CN110400316A - A kind of orthopaedics image measuring method and device based on deep learning - Google Patents
A kind of orthopaedics image measuring method and device based on deep learning Download PDFInfo
- Publication number
- CN110400316A CN110400316A CN201910711742.8A CN201910711742A CN110400316A CN 110400316 A CN110400316 A CN 110400316A CN 201910711742 A CN201910711742 A CN 201910711742A CN 110400316 A CN110400316 A CN 110400316A
- Authority
- CN
- China
- Prior art keywords
- key point
- image
- orthopaedics
- bone
- thermal map
- 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 27
- 238000013135 deep learning Methods 0.000 title claims abstract description 14
- 210000000988 bone and bone Anatomy 0.000 claims abstract description 48
- 238000001514 detection method Methods 0.000 claims abstract description 27
- 238000012549 training Methods 0.000 claims abstract description 17
- 238000005259 measurement Methods 0.000 claims abstract description 12
- 230000004044 response Effects 0.000 claims description 25
- 238000013527 convolutional neural network Methods 0.000 claims description 21
- 238000010606 normalization Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000000399 orthopedic effect Effects 0.000 claims description 2
- 230000001537 neural effect Effects 0.000 claims 1
- 238000003745 diagnosis Methods 0.000 abstract description 6
- 238000013528 artificial neural network Methods 0.000 abstract 1
- 238000000691 measurement method Methods 0.000 abstract 1
- 238000002591 computed tomography Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 238000002595 magnetic resonance imaging Methods 0.000 description 5
- 201000010099 disease Diseases 0.000 description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 208000014674 injury Diseases 0.000 description 3
- 210000000629 knee joint Anatomy 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 210000003484 anatomy Anatomy 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002683 hand surgery Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 230000008733 trauma Effects 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 239000002671 adjuvant Substances 0.000 description 1
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000007598 dipping method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 201000011066 hemangioma Diseases 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- 238000003475 lamination Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000001356 surgical procedure Methods 0.000 description 1
- 238000002604 ultrasonography Methods 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
-
- 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
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4007—Scaling of whole images or parts thereof, e.g. expanding or contracting based on interpolation, e.g. bilinear interpolation
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
- G06V10/464—Salient features, e.g. scale invariant feature transforms [SIFT] using a plurality of salient features, e.g. bag-of-words [BoW] representations
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30008—Bone
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Public Health (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Primary Health Care (AREA)
- Epidemiology (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Quality & Reliability (AREA)
- Pathology (AREA)
- Multimedia (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of orthopaedics image measuring method and device based on deep learning, compared to traditional manual measurement method, hence it is evident that improve the efficiency of orthopaedics radiographic measurement and do not influence precision, realize full automatic orthopaedics radiographic measurement.Wherein the key step of summary of the invention includes: the orthopaedics image for obtaining examinate；Orthopaedics image is pre-processed；Pretreated orthopaedics image is inputted into default bone critical point detection model and obtains bone key point coordinate；Index of correlation is measured based on key point.Due to using volume neural network, the training and detection of bone key point end to end are realized, manual intervention is not necessarily to and detection measuring speed is fast, improve the efficiency of diagnosis.
Description
Technical field
The present invention relates to medical image processing field, in particular to a kind of orthopaedics image measuring method based on deep learning
And device.
Background technique
Iconography can greatly improve medical diagnosis on disease rate, and especially in terms of orthopaedics and most important, iconography is in joint
The fields such as orthopaedics, orthopaedic trauma, spinal surgery, hand surgery related disease all have extremely important application.In Bones and joints section
It is now mostly total with x-ray photo, CT scan (Computed Tomography, CT) image, magnetic in clinic diagnosis
Vibration imaging (Magnetic Resonance Imaging, MRI) is used as diagnosis basis, and adjuvant clinical inspection is made a definite diagnosis.It is creating
In terms of injury of the bone section, mostly diagnosed according to x-ray, CT three-dimensional reconstruction.Spinal trauma then mostly shows spinal anatomy, CT with MRI
Scanning display bone canalis spinalis anatomical structure, x-ray checks long bone growing state, and hospital can check the form combined inspection with a variety of
Survey disease.And hand surgery can usually offer a clear explanation by x-ray, or diagnose (hemangioma of such as hand) by B ultrasound.
In view of the complexity that extremely diagnoses of diversification of orthopaedic disease, by every video diagnostic technology also show
Diversity and complexity, this proposes huge challenge for medical staff.And currently, be directed to orthopaedics image measure it is normal
Square method is concentrated mainly on hand dipping method, threshold method, region growth method, classification based on feature etc., the above each method or more
Or certain priori knowledge is required less, it can not accomplish that not having to any priori knowledge is automatically and accurately detected, local ginseng
It is several, manual setting is needed, versatility, detection efficiency and accuracy are relatively low.And with the rise of deep learning, various depths
The algorithm of degree study is applied to field of medical imaging.
Summary of the invention
In view of this, the present invention provides a kind of orthopaedics image measuring method and device based on deep learning, on solving
State problem.
For up to above-mentioned and other purposes, the embodiment of the present invention provides a kind of orthopaedics radiographic measurement side based on deep learning
Method includes the following steps:
S1 the orthopaedics image of subject) is obtained；
S2) orthopaedics image is pre-processed；
S3 pretreated orthopaedics image) is inputted into default bone critical point detection model and obtains bone key point coordinate, wherein presetting
Bone critical point detection model is made of convolutional neural networks model and passes through training and obtains；
S4 it) is based on key point coordinate, predetermined orthopaedics parameter value is measured, one group of data value is obtained.
Optionally, in step S1, orthopaedics image includes x-ray plain film, CT scan (CT) image, magnetic resonance
(MRI) etc. is imaged.
Optionally, in step S2, pretreatment includes:
Select image to be measured；
And image scaled carried out to image, the operation such as image pixel value normalization or standardization, wherein scaling can be used it is double
Linear interpolation, arest neighbors interpolation etc., normalization includes that pixel value is normalized between 0 to 1, and standardization includes subtracting pixel value
Go mean value and except variance.
Optionally, in step S3, default bone critical point detection model is using convolutional neural networks to the orthopaedics marked
Image data training obtains, comprising:
The input of convolutional neural networks is pretreated image to be measured, exports and responds thermal map for multichannel bone key point；
Wherein, the prediction of the corresponding key point in each channel of multichannel bone key point response thermal map, the key in each channel
Point response thermal map is Gaussian Profile, wherein the maximum value of each channel response thermal map and the position of maximum value are confidence level and pass
Key point position coordinates, then corresponding response thermal map is sky to some crucial spill tag if it exists；
The parameter of convolutional neural networks is obtained by backpropagation training.
Optionally, in step S3, pretreated orthopaedics image is inputted into default bone critical point detection model and obtains bone pass
Key point coordinate, comprising:
Pretreated orthopaedics image is inputted into default convolutional neural networks model, obtains multichannel bone key point response heat
Figure；
The key point response thermal map in each channel is parsed, the coordinate of the corresponding key point of current channel is obtained, wherein often
The maximum value of a channel response thermal map and the position of maximum value are confidence level and key point position coordinates；
If confidence level is less than preset threshold, key point is not present.
Optionally, in step S4, the calculation of the measured value are as follows:
Measure the distance between key point；
The angle between straight line is measured, wherein straight line is connected to obtain by two key points；
Distance of the measurement point to straight line；
Relevant parameter is measured by way of constructing auxiliary line；
Pass through distance or angle survey calculation relevant parameter.
Correspondingly, the embodiment of the invention also provides a kind of orthopaedics image measuring device based on deep learning, comprising:
Acquiring unit, for obtaining the x-ray plain film by proofer, CT or MRI orthopaedics image；
Pretreatment unit for selecting image to be measured, and carries out image scaled to it, and image pixel value normalization etc. is pre-
Processing operation；
Detection unit, for pretreated orthopaedics image input bone critical point detection model to be obtained bone key point coordinate,
Middle bone critical point detection model is made of convolutional neural networks；
Measuring unit obtains one group of data value for measuring scheduled orthopedic parameter value.
Optionally, bone critical point detection model is to the orthopaedics image data marked trained using convolutional neural networks
It arrives, comprising:
The input of convolutional neural networks is image to be measured, exports and responds thermal map for multichannel bone key point；
Multichannel bone key point responds the corresponding key point in each channel of thermal map, and the key point response thermal map in each channel is
Dimensional gaussian distribution, wherein the maximum value of each channel response thermal map and the position of maximum value are confidence level and key point position
Coordinate；
The parameter of convolutional neural networks is obtained by backpropagation training.
Optionally, detection unit is specifically used for:
Pretreated orthopaedics image is inputted into default convolutional neural networks model, obtains multichannel bone key point response heat
Figure；
The key point response thermal map in each channel is parsed, the coordinate of the corresponding key point of current channel is obtained, wherein often
The maximum value of a channel response thermal map and the position of maximum value are confidence level and key point position coordinates；
If confidence level is less than preset threshold, key point is not present.
Optionally, measuring unit is specifically used for:
Measure the distance between key point；
The angle between straight line is measured, wherein straight line is connected to obtain by two key points；
Distance of the measurement point to straight line；
Relevant parameter is measured by way of constructing auxiliary line；
Pass through distance or angle survey calculation relevant parameter.
Correspondingly, the embodiment of the invention also provides a kind of calculating equipment, comprising:
Memory, for storing program instruction；
Processor, for executing program instructions, according to the method for the program execution orthopaedics radiographic measurement of acquisition；
A kind of computer-readable non-volatile memory medium, including computer-readable instruction, when computer is read and executes calculating
When machine readable instruction, so that computer executes above-mentioned orthopaedics image measuring method method.
Compared with the prior art, the invention has the following advantages:
The present invention, come the relevant parameter of automatic measurement orthopaedics image, substantially increases the diagnosis effect of doctor using depth learning technology
Rate, and doctor is assisted to provide more accurate judgement；
The present invention is compared to traditional images processing technique with better robustness, accuracy, and can increase more
Training data is marked to keep model more accurate.
Detailed description of the invention
Fig. 1 is a kind of flow chart of orthopaedics image measuring method based on deep learning of the invention.
Fig. 2 is that the key point of prediction of the present invention responds thermal map.
Fig. 3 is that orthopaedics image of the present invention and key point visualize.
Fig. 4 is a kind of system architecture diagram of orthopaedics image measuring device based on deep learning of the invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention make into
It is described in detail to one step, it is clear that described embodiments are only a part of the embodiments of the present invention, rather than whole implementation
Example.Based on the embodiments of the present invention, obtained by those of ordinary skill in the art without making creative efforts
All other embodiment, shall fall within the protection scope of the present invention.
Fig. 1 is a kind of flow chart of orthopaedics image measuring method and device based on deep learning.Its key step includes:
Obtain the orthopaedics image of examinate；Orthopaedics image is pre-processed；Pretreated orthopaedics image is inputted default bone to close
Key point detection model obtains bone key point coordinate；Index of correlation is measured based on key point.It is thin for convenience of the items understood in invention
Section, by taking the measurement of knee joint x-ray plain film orthopaedics image as an example, is described in detail.
(1) the orthopaedics image of examinate is obtained
Knee joint x-ray plain film orthopaedics image is read from database or disk, as shown in Figure 3.
(2) orthopaedics image is pre-processed
Obtained knee joint x-ray plain film is pre-processed, a height of 384 pixel is specially scaled the images to and width is 288 pictures
The image of plain size, and normalized to pixel value between 0 to 1 using maximum-minimum method for normalizing.
(3) pretreated orthopaedics image is input to default bone critical point detection model and obtains bone key point coordinate
Pretreated orthopaedics image is input to preset critical point detection model, wherein preset critical point detection model is
It obtains by training and is made of convolutional neural networks.Training process is as described below, closes firstly, obtaining 100 to 1000 knees
X-ray plain film orthopaedics image is saved, then the key point to be measured on bone is labeled by veteran doctor, then by another
An outer doctor is confirmed, labeled data concentration is added to if mark is errorless, then by data set according to the ratio of 8:2
Example is divided into training set and verifying collection.
Select class U-Net network as the convolutional neural networks of critical point detection, network includes down-sampling layer (maximum pond
Layer) and up-sampling layer (warp lamination), and it is intermediate (Concatenate) is connected by stacking, each down-sampling layer and above adopt
2 convolution blocks of sample layer heel, each convolution block include 2 dimensions convolution (2DConv), crowd normalization (Batch
Normalization), nonlinear activation (ReLU).
The output of network is multichannel key point response thermal map (as shown in Fig. 2, Fig. 2 is multichannel Overlapping display effect),
Each channel represents a coordinate, and key point coordinate is converted into key point response thermal map as the supervision of network when training and is believed
Breath.The pretreatment in step (2) is carried out to input picture when training, and data are enhanced, such as carries out left and right overturning, rotation
Turn, the operation such as image scaled.Select L2 loss as loss function when training, Adam is as optimizer.Final choice exists
Verifying collects the model of upper loss reduction as final mask.
Image to be tested is pre-processed by step (2), is input to the multi-pass that the final mask of above-mentioned selection is predicted
Road responds thermal map, then is parsed to obtain key point coordinate to multichannel response thermal map, concrete mode is to each channel
It responds thermal map and carries out gaussian filtering, then obtain the maximum value of the response thermal map in each channel and the coordinate of maximum value, this seat
Mark is the coordinate of the corresponding key point of current channel, and key point is not present if maximum value is less than some threshold value.Finally obtain
The key point coordinate detected.
(4) index of correlation is measured based on key point
The measurement index that many orthopaedics can be calculated by key point, including: bone length, bone spacing, bone angle,
The ratio etc. of correlation values.
Specifically there is the pixel Euclidean distance by two key points and be multiplied by the spacing information of image, between calculating bone
Away from bone length；The straight line that two key points are constituted, the angle of straight line can be determined by two straight lines；Pass through related angle
Degree, distance can calculate the index of correlation in some orthopaedics measurements.
Fig. 4 is a kind of system architecture diagram of the orthopaedics measuring device based on deep learning of the present invention, and Fig. 4 is only one and shows
Example, should not function to the embodiment of the present application and use scope bring any restrictions.
The system architecture as shown in Figure 4 can be divided into processor, memory and communication interface.
Wherein communication interface is communicated for terminal device, receives and dispatches the information of terminal device transmission, realizes communication.
Processor is the control centre of system architecture, using the various pieces of various interfaces and line connection system framework,
By running or executing the software program or module that are stored in memory, and the data that calling is stored in memory, it is
The various functions and processing data of system framework.Optionally, processor may include one or more processing units.
Memory can be used for storing software program and module, and processor is stored in the software program of memory by operation
And module, thereby executing various function application and data processing.Memory can mainly include storing program area and storage number
According to area, wherein storing program area can application program needed for storage program area, at least one function etc.；Storage data area can deposit
Store up the data etc. created according to business processing.In addition, memory may include rapid random access memory, can also include
Nonvolatile memory, for example, at least a disk memory, flush memory device or other volatile solid-state parts.
Claims (12)
1. a kind of orthopaedics image measuring method and device based on deep learning, which is characterized in that key step includes:
S1 the orthopaedics image of subject) is obtained；
S2) the orthopaedics image is pre-processed；
S3 the pretreated orthopaedics image) is inputted into default bone critical point detection model and obtains bone key point coordinate, wherein
Default bone critical point detection model is made of convolutional neural networks model and passes through training and obtains；
S4 it) is based on the key point coordinate, predetermined orthopaedics parameter value is measured, one group of data value is obtained.
2. the method as described in claim 1, which is characterized in that in the step S1, the orthopaedics image include x-ray plain film,
CT, MRI etc..
3. the method as described in claim 1, which is characterized in that in the step S2, the pretreatment includes:
Select image to be measured；
And image scaled, the operation such as image pixel value normalization or standardization are carried out to described image.
4. the method as described in claim 1, which is characterized in that in the step S3, the default bone critical point detection model
It is to be obtained using convolutional neural networks to the orthopaedics image data training marked, comprising:
The input of the convolutional neural networks is image to be measured, exports and responds thermal map for multichannel bone key point；
The corresponding key point in each channel of the multichannel bone key point response thermal map, the key point in each channel respond heat
Figure is Gaussian Profile, wherein the maximum value of each channel response thermal map and the position of maximum value are confidence level and key point position
Coordinate；
The parameter of the convolutional neural networks is obtained by backpropagation training.
5. the method as described in claim 1, which is characterized in that in the step S3, by the pretreated orthopaedics shadow
Bone key point coordinate is obtained as inputting default bone critical point detection model, comprising:
The pretreated orthopaedics image is inputted into default convolutional neural networks model, obtains the response of multichannel key point
Thermal map；
The key point response thermal map in each channel is parsed, the coordinate of the corresponding key point of current channel is obtained, wherein often
The maximum value of a channel response thermal map and the position of maximum value are confidence level and key point position coordinates；
If the confidence level is less than preset threshold, key point is not present.
6. the method as described in claim 1, which is characterized in that in the step S4, the calculation of the measured value are as follows:
Measure the distance between key point；
The angle between straight line is measured, wherein straight line is connected to obtain by two key points；
Distance of the measurement point to straight line；
Relevant parameter is measured by way of constructing auxiliary line；
Pass through distance or angle survey calculation relevant parameter.
7. a kind of orthopaedics image measuring device based on deep learning, which is characterized in that described device includes:
Acquiring unit, for obtaining the x-ray plain film by proofer, CT or MRI orthopaedics image；
Pretreatment unit for selecting image to be measured, and carries out image scaled to it, image pixel value normalization or mark
The operation such as standardization；
Detection unit is sat for the pretreated orthopaedics image input bone critical point detection model to be obtained bone key point
Mark, wherein bone critical point detection model is made of convolutional neural networks；
Measuring unit obtains one group of data value for measuring scheduled orthopedic parameter value.
8. device according to claim 7, which is characterized in that the bone critical point detection model is using convolutional Neural net
Network obtains the orthopaedics image data training marked, comprising:
The input of the convolutional neural networks is image to be measured, exports and responds thermal map for multichannel bone key point；
The corresponding key point in each channel of the multichannel bone key point response thermal map, the key point in each channel respond heat
Figure is Gaussian Profile, wherein the maximum value of each channel response thermal map and the position of maximum value are confidence level and key point position
Coordinate；
The parameter of the convolutional neural networks is obtained by backpropagation training.
9. device according to claim 7, which is characterized in that the detection unit is specifically used for:
The pretreated orthopaedics image is inputted into default convolutional neural networks model, multichannel bone key point is obtained and rings
Answer thermal map；
The key point response thermal map in each channel is parsed, the coordinate of the corresponding key point of current channel is obtained, wherein often
The maximum value of a channel response thermal map and the position of maximum value are confidence level and key point position coordinates；
If the confidence level is less than preset threshold, key point is not present.
10. to go 7 described devices according to right, which is characterized in that the measuring unit is specifically used for:
Measure the distance between key point；
The angle between straight line is measured, wherein straight line is connected to obtain by two key points；
Distance of the measurement point to straight line；
Relevant parameter is measured by way of constructing auxiliary line；
Pass through distance or angle survey calculation relevant parameter.
11. a kind of calculating equipment characterized by comprising
Memory, for storing program instruction；
Processor requires 1 to 6 according to the program execution benefit of acquisition for calling the program instruction stored in the memory
Described in any item methods.
12. a kind of computer-readable non-volatile memory medium, which is characterized in that including computer-readable instruction, work as computer
When reading and executing the computer-readable instruction, so that computer executes such as method as claimed in any one of claims 1 to 6.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910317167 | 2019-04-19 | ||
| CN2019103171673 | 2019-04-19 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110400316A true CN110400316A (en) | 2019-11-01 |
Family
ID=68327157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910711742.8A Pending CN110400316A (en) | 2019-04-19 | 2019-08-02 | A kind of orthopaedics image measuring method and device based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110400316A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111340800A (en) * | 2020-03-18 | 2020-06-26 | 联影智能医疗科技(北京)有限公司 | Image detection method, computer device, and storage medium |
| CN111383222A (en) * | 2020-03-18 | 2020-07-07 | 桂林理工大学 | Intervertebral disc MRI image intelligent diagnosis system based on deep learning |
| CN111915554A (en) * | 2020-06-19 | 2020-11-10 | 杭州深睿博联科技有限公司 | Fracture detection and positioning integrated method and device based on X-ray image |
| CN112733767A (en) * | 2021-01-15 | 2021-04-30 | 西安电子科技大学 | Human body key point detection method and device, storage medium and terminal equipment |
| CN112950464A (en) * | 2021-01-25 | 2021-06-11 | 西安电子科技大学 | Binary super-resolution reconstruction method without regularization layer |
| CN113674261A (en) * | 2021-08-26 | 2021-11-19 | 上海脊影慧智能科技有限公司 | Bone detection method, system, electronic device and storage medium |
| CN113706600A (en) * | 2021-08-19 | 2021-11-26 | 北京欧应信息技术有限公司 | Method, apparatus and medium for measuring critical dimensions of a body part |
| CN114638878A (en) * | 2022-03-18 | 2022-06-17 | 北京安德医智科技有限公司 | Two-dimensional echocardiogram pipe diameter detection method and device based on deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107492097A (en) * | 2017-08-07 | 2017-12-19 | 北京深睿博联科技有限责任公司 | A kind of method and device for identifying MRI image area-of-interest |
| CN108227912A (en) * | 2017-11-30 | 2018-06-29 | 北京市商汤科技开发有限公司 | Apparatus control method and device, electronic equipment, computer storage media |
| CN109544536A (en) * | 2018-11-26 | 2019-03-29 | 中国科学技术大学 | The fast automatic analysis method of hip joint x-ray image |
| CN109583341A (en) * | 2018-11-19 | 2019-04-05 | 清华大学深圳研究生院 | To more people's bone bone critical point detection method and devices of the image comprising portrait |
-
2019
- 2019-08-02 CN CN201910711742.8A patent/CN110400316A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107492097A (en) * | 2017-08-07 | 2017-12-19 | 北京深睿博联科技有限责任公司 | A kind of method and device for identifying MRI image area-of-interest |
| CN108227912A (en) * | 2017-11-30 | 2018-06-29 | 北京市商汤科技开发有限公司 | Apparatus control method and device, electronic equipment, computer storage media |
| CN109583341A (en) * | 2018-11-19 | 2019-04-05 | 清华大学深圳研究生院 | To more people's bone bone critical point detection method and devices of the image comprising portrait |
| CN109544536A (en) * | 2018-11-26 | 2019-03-29 | 中国科学技术大学 | The fast automatic analysis method of hip joint x-ray image |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111340800A (en) * | 2020-03-18 | 2020-06-26 | 联影智能医疗科技(北京)有限公司 | Image detection method, computer device, and storage medium |
| CN111383222A (en) * | 2020-03-18 | 2020-07-07 | 桂林理工大学 | Intervertebral disc MRI image intelligent diagnosis system based on deep learning |
| CN111340800B (en) * | 2020-03-18 | 2024-02-27 | 联影智能医疗科技(北京)有限公司 | Image detection method, computer device, and storage medium |
| CN111915554A (en) * | 2020-06-19 | 2020-11-10 | 杭州深睿博联科技有限公司 | Fracture detection and positioning integrated method and device based on X-ray image |
| CN111915554B (en) * | 2020-06-19 | 2024-05-14 | 杭州深睿博联科技有限公司 | Fracture detection and positioning integrated method and device based on X-ray image |
| CN112733767A (en) * | 2021-01-15 | 2021-04-30 | 西安电子科技大学 | Human body key point detection method and device, storage medium and terminal equipment |
| CN112733767B (en) * | 2021-01-15 | 2022-05-31 | 西安电子科技大学 | Human body key point detection method and device, storage medium and terminal equipment |
| CN112950464A (en) * | 2021-01-25 | 2021-06-11 | 西安电子科技大学 | Binary super-resolution reconstruction method without regularization layer |
| CN112950464B (en) * | 2021-01-25 | 2023-09-01 | 西安电子科技大学 | Binary super-resolution reconstruction method without regularization layer |
| CN113706600A (en) * | 2021-08-19 | 2021-11-26 | 北京欧应信息技术有限公司 | Method, apparatus and medium for measuring critical dimensions of a body part |
| CN113674261A (en) * | 2021-08-26 | 2021-11-19 | 上海脊影慧智能科技有限公司 | Bone detection method, system, electronic device and storage medium |
| CN114638878A (en) * | 2022-03-18 | 2022-06-17 | 北京安德医智科技有限公司 | Two-dimensional echocardiogram pipe diameter detection method and device based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110400316A (en) | A kind of orthopaedics image measuring method and device based on deep learning | |
| CN112861598B (en) | System and method for human body model estimation | |
| CN107909622B (en) | Model generation method, medical imaging scanning planning method and medical imaging system | |
| JP5054669B2 (en) | Method and system for characterization of knee joint morphology | |
| CN107480677A (en) | The method and device of area-of-interest in a kind of identification three-dimensional CT image | |
| CN107492097A (en) | A kind of method and device for identifying MRI image area-of-interest | |
| CN110246580B (en) | Cranial image analysis method and system based on neural network and random forest | |
| WO2020114632A1 (en) | Three-dimensional shape reconstruction from a topogram in medical imaging | |
| CN110634133A (en) | Knee joint orthopedic measurement method and device based on X-ray plain film | |
| JP2011520190A (en) | System for mapping objects across different images using image alignment | |
| KR102389067B1 (en) | Method for evaluating scoliosis and device evaluating scoliosis using the same | |
| JP4559460B2 (en) | Size measuring device, image display device, size measuring program, and image display program | |
| JP4938427B2 (en) | Cerebral hemorrhage volume calculator | |
| WO2011119935A2 (en) | Systems and methods for improved tractographic processing | |
| CN110555860A (en) | Method, electronic device and storage medium for marking rib region in medical image | |
| US11544848B2 (en) | Deep learning based methods and systems for automated subject anatomy and orientation identification | |
| CN116958217B (en) | MRI and CT multi-mode 3D automatic registration method and device | |
| EP3758602A1 (en) | Neural network classification of osteolysis and synovitis near metal implants | |
| CN106981090A (en) | Stepping unidirectional light beam scans the three-dimensional rebuilding method of faultage image in a kind of pipe | |
| WO2021011775A1 (en) | Systems and methods for generating classifying and quantitative analysis reports of aneurysms from medical image data | |
| CN115861656A (en) | Method, apparatus and system for automatically processing medical images to output an alert | |
| CN111445575A (en) | Image reconstruction method and device of Wirisi ring, electronic device and storage medium | |
| CN111918611A (en) | Abnormal display control method for chest X-ray image, abnormal display control program, abnormal display control device, and server device | |
| CN117408908A (en) | Preoperative and intraoperative CT image automatic fusion method based on deep neural network | |
| CN114266848A (en) | Medical image three-dimensional reconstruction method and device |
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 | ||
| CB03 | Change of inventor or designer information |
Inventor after: Jiang Zhiqiang Inventor after: Gong Xiangyang Inventor after: Cheng Guohua Inventor after: He Linyang Inventor after: Ji Hongli Inventor before: Cheng Guohua Inventor before: Xia Haiqi Inventor before: He Linyang Inventor before: Ji Hongli |
|
| CB03 | Change of inventor or designer information | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191101 |
|
| WD01 | Invention patent application deemed withdrawn after publication |