秀硕士学位论文全文数据库 医药卫生科技辑（月刊）》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110279433B (en) | 2020-03-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3543880A1 (en) | Stroke diagnosis and prognosis prediction method and system | |
| Kalyan et al. | Artificial neural network application in the diagnosis of disease conditions with liver ultrasound images | |
| Dundar et al. | Computerized classification of intraductal breast lesions using histopathological images | |
| Xu et al. | Medical breast ultrasound image segmentation by machine learning | |
| CN103607947B (en) | The detection of anatomic landmark | |
| KR20190028422A (en) | Systems and methods for automatic detection, localization, and semantic segmentation of anatomical objects | |
| CA2575517C (en) | System and method for utilizing shape analysis to assess fetal abnormality | |
| CN105009174B (en) | Method and system for assessing tissue fibrosis | |
| CN105513077B (en) | A kind of system for diabetic retinopathy screening | |
| CN101706843B (en) | Interactive film Interpretation method of mammary gland CR image | |
| CN102800089B (en) | Main carotid artery blood vessel extraction and thickness measuring method based on neck ultrasound images | |
| US20150003706A1 (en) | Probability mapping for visualisation and analysis of biomedical images | |
| Barbosa et al. | Fast and fully automatic 3-d echocardiographic segmentation using B-spline explicit active surfaces: feasibility study and validation in a clinical setting | |
| US20080021502A1 (en) | Systems and methods for automatic symmetry identification and for quantification of asymmetry for analytic, diagnostic and therapeutic purposes | |
| Hiremath et al. | Automatic detection of follicles in ultrasound images of ovaries using edge based method | |
| Alansary et al. | Fast fully automatic segmentation of the human placenta from motion corrupted MRI | |
| CN102871686B (en) | The apparatus and method of physiological parameter are measured based on 3D medical image | |
| Qiu et al. | Automatic segmentation approach to extracting neonatal cerebral ventricles from 3D ultrasound images | |
| EP2365356B1 (en) | Three-dimensional (3D) ultrasound system for scanning object inside human body and method for operating 3D ultrasound system | |
| Casciaro et al. | Fully automatic segmentations of liver and hepatic tumors from 3-D computed tomography abdominal images: comparative evaluation of two automatic methods | |
| CN100562291C (en) | A kind of at CT treatment of picture device, method and system | |
| CN107492097B (en) | Method and device for identifying region of interest of MRI (magnetic resonance imaging) image | |
| Mansourvar et al. | Automated bone age assessment: motivation, taxonomies, and challenges | |
| US8634610B2 (en) | System and method for assessing cancer risk | |
| CN101596111B (en) | Automatic localization of the left ventricle in cardiac cine magnetic resonance imaging |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |