CN110458831B - Scoliosis image processing method based on deep learning - Google Patents
Scoliosis image processing method based on deep learning Download PDFInfo
- Publication number
- CN110458831B CN110458831B CN201910748283.0A CN201910748283A CN110458831B CN 110458831 B CN110458831 B CN 110458831B CN 201910748283 A CN201910748283 A CN 201910748283A CN 110458831 B CN110458831 B CN 110458831B
- Authority
- CN
- China
- Prior art keywords
- image
- spine
- midline
- scoliosis
- processing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 206010039722 scoliosis Diseases 0.000 title claims abstract description 27
- 238000013135 deep learning Methods 0.000 title claims abstract description 15
- 238000003672 processing method Methods 0.000 title claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 24
- 230000011218 segmentation Effects 0.000 claims abstract description 24
- 239000011436 cob Substances 0.000 claims abstract description 14
- 238000012805 post-processing Methods 0.000 claims abstract description 6
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000000034 method Methods 0.000 claims description 22
- 238000002601 radiography Methods 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 11
- 238000000605 extraction Methods 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 7
- 238000005452 bending Methods 0.000 claims description 6
- 238000013527 convolutional neural network Methods 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 230000000877 morphologic effect Effects 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 238000013519 translation Methods 0.000 claims description 3
- 238000012800 visualization Methods 0.000 abstract 1
- 238000013528 artificial neural network Methods 0.000 description 8
- 238000000691 measurement method Methods 0.000 description 5
- 230000002708 enhancing effect Effects 0.000 description 3
- 238000000053 physical method Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000000472 traumatic effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- 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
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30008—Bone
- G06T2207/30012—Spine; Backbone
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Evolutionary Biology (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Multimedia (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the technical field of medical image processing, and discloses a scoliosis image processing method based on deep learning, which comprises the following steps: preprocessing image data; semantic segmentation of a chest picture spine; image post-processing; and obtaining the inflection point of the spinal midline and solving the Cobb angle. The invention realizes the visualization processing of the X-ray film and reduces the labor workload.
Description
Technical Field
The invention belongs to the technical field of medical image processing, and particularly relates to a scoliosis image based on deep learning
And (4) processing the method.
Background
The spine is the medial column of the human body and can be traumatic to the physical appearance and spirit of the human body if a lateral bend occurs. Side of the spine
The curvature is also called scoliosis, which is caused by the deviation of the spinal segment on the coronal plane of the human body from the central line of the back and the lateral curvature
In (3). There are many methods for examining the scoliosis, and the methods can be roughly classified into a physical measurement method and an image measurement method.
The physical measurement method is a method for directly contacting with the back of a human body when measuring the scoliosis, and Adams is mainly used
Forward stooping test, measuring trunk rotation angle by applying scoliosis ruler, measuring rib protuberance and the like; image measuring method
The method is a method that does not directly contact with the back of the human body during examination, and mainly includes Moire (Moire) image measurement method and X-ray film
Measurement methods, structured light measurement methods, laser scanner measurement methods, and the like. Adams forward bending test and X are generally adopted in general survey
The light sheet measurements were combined.
Since the physical measurement method is based on manual detection, manual detection becomes a common survey of a large number of people
Is rather cumbersome, inefficient, and can cause erroneous determinations due to fatigue of physicians.
Disclosure of Invention
The embodiment of the invention aims to provide a scoliosis image processing method based on deep learning, which can process X-ray image
The light sheet is directly visualized, and the labor workload is reduced.
The embodiment of the invention is realized as follows:
a scoliosis image processing method based on deep learning comprises the following steps:
image data preprocessing:
converting X-ray chest radiograph with DICOM image format into PNG format, and adjusting by histogram equalization during conversion
Storing the image in an RGB mode until the image is in the clearest display state; normalizing the image, dividing the image into three color channels of R, G and B, respectively subtracting the average value corresponding to each channel from the pixel value of each pixel point of each channel, dividing the average value by the standard deviation, normalizing all the pixel values from 0 to 255 to 0 to 1, and finally reducing the image to the size of 512 multiplied by 512; data enhancement, adjusting the image data by translation, cutting, rotation, brightness and contrast
C, processing; preparing a data label, marking the data, delineating a spine region, generating a binary data label, wherein each X-ray chest radiography image corresponds to one label, and the label records the spine region needing characteristic extraction;
semantic segmentation of a spine of a chest picture:
the processed X-ray chest radiography image is characterized by extracting by utilizing a convolution neural network, training the network by gradient descent,
dividing the image to be predicted into R, G and B three-channel matrixes, inputting the matrixes into a convolutional neural network, and obtaining an output of
262144, the feature vector of dimensions, reducing this vector to a matrix of 512 by 512, the pixel value range to 0 to 255,
the obtained result is a spine semantic segmentation result graph of the X-ray chest radiography;
image post-processing:
further processing the obtained spine semantic segmentation result picture of the X-ray chest picture, performing morphological processing, and performing
Opening operation processing is carried out, so that the outline becomes smooth, the narrow connection area is disconnected, and burrs and noise are eliminated; by calculating each connected region
Setting an area threshold value, wherein the size of the area threshold value is 1800-2200 pixels, removing a connected region with a small area,
and reserving a communication area with large area; setting a threshold value of pixel value, and performing binarization processing on the whole image, wherein the threshold value is larger than
Setting the pixel value to be 255 and setting the pixel value to be 0 when the pixel value is smaller than the threshold value, and obtaining a binary image of spine semantic segmentation; segmented according to spine semantics
The method comprises the steps of binarizing an image, fitting a midline of a spine, setting an interval with the size of 10-20 pixels, traversing the binarized image from top to bottom according to the interval, solving the coordinate average value of an x axis of which all pixel values are not zero every time, wherein the value is the mid point of the spine after y axis down sampling, and finally obtaining a down-sampled midline of the spine; fitting a curve by using a 6 th order polynomial according to the down-sampled spinal midline to obtain a reconstructed spinal fitted midline;
obtaining a spinal midline inflection point, and solving a Cobb angle:
obtaining a midline inflection point, and solving a second derivative of the spinal midline obtained by reconstruction and fitting to obtain the inflection point of the midline; if there is only one crutch
Point, there is no scoliosis; if the number of turning points is N, N is more than or equal to 2, the number of scoliosis angles is equal to N-1, according to
The inflection points calculate the included angle Cobb angle between two adjacent inflection points, for each two adjacent inflection points, calculate the tangent line of the inflection point to the central line of the spine,
the included angle of intersection of the two tangent lines is the Cobb angle, namely the angle of scoliosis.
Wherein the area threshold size is 2000 pixels.
Wherein the interval size is 14 pixels.
Embodiments of the present invention perform automatic feature extraction by employing a convolutional neural network specifically adapted for medical images, in contrast to
The traditional manual extraction of focus features greatly shortens the time and difficulty of algorithm development, has higher robustness and accuracy,
secondly, the scheme of fitting and reconstructing the midline of the spine according to the spine segmentation result is adopted, so that the whole process is more direct and simple
And the error is small, and the reconstruction precision is high.
Drawings
FIG. 1 is a flow chart of a scoliosis image processing method based on deep learning according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more clearly understood, the following description is made with reference to the accompanying drawings and embodiments
The present invention will be described in further detail. It should be understood that the specific embodiments described herein are merely illustrative of the invention and that
And are not intended to limit the invention.
In order to solve the problems of large workload, complexity and the like in the prior art, the invention provides a spine based on deep learning
A method for measuring the lateral bending angle of a column is a technique for accurately detecting and quantitatively measuring the lateral bending of the spine on an X-ray chest film, in particular to a method for measuring the lateral bending angle of the column
A chest radiography spine semantic segmentation method based on deep learning and image processing and an image processing curve fitting method. As shown in figure 1
The technical scheme of the invention comprises the following steps:
preprocessing images, converting original images of medical images in a DICOM format into a PNG format, enhancing the images, enhancing training data, and labeling data to prepare labels;
training a convolutional neural network, and predicting chest slices to obtain a semantic segmentation graph of the spine;
performing image post-processing, and fitting a central line of the spine according to the semantic segmentation map of the spine;
solving the inflection point of the curve according to the spinal midline;
tangent lines are calculated at the turning points, and the included angle between every two tangent lines is the cobb angle, namely the quantitative angle for measuring the scoliosis;
the invention is characterized in that a convolution neural network specially suitable for medical images is adopted for automatic feature extraction, compared with the traditional method
The focus characteristics are manually extracted, the time and the difficulty of algorithm development are greatly shortened, and the robustness and the accuracy are higher. Secondly, the first step is to carry out the first,
the scheme of fitting and reconstructing the midline of the spine according to the spine segmentation result is adopted, so that the whole process is more direct and simpler, and the error is caused
The difference is small, and the reconstruction precision is high.
A scoliosis image processing method based on deep learning comprises the following steps:
image data preprocessing:
converting X-ray chest radiograph with DICOM image format into PNG format, and adjusting by histogram equalization during conversion
Storing the image in an RGB mode until the image is in the clearest display state; normalizing the image, dividing the image into three color channels of R, G and B, respectively subtracting the average value corresponding to each channel from the pixel value of each pixel point of each channel, dividing the average value by the standard deviation, normalizing all the pixel values from 0 to 255 to 0 to 1, and finally reducing the image to the size of 512 multiplied by 512; data enhancement, adjusting the image data by translation, cutting, rotation, brightness and contrast
C, processing; preparing a data label, marking the data, outlining a spine region, generating a binary data label, wherein each sheet is provided with a plurality of data labels
The X-ray chest radiography image corresponds to a label, and the label records a spine region needing characteristic extraction;
semantic segmentation of the chest spine:
the processed X-ray chest radiography image is characterized by utilizing a convolution neural network to extract features, training the network by gradient descent, dividing the image to be predicted into a matrix with three channels of R, G and B, inputting the matrix into the convolution neural network, and obtaining the output of the matrix as
262144, the feature vector of the dimension, reducing this vector to a matrix of 512 times 512, the pixel value range to 0 to 255,
the obtained result is a spine semantic segmentation result graph of the X-ray chest radiography;
and (3) image post-processing:
further processing the obtained spine semantic segmentation result picture of the X-ray chest picture, performing morphological processing, and performing
Opening operation processing is carried out, so that the outline becomes smooth, the narrow connection area is disconnected, and burrs and noise are eliminated; by calculating each connected region
Setting an area threshold value, wherein the size of the area threshold value is 1800-2200 pixels, removing a connected region with small area,
and reserving a communication area with a large area; setting a threshold value of pixel value, and performing binarization processing on the whole image, wherein the threshold value is larger than
Setting the pixel value to be 255 and setting the pixel value to be 0 when the pixel value is smaller than the threshold value to obtain a binary image of spine semantic segmentation; segmented according to spinal semantics
The method comprises the steps of binarizing an image, fitting a midline of a spine, setting an interval with the size of 10-20 pixels, traversing the binarized image from top to bottom according to the interval, solving the coordinate average value of an x axis of which all pixel values are not zero every time, wherein the value is the mid point of the spine after y axis down sampling, and finally obtaining a down-sampled midline of the spine; fitting a curve by using a 6 th order polynomial according to the down-sampled spinal midline to obtain a reconstructed spinal fitted midline;
obtaining a spinal midline inflection point, and solving a Cobb angle:
obtaining a midline inflection point, and solving a second derivative of the spinal midline which is obtained by reconstruction fitting to obtain the inflection point of the midline; if there is only one crutch
Point, there is no scoliosis; if the number of the turning points is N, and N is more than or equal to 2, the number of the scoliosis angles is equal to N-1, according to
The inflection points calculate the included angle Cobb angle between two adjacent inflection points, for each two adjacent inflection points, calculate the tangent line of the inflection point to the central line of the spine,
the included angle of intersection of the two tangent lines is the Cobb angle, namely the angle of scoliosis.
Wherein the area threshold size is 2000 pixels.
Wherein the interval size is 14 pixels.
The following detailed description of specific implementations of the present invention is provided in conjunction with specific embodiments:
a scoliosis image processing method based on deep learning comprises the following steps:
(1) Image data pre-processing
(1.1) converting the X-ray chest film with the image format of DICOM into PNG format. Equalization by histogram during conversion
And adjusting to the clearest display state. The image is saved in RGB form.
And (1.2) carrying out normalization processing on the image. The image is divided into three color channels R, G, B, one for each channel
The pixel value of each pixel point subtracts the average value corresponding to the channel, and then the average value is divided by the standard deviation. Then all the pixel values are processed
Normalized from 0 to 255 to 0 to 1. Finally the image is reduced to a size of 512 times 512.
And (1.3) enhancing data. For training the deep neural network, data enhancement needs to be performed on data to improve spirit
Robustness over a network. Specifically, the data is randomly translated, cut, rotated, and adjusted in brightness and contrast.
And (1.4) preparing a data label. Marking the data, outlining the spine region, and generating the binary data mark
And labeling, wherein each X-ray chest radiography image corresponds to a label, and the label records the spine region needing the neural network to extract the characteristics.
(2) Chest piece spine semantic segmentation
(2.1) feature extraction is carried out on the processed X-ray chest radiography image by utilizing a convolution neural network, and gradient descent is adopted
To train the network. Gradient descent is a common means for training the network in deep learning, specifically setting a loss
Function, calculating the loss of each iteration network, back-propagating and updating parameters, so that the next loss function can be continuously realized
And becomes smaller. Dividing the image to be predicted into R, G and B three-channel matrix, and inputting the matrix into the trained convolutional neural network
In (3), a 262144-dimensional feature vector is obtained as an output.
And (2.2) reducing the vector into a matrix of 512 multiplied by 512, and reducing the pixel value range into 0 to 255, wherein the spine semantic segmentation result of the X-ray chest picture is obtained.
(3) Image post-processing
And (3.1) the obtained neural network original segmentation map needs further image processing. Firstly, the morphological treatment is carried out,
and performing opening operation processing to smooth the contour, disconnecting the narrow connection area and eliminating burrs and noise.
(3.2) removing the connected regions having a small area, e.g., by calculating the area of each connected region and setting a threshold value
If all connected regions in the spine segmentation map of one chest image are small, we consider the spine analysis of the chest image to fail.
And (3.3) setting a threshold value and carrying out binarization processing on the image. The pixel value greater than the threshold is set to 255 and less than
The threshold is set to 0. Obtaining a binary image of spine semantic segmentation.
And (3.4) fitting the midline of the spine according to the binary image of the semantic segmentation of the spine. Setting an interval, first holding two
The valued image is traversed from top to bottom according to the interval, and the coordinate mean value of the x axis of all the pixel values which are not zero at each time is obtained, and the value is the spine midpoint after the y axis down sampling. Finally, a down-sampled spinal midline is obtained.
(3.5) fitting a curve with a 6 th order polynomial based on the downsampled spinal midline to obtain a reconstructed spinal column
The spine is fitted to a midline.
(4) Calculating the inflection point of the spinal midline and calculating the Cobb angle
And (4.1) calculating a midline inflection point. And solving a second derivative of the spinal midline which is fit by reconstruction to obtain an inflection point of the midline. If only
If there is an inflection point, there is no scoliosis, if the number of inflection points N is greater than or equal to 2, the number of scoliosis angles is equal to N-1.
And (4.2) calculating an included angle Cobb angle between two adjacent inflection points according to the inflection points. For every two adjacent inflection points, the tangent to the midline of the spine at the inflection points is solved, and the intersected included angle of the two tangents is the Cobb angle, namely the angle of the scoliosis.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention in any way
And any modifications, equivalents, improvements and the like made within the principles are intended to be included within the scope of the present invention.
Claims (3)
1. A scoliosis image processing method based on deep learning is characterized by comprising the following steps:
image data preprocessing:
converting the X-ray chest radiograph with the DICOM format into a PNG format, adjusting the X-ray chest radiograph to be in the clearest display state through histogram equalization in the conversion process, and storing the image in an RGB (red, green and blue) form; normalizing the image, dividing the image into R, G and B color channels, respectively subtracting the average value corresponding to each channel from the pixel value of each pixel point of each channel, dividing the average value by the standard deviation, normalizing all the pixel values from 0 to 255 to 0 to 1, and finally reducing the image to 512 times 512; data enhancement, namely adjusting translation, cutting, rotation, brightness and contrast of image data; preparing a data label, marking data, outlining a spine region, generating a binary data label, wherein each X-ray chest radiography image corresponds to one label, and the label records the spine region needing characteristic extraction;
semantic segmentation of a spine of a chest picture:
performing feature extraction on the processed X-ray chest radiography image by using a convolutional neural network, training the network by using gradient descent, dividing the image to be predicted into a matrix with three channels of R, G and B, inputting the matrix into the convolutional neural network to obtain a feature vector with 262144 dimensions as output, reducing the vector into a matrix with 512 multiplied by 512, and reducing the pixel value range into 0 to 255, wherein the obtained result is a spine semantic segmentation result image of the X-ray chest radiography;
and (3) image post-processing:
carrying out further image processing on the spine semantic segmentation result picture of the obtained X-ray chest picture, carrying out morphological processing, carrying out opening operation processing to enable the outline to become smooth, disconnecting a narrow connection area, and eliminating burrs and noise; setting an area threshold value by calculating the area of each connected region, wherein the size of the area threshold value is 1800-2200 pixels, removing the connected regions with small areas and reserving the connected regions with large areas; setting a pixel value threshold, carrying out binarization processing on the whole image, setting the pixel value larger than the threshold to be 255 and setting the pixel value smaller than the threshold to be 0, and obtaining a spine semantic segmentation binarization image; fitting a central line of a spine according to a binary image segmented by spine semantics, setting an interval, wherein the interval is 10-20 pixels, traversing the binary image from top to bottom according to the interval, solving a coordinate average value of an x axis of which all pixel values are not zero every time, wherein the value is a central point of the spine after y-axis down-sampling, and finally obtaining a down-sampled central line of the spine; fitting a curve by using a 6 th-order polynomial according to the down-sampled spinal midline to obtain a reconstructed spinal fitted midline;
obtaining a spinal midline inflection point, and solving a Cobb angle:
obtaining a midline inflection point, and solving a second derivative of the spinal midline which is obtained by reconstruction fitting to obtain the inflection point of the midline; if there is only one inflection point, there is no scoliosis; if the number of the turning points is N, and N is more than or equal to 2, the number of the lateral bending angles of the spine is equal to N-1, an included angle Cobb angle between two adjacent turning points is solved according to the turning points, for each two adjacent turning points, a tangent line to the central line of the spine at the turning point is solved, and the intersected included angle of the two tangent lines is the Cobb angle, namely the lateral bending angle of the spine.
2. The scoliosis image processing method based on deep learning of claim 1, wherein: the area threshold size is 2000 pixels.
3. The scoliosis image processing method based on deep learning of claim 1, wherein: the spacing is 14 pixels in size.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910748283.0A CN110458831B (en) | 2019-08-12 | 2019-08-12 | Scoliosis image processing method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910748283.0A CN110458831B (en) | 2019-08-12 | 2019-08-12 | Scoliosis image processing method based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110458831A CN110458831A (en) | 2019-11-15 |
| CN110458831B true CN110458831B (en) | 2023-02-03 |
Family
ID=68486444
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910748283.0A Active CN110458831B (en) | 2019-08-12 | 2019-08-12 | Scoliosis image processing method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110458831B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111222468A (en) * | 2020-01-08 | 2020-06-02 | 浙江光珀智能科技有限公司 | People stream detection method and system based on deep learning |
| CN111666890B (en) * | 2020-06-08 | 2023-06-30 | 平安科技（深圳）有限公司 | Spine deformation crowd identification method and device, computer equipment and storage medium |
| CN112274164B (en) * | 2020-07-29 | 2023-02-21 | 深圳市智影医疗科技有限公司 | Scoliosis prediction method, scoliosis prediction device, electronic device, and storage medium |
| CN112381757A (en) * | 2020-10-09 | 2021-02-19 | 温州医科大学附属第二医院、温州医科大学附属育英儿童医院 | System and method for measuring and calculating scoliosis Cobb angle through full-length X-ray film of spine based on artificial intelligence-image recognition |
| CN112381869B (en) * | 2020-10-09 | 2022-12-13 | 温州医科大学附属第二医院、温州医科大学附属育英儿童医院 | Measuring and calculating device for measuring and calculating scoliosis by using full-length X-ray film of spine |
| CN114549619A (en) * | 2020-11-24 | 2022-05-27 | 奥泰医疗系统有限责任公司 | Deep learning-based MRI (magnetic resonance imaging) spine image intervertebral disc scanning positioning and vertebra identification method |
| CN112801979B (en) * | 2021-01-28 | 2023-09-05 | 广东工业大学 | Scoliosis detection method and system based on uncertainty reasoning |
| CN112734757B (en) * | 2021-03-29 | 2021-06-25 | 成都成电金盘健康数据技术有限公司 | Spine X-ray image cobb angle measuring method |
| CN113628740B (en) * | 2021-06-22 | 2023-05-02 | 四川大学 | Children spine morphology rapid analysis method based on 3D photographing technology |
| CN113781453B (en) * | 2021-09-15 | 2024-04-02 | 南京大学 | Scoliosis advancing and expanding prediction method and device based on X-ray film |
| CN114078125A (en) * | 2021-11-29 | 2022-02-22 | 开封市人民医院 | Vertebra image processing method based on nuclear magnetic image |
| CN114287915B (en) * | 2021-12-28 | 2024-03-05 | 深圳零动医疗科技有限公司 | Noninvasive scoliosis screening method and system based on back color images |
| CN114511468B (en) * | 2022-03-31 | 2022-11-11 | 西北民族大学 | Distorted bone image correction method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109431511A (en) * | 2018-11-14 | 2019-03-08 | 南京航空航天大学 | A kind of human body back scoliosis angle measurement method based on Digital Image Processing |
| CN109903277A (en) * | 2019-02-25 | 2019-06-18 | 电子科技大学 | A kind of scoliosis detection method based on polynomial curve fitting |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070242869A1 (en) * | 2006-04-12 | 2007-10-18 | Eastman Kodak Company | Processing and measuring the spine in radiographs |
| EP3107031A1 (en) * | 2015-06-18 | 2016-12-21 | Agfa HealthCare | Method, apparatus and system for spine labeling |
| EP3470006B1 (en) * | 2017-10-10 | 2020-06-10 | Holo Surgical Inc. | Automated segmentation of three dimensional bony structure images |
-
2019
- 2019-08-12 CN CN201910748283.0A patent/CN110458831B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109431511A (en) * | 2018-11-14 | 2019-03-08 | 南京航空航天大学 | A kind of human body back scoliosis angle measurement method based on Digital Image Processing |
| CN109903277A (en) * | 2019-02-25 | 2019-06-18 | 电子科技大学 | A kind of scoliosis detection method based on polynomial curve fitting |
Non-Patent Citations (2)
| Title |
|---|
| 基于深度学习的脊柱CT图像分割;刘忠利等;《计算机应用与软件》;20181012(第10期);全文 * |
| 数字化X线胸片计算机辅助诊断鉴别肺结节良恶性研究;权申文等;《新发传染病电子杂志》;20180531;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110458831A (en) | 2019-11-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110458831B (en) | Scoliosis image processing method based on deep learning | |
| CN108537784B (en) | CT image pulmonary nodule detection method based on deep learning | |
| CN108665456B (en) | Method and system for real-time marking of breast ultrasound lesion region based on artificial intelligence | |
| CN111784671B (en) | Pathological image focus region detection method based on multi-scale deep learning | |
| CN112150428B (en) | Medical image segmentation method based on deep learning | |
| CN108765363B (en) | Coronary artery CTA automatic post-processing system based on artificial intelligence | |
| CN111407245B (en) | Non-contact heart rate and body temperature measuring method based on camera | |
| CN108615239B (en) | Tongue image segmentation method based on threshold technology and gray level projection | |
| WO2021136368A1 (en) | Method and apparatus for automatically detecting pectoralis major region in molybdenum target image | |
| CN110782428A (en) | Method and system for constructing clinical brain CT image ROI template | |
| WO2022048171A1 (en) | Method and apparatus for measuring blood vessel diameter in fundus image | |
| CN113139929A (en) | Gastrointestinal tract endoscope image preprocessing method comprising information screening and fusion repairing | |
| CN104933723B (en) | Tongue image dividing method based on rarefaction representation | |
| CN110211200A (en) | A kind of arch wire generation method and its system based on nerual network technique | |
| CN116758336A (en) | Medical image intelligent analysis system based on artificial intelligence | |
| CN114693672A (en) | Mammary gland molybdenum target image skin gland and nipple removing method based on image processing | |
| CN114708237A (en) | Detection algorithm for hair health condition | |
| CN109785325A (en) | A method of the Multimodal medical image based on deep learning | |
| CN108629780B (en) | Tongue image segmentation method based on color decomposition and threshold technology | |
| CN117522862A (en) | Image processing method and processing system based on CT image pneumonia recognition | |
| CN109949906B (en) | Pathological section image processing and model training method and equipment | |
| Klinder et al. | Lobar fissure detection using line enhancing filters | |
| US20230115927A1 (en) | Systems and methods for plaque identification, plaque composition analysis, and plaque stability detection | |
| CN117252825A (en) | Dental caries identification method and device based on oral panoramic image | |
| WO2022126923A1 (en) | Asc-us diagnosis result identification method and apparatus, computer device, and storage medium |
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 |