CN109978861B - Polio detection method, apparatus, device and computer readable storage medium - Google Patents
Polio detection method, apparatus, device and computer readable storage medium Download PDFInfo
- Publication number
- CN109978861B CN109978861B CN201910237817.3A CN201910237817A CN109978861B CN 109978861 B CN109978861 B CN 109978861B CN 201910237817 A CN201910237817 A CN 201910237817A CN 109978861 B CN109978861 B CN 109978861B
- Authority
- CN
- China
- Prior art keywords
- image
- area
- bone marrow
- mri
- initial
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Quality & Reliability (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Image Processing (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention discloses a method, a device, equipment and a computer readable storage medium for detecting poliosis, wherein the method comprises the following steps: sequentially inputting each MRI image of the vertebra into a pre-trained U-Net model for bone marrow gray segmentation to obtain at least one initial bone marrow gray image; determining a pixel point from the selected initial bone marrow gray image corresponding to the MRI image belonging to the middle part of the vertebra, and collecting position information of the pixel point; respectively selecting seed points from the MRI images to be detected according to the position information, and respectively extracting segmentation images containing the seed points from the MRI images to be detected by using a region growing algorithm; and when the error between the area of the extracted segmentation image and the area of the corresponding initial bone marrow grey matter image is smaller than or equal to a preset error, determining that the bone marrow grey matter area exists in the MRI image to be detected corresponding to the segmentation image. The invention is suitable for detecting the spinal polio area, further improves the speed of processing detection data, and has high automation and intelligence degree.
Description
Technical Field
The invention relates to the technical field of gray matter detection, in particular to a method, a device, equipment and a computer readable storage medium for detecting bone marrow gray matter.
Background
The medical image can be used for describing the morphology of different tissues and organs of a human body and reflecting the health condition of the human body, and is helpful for auxiliary diagnosis and treatment. In the conventional art, the mode that adopts artifical the detection to the detection of bone marrow grey matter for the majority to lead to inefficiency and error big, even if there is the technique of machine automated inspection bone marrow grey matter, also there is intelligent degree not high when detecting in the face of a large amount of data, cause with high costs and detect the low not good problem that leads to the detection effect of precision.
Disclosure of Invention
The present invention is directed to provide a method, an apparatus, a device and a computer-readable storage medium for detecting gray matter in bone marrow.
The method for detecting the polio of the embodiment of the invention comprises the following steps:
sequentially inputting each MRI image of the vertebra into a pre-trained U-Net model for bone marrow gray segmentation to obtain at least one initial bone marrow gray image;
determining a pixel point from the selected initial bone marrow gray image corresponding to the MRI image belonging to the middle part of the vertebra, and collecting position information of the pixel point;
respectively selecting seed points from the MRI images to be detected according to the position information, and respectively extracting segmentation images containing the seed points from the MRI images to be detected by using a region growing algorithm, wherein each MRI image to be detected is an MRI image corresponding to each initial grey bone matter image;
and when the error between the area of the extracted segmentation image and the area of the corresponding initial bone marrow grey matter image is smaller than or equal to a preset error, determining that the bone marrow grey matter area exists in the MRI image to be detected corresponding to the segmentation image.
In one embodiment, "when an error between an area of an extracted segmented image and an area of a corresponding initial gray matter image is less than or equal to a preset error, determining that an MRI image to be measured corresponding to the segmented image has a gray matter region", the method further includes:
taking the area of the initial bone marrow gray matter image corresponding to each MRI image with the bone marrow gray matter area as a first area, and taking the area of the corresponding segmentation image as a second area;
the average value of the first area and the corresponding second area is used as the final area of the gray matter region of the bone marrow of the MRI image in which the gray matter region exists.
In one embodiment, after "taking the average of the first area and the corresponding second area as the final area of the gray bone area of the MRI image where the gray bone area exists", the method further includes:
acquiring the corresponding layer thickness of the MRI image of each area with the bone marrow gray matter;
calculating the product of each layer thickness and the corresponding final area of the polio area;
the sum of the products is calculated and taken as the total volume of the grey matter of the spinal bone.
In one embodiment, the training process of the trained U-Net model comprises the following steps:
inputting the obtained MRI sample images of a preset number into a preset initial U-Net model to obtain each gray bone matter training image output after the MRI sample images are segmented by the initial U-Net model;
calculating the deviation of the bone marrow gray matter training image and the bone marrow gray matter standard image of the corresponding MRI sample image;
and updating the weight coefficient of the initial U-Net model by using a back propagation algorithm based on the deviation until the updated initial U-Net model meets the preset condition, and taking the finally updated initial U-Net model as the trained U-Net model.
In one embodiment, "selecting seed points from each MRI image to be measured according to the position information, and extracting segmented images including the seed points from each MRI image to be measured by using a region growing algorithm", includes:
selecting initial pixel points with the same position information from an MRI image to be detected, and taking the initial pixel points as seed points;
taking the seed points as growth starting points, and gradually merging adjacent pixel points meeting a preset growth rule into a current image area;
and extracting the current image area which is combined completely as a segmentation image.
On the other hand, the embodiment of the invention also provides a polio detecting device, which comprises:
the initial image acquisition module is used for sequentially inputting each MRI image of the vertebra into a pre-trained U-Net model for bone marrow gray matter segmentation to acquire at least one initial bone marrow gray matter image;
the selection module is used for determining a pixel point from the selected initial bone marrow gray image corresponding to the MRI image belonging to the middle part of the vertebra and collecting the position information of the pixel point;
the region growing module is used for respectively selecting seed points from the MRI images to be detected according to the position information and respectively extracting segmentation images containing the seed points from the MRI images to be detected by using a region growing algorithm, wherein each MRI image to be detected is an MRI image corresponding to each initial bone marrow gray matter image;
and the bone marrow grey matter determining module is used for determining that the MRI image to be detected corresponding to the segmentation image has a bone marrow grey matter area when the error between the area of the extracted segmentation image and the area of the corresponding initial bone marrow grey matter image is less than or equal to a preset error.
In one embodiment, the method further comprises the following steps:
the area determination module is used for taking the area of the initial bone marrow gray matter image corresponding to each MRI image with the bone marrow gray matter area as a first area and taking the area of the corresponding segmentation image as a second area;
and the calculation module is used for taking the average value of the first area and the corresponding second area as the final area of the polio area of the MRI image of the polio area.
In one embodiment, the method further comprises the following steps:
the layer thickness information acquisition module is used for acquiring the layer thickness corresponding to the MRI image of each area with the bone marrow gray matter;
a layer volume calculation module for calculating the product of each layer thickness and the corresponding final area of the polio area;
and the gray matter volume calculation module is used for calculating the sum of the products and taking the sum as the total volume of the bone marrow matter of the vertebra.
On the other hand, the embodiment of the invention also provides bone marrow ash detection equipment, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the bone marrow ash detection method when executing the computer program.
In another aspect, an embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the computer program implements a gray bone detection method.
One of the above technical solutions has the following advantages and beneficial effects:
the method, the device, the equipment and the computer readable storage medium for detecting the gray matter of the bone marrow can segment the initial gray matter image in the MRI image data of the vertebra based on the trained U-Net model. Because bone marrow is present in the middle of the spine, it is helpful to select a corresponding image in the middle of the spine from the initial bone marrow gray images. In order to improve the precision of detecting the vertebra gray matter region and prevent misjudgment, a pixel point is determined in the selected initial gray matter image, the position information of the pixel point is collected, the pixel point is used as the basis for selecting a seed point, and a corresponding segmentation image is provided in each MRI image to be detected based on a region growing algorithm. And through comparison of the segmented image and the initial bone marrow grey image, if the area error is within a preset error range, determining that the corresponding MRI image to be detected has a bone marrow grey area. According to the embodiment of the invention, the double detection is carried out on the poliomyelike area through the U-Net model in combination with the area growth algorithm and according to the poliomyelike distribution characteristics of the vertebra, so that the poliomyelike area can be quickly positioned, and meanwhile, the problem of false detection and false judgment can be prevented. The invention is particularly suitable for detecting the spinal polio area, further improves the speed of processing detection data and has high automation and intelligence degree.
Drawings
FIG. 1 is a schematic flow chart of a gray matter detection method according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of gray matter volume calculation provided by an embodiment of the gray matter detection method of the present invention;
FIG. 3 is a schematic flow chart of region growing provided by an embodiment of the gray matter detection method of the present invention;
FIG. 4 is a schematic diagram of the configuration of the gray matter detection apparatus of the present invention;
fig. 5 is a schematic structural view of the gray matter detection apparatus of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail below with reference to the accompanying drawings in conjunction with embodiments.
Referring to fig. 1, in one embodiment, the present invention provides a polio detection method, comprising the steps of:
step S110: and sequentially inputting each MRI image of the vertebra into a pre-trained U-Net model for bone marrow gray matter segmentation, and acquiring at least one initial bone marrow gray matter image.
The U-Net model is a convolution neural network for biomedical image segmentation, and is suitable for segmentation of medical images. The pre-trained U-Net model in the embodiment of the invention is obtained by training an initial U-Net model through MRI (Magnetic Resonance Imaging) sample images of multiple vertebras, and can segment a polio area in the MRI sample image. Therefore, each MRI image of the spine is input to the trained U-Net model, and the trained U-Net model can segment an initial bone marrow gray image that is considered to contain bone marrow gray. The spine MRI image of each detected person includes a plurality of pages.
The method for detecting the gray matter of the bone marrow can automatically identify the gray matter of the bone marrow in each MRI image of the vertebra based on the pre-trained U-Net model, so that the initial gray matter image of the bone marrow is quickly segmented.
Step S120: and determining a pixel point from the selected MRI image belonging to the middle part of the vertebra corresponding to the initial gray bone image, and collecting the position information of the pixel point.
Since the gray bone is present in the middle of the spine, when MRI scanning is performed on the spine, the spine is scanned from the top down or from the bottom up until the entire spine portion is scanned. Therefore, based on the initial bone marrow gray image obtained in step S110, the MRI image corresponding to the middle of the spine can be selected, because the MRI image of the middle of the spine is determined to have bone marrow gray. Wherein, each page of MRI image is marked with a sequence number to show the sequence of the graph, and the sequence of the sequence number is determined according to the sequence of scanning the spine. Therefore, the initial bone marrow gray matter image corresponding to the MRI image of the central vertebra can be selected according to the serial number, for example, the page corresponding to the middle of the serial number is the MRI image of the central vertebra. If a total of 200 MRI images of the vertebra of one examiner are scanned, about 100 central pages are scanned, and if the 100 th page is the MRI image corresponding to the middle part of the vertebra, the initial bone marrow gray matter image corresponding to the MRI image is further obtained. The selection mode can be manual selection or automatic selection by a computer.
Specifically, a pixel point determined in the initial gray matter image corresponding to the MRI image in the middle of the vertebra is any point in the gray matter region, and preferably, may be a center of mass point of the gray matter region. Determining a pixel point in an initial gray bone matter image corresponding to the MRI image in the middle of the vertebra, and collecting position information of the pixel point, wherein the position information can be a pixel point coordinate, because the gray bone matter is distributed in the middle of the vertebra based on the distribution condition of the gray bone matter in the vertebra, the MRI image in the middle of the vertebra is determined to have the gray bone matter, and then the position of the pixel point is one of the positions indicating the existence of the gray bone matter.
On the other hand, the pre-trained U-Net model performs segmentation of the polio area on the spine MRI image of the person to be detected page by page, but misjudgment may occur because all MRI images may not have the polio. Therefore, the trained U-Net model can mistakenly segment an initial bone marrow gray image from an MRI image without bone marrow gray matter. The position information can be used as a reference basis for detecting whether the MRI images to be detected determine the existence of the gray bone matter. Therefore, the position information of the pixel point can be used as a detection standard, and the gray bone area can be further detected in the MRI image corresponding to each initial gray bone image, so that the accuracy of gray bone detection can be guaranteed.
Step S130: and respectively selecting seed points from the MRI images to be detected according to the position information, and respectively extracting segmentation images containing the seed points from the MRI images to be detected by using a region growing algorithm, wherein each MRI image to be detected is an MRI image corresponding to each initial polio image.
The region growing algorithm can specify a seed point in the image as a growing starting point, then compare pixel points in the field around the growing starting point with the growing starting point, combine points with similar attributes to continue growing outwards until pixel points which do not meet the conditions are included. Specifically, according to the position information collected in step S120, a relatively close position is quickly located in each MRI image to be measured, for example, the distance between the pixel coordinates of the position and the pixel coordinates in the position information is less than or equal to a preset distance, and preferably, the located position is the same as the position information, that is, the pixel coordinates of the two are the same. Therefore, the positioned pixel points corresponding to the positions are used as seed points, and region growth is carried out based on the seed points to obtain a segmentation image containing the seed points.
The method for detecting the gray matter of bone marrow of the embodiment of the invention can effectively position the corresponding position in each MRI image to be detected based on the position information determined in the step S120, thereby accurately finding out the position where the gray matter of bone marrow should exist to select the seed point, having clear selection target and reducing the redundancy of the program. Meanwhile, pixel points with similar attributes to the seed points can be merged in each MRI image to be detected by using a region growing algorithm, and whether the polio exists in the MRI image to be detected can be further detected, so that misjudgment caused by a single detection mode is avoided.
Step S140: and when the error between the area of the extracted segmentation image and the area of the corresponding initial bone marrow grey matter image is smaller than or equal to a preset error, determining that the bone marrow grey matter area exists in the MRI image to be detected corresponding to the segmentation image.
If the MRI image to be detected has pixel points with similar attributes to the seed points, the segmented images with certain area and containing the seed points can be merged through the region growing algorithm. Further, if the MRI image to be detected has a bone marrow gray image, the area of the obtained segmented image is smaller than the area error of the initial bone marrow gray image corresponding to the MRI image to be detected; if the MRI image to be detected does not have the polio image originally, the error of the segmentation image and the initial polio image corresponding to the MRI image to be detected is large, and the MRI image to be detected in the area where the polio does not exist can be abandoned. Preferably, the preset error may be set to 20%.
The method for detecting the gray matter of the bone marrow can segment the initial gray matter image in the MRI image data of the vertebra based on the trained U-Net model. Because bone marrow is present in the middle of the spine, it is helpful to select a corresponding image in the middle of the spine from the initial bone marrow gray images. In order to improve the precision of detecting the vertebra gray matter region and prevent misjudgment, a pixel point is determined in the selected initial gray matter image, the position information of the pixel point is collected, the pixel point is used as the basis for selecting a seed point, and a corresponding segmentation image is provided in each MRI image to be detected based on a region growing algorithm. And through comparison of the segmented image and the initial bone marrow grey image, if the area error is within a preset error range, determining that the corresponding MRI image to be detected has a bone marrow grey area. According to the embodiment of the invention, the dual detection is carried out on the poliomyelike area through the U-Net model in combination with the area growth algorithm and according to the poliomyelike distribution characteristics of the vertebra, so that the poliomyelike area can be quickly positioned, and meanwhile, the problem of false detection and false judgment can be prevented. The invention is particularly suitable for detecting the spinal polio area, further improves the speed of processing detection data and has high automation and intelligence degree.
In a specific embodiment, "when an error between an area of the extracted segmented image and an area of the corresponding initial gray matter image is less than or equal to a preset error, determining that the MRI image to be tested corresponding to the segmented image has a gray matter area", the method further includes:
step S2: the area of the initial bone marrow gray matter image corresponding to each MRI image in which a bone marrow gray matter region exists is taken as a first area, and the area of the corresponding segmented image is taken as a second area.
Step S4: the average value of the first area and the corresponding second area is used as the final area of the gray matter region of the bone marrow of the MRI image in which the gray matter region exists.
According to the method for detecting the gray matter of bone marrow, disclosed by the embodiment of the invention, after the fact that the gray matter of bone marrow exists in the MRI image to be detected is detected, the first area of the initial gray matter of bone marrow corresponding to the MRI image in the region where the gray matter of bone marrow exists and the mean value of the second areas of the corresponding segmentation images can be comprehensively determined. This can further improve the detection accuracy of the gray matter of bone marrow, and prevent erroneous judgment and erroneous detection.
Referring to fig. 2, in a specific embodiment, after "taking the average of the first area and the corresponding second area as the final area of the bone marrow gray area of the MRI image where the bone marrow gray area exists", the method further includes:
step S210: the slice thickness corresponding to the MRI image is acquired for each region where bone marrow ash is present.
The layer thickness corresponding to the MRI image refers to the thickness of the excited layer when performing the MRI scan.
Step S220: the product of each layer thickness and the corresponding final area of the polio area is calculated.
Based on the steps, the volume of the bone marrow ash of the layer corresponding to the excited layer where the final area of the bone marrow ash region is located can be obtained, and the resolution of the calculation of the total volume of the bone marrow ash is improved.
Step S230: the sum of the products is calculated and taken as the total volume of the grey matter of the spinal bone.
According to the method for detecting the poliomy matter, the volume of each excited layer of the spinal poliomy matter is calculated firstly by utilizing the corresponding layer thickness of the MRI image of each area with the poliomy matter, so that the total volume of the poliomy matter is obtained. The embodiment of the invention has high automation and intelligence degree, can effectively and accurately calculate the volume of the gray matter of the bone marrow, has high resolution and is particularly suitable for diagnosing the vertebra.
In a specific embodiment, the training process of the trained U-Net model comprises the following steps:
step S8: and inputting the obtained MRI sample images of the preset number into a preset initial U-Net model to obtain each gray bone matter training image output after the MRI sample images are segmented by the initial U-Net model.
The MRI sample image is an MRI image containing a bone marrow gray matter, and is used for training a U-Net model so that the U-Net model can recognize and segment the bone marrow gray matter in the MRI sample image, and preferably, the MRI sample image can be an MRI image containing a spinal bone marrow gray matter.
Step S10: calculating the deviation of the grey matter training image from the grey matter standard image of the corresponding MRI sample image.
The standard image of the grey matter of the bone marrow of the MRI sample image is an image of the grey matter of the bone marrow in the MRI image, and is used for judging whether the training image of the grey matter of the bone marrow is the MRI image of the grey matter of the bone marrow. The deviation is used for representing the similarity between the bone marrow gray matter training image and the corresponding bone marrow gray matter standard image, the smaller the deviation of the two is, the more similar the two are, the higher the precision of the U-Net model for segmenting the bone marrow gray matter in the MRI image is, and the closer the U-Net model is to the training target.
Step S12: and updating the weight coefficient of the initial U-Net model by using a back propagation algorithm based on the deviation until the updated initial U-Net model meets the preset condition, and taking the finally updated initial U-Net model as the trained U-Net model.
And the initial U-Net model comprises an input layer, a hidden layer and an output layer, and is sequentially updated and the weight coefficient of each node is adjusted from the output layer, the hidden layer to the input layer by using a back propagation algorithm through the deviation of the output gray matter training image and the corresponding gray matter standard image until the initial U-Net model reaches the preset condition, so that the finally updated initial U-Net model meeting the preset condition is used as the trained U-Net model. The preset condition is that the deviation is smaller than or equal to the preset deviation, that is, the accuracy of the updated initial U-Net model for segmenting the bone marrow gray image reaches the preset value, and preferably, the accuracy can be set to 99%.
The method for detecting the gray matter of the bone marrow can train the U-Net model which can accurately segment the gray matter image of the bone marrow from the MRI image in an artificial intelligence mode, has high intelligence degree, can realize automatic detection of the gray matter area of the bone marrow, and can improve the diagnosis efficiency.
Referring to fig. 3, in a specific embodiment, "selecting seed points from each MRI image to be measured according to the position information, and extracting segmented images including the seed points from each MRI image to be measured by using a region growing algorithm", includes:
step S310: and selecting initial pixel points with the same position information from the MRI image to be detected, and taking the initial pixel points as seed points.
Step S320: and combining adjacent pixel points which meet a preset growth rule into the current image area step by taking the seed points as growth starting points.
The seed point is a growth starting point of the region growth and is also an initial current image region, and when the region growth is carried out, the seed point is used as a center to traverse adjacent pixel points. For example, if the difference between the gray value of the adjacent pixel and the gray value of the seed point is less than or equal to the preset gray difference, the adjacent pixel and the seed point are merged into a new current image area, then the pixel adjacent to the pixel merged into the current image area is traversed, if the difference between the gray value of the traversed pixel and the gray value of the pixel merged into the current image area is less than or equal to the preset gray difference, the traversed pixel is merged into the current image area, and the pixel adjacent to the pixel merged into the current image area is continuously traversed, so that the adjacent pixels with similar attributes are gradually merged into the current image area in an iterative manner until no pixel meeting the preset growth rule exists. The preset growth rule is that the gray value difference value between a pixel point which is not merged into the current image area and an adjacent pixel point which belongs to the current image area is smaller than or equal to the preset gray value difference value.
Step S330: and extracting the current image area which is combined completely as a segmentation image.
According to the method for detecting the gray matter of bone marrow, disclosed by the embodiment of the invention, the connected regions with similar characteristics can be identified in the MRI image to be detected through the attribute characteristics of the seed points, so that the gray matter of bone marrow regions meeting the conditions can be conveniently extracted, the efficiency is high, the complete image can be obtained, and the continuity of the image is ensured. The embodiment of the invention particularly improves the detection precision of the spinal poliosis.
Referring to fig. 4, in one embodiment, the present invention also provides a polio detecting device, including:
an initial image obtaining module 410, configured to sequentially input each MRI image of a vertebra into a pre-trained U-Net model to perform gray matter segmentation, so as to obtain at least one initial gray matter image.
The selecting module 420 is configured to determine a pixel point from the initial gray bone image corresponding to the selected MRI image belonging to the middle of the spine, and collect position information of the pixel point.
And the region growing module 430 is configured to select seed points from the MRI images to be detected respectively according to the position information, and extract segmented images including the seed points from the MRI images to be detected respectively by using a region growing algorithm, where each MRI image to be detected is an MRI image corresponding to each initial bone marrow gray matter image.
And a bone marrow gray determining module 440, configured to determine that a bone marrow gray region exists in the MRI image to be detected corresponding to the segmented image when an error between the area of the extracted segmented image and the area of the corresponding initial bone marrow gray image is less than or equal to a preset error.
The bone marrow gray matter detection device provided by the embodiment of the invention can segment the initial bone marrow gray matter image in the MRI image data of the vertebra based on the trained U-Net model. Because bone marrow is present in the middle of the spine, it is helpful to select a corresponding image in the middle of the spine from the initial bone marrow gray images. In order to improve the precision of detecting the vertebra gray matter region and prevent misjudgment, a pixel point is determined in the selected initial gray matter image, the position information of the pixel point is collected, the pixel point is used as the basis for selecting a seed point, and a corresponding segmentation image is provided in each MRI image to be detected based on a region growing algorithm. And through comparison of the segmented image and the initial bone marrow grey image, if the area error is within a preset error range, determining that the corresponding MRI image to be detected has a bone marrow grey area. According to the embodiment of the invention, the dual detection is carried out on the poliomyelike area through the U-Net model in combination with the area growth algorithm and according to the poliomyelike distribution characteristics of the vertebra, so that the poliomyelike area can be quickly positioned, and meanwhile, the problem of false detection and false judgment can be prevented. The invention is particularly suitable for detecting the spinal polio area, further improves the speed of processing detection data and has high automation and intelligence degree.
In a specific embodiment, the method further comprises the following steps:
and the area determining module is used for taking the area of the initial bone marrow gray matter image corresponding to each MRI image with the bone marrow gray matter area as a first area and taking the area of the corresponding segmentation image as a second area.
And the calculation module is used for taking the average value of the first area and the corresponding second area as the final area of the polio area of the MRI image of the polio area.
According to the bone marrow gray matter detection device provided by the embodiment of the invention, after the fact that the bone marrow gray matter exists in the MRI image to be detected is detected, the first area of the initial bone marrow gray matter image corresponding to the MRI image in the region where the bone marrow gray matter exists and the mean value of the second area of the corresponding segmentation image can be comprehensively determined. This can further improve the detection accuracy of the gray matter of bone marrow, and prevent erroneous judgment and erroneous detection.
In a specific embodiment, the method further comprises the following steps:
and the layer thickness information acquisition module is used for acquiring the layer thickness corresponding to each MRI image of the area with the bone marrow gray matter.
And the layer volume calculation module is used for calculating the product of each layer thickness and the corresponding final bone marrow gray area.
And the gray matter volume calculation module is used for calculating the sum of the products and taking the sum as the total volume of the bone marrow matter of the vertebra.
According to the bone marrow gray matter detection device provided by the embodiment of the invention, the volume of each layer of spinal bone marrow gray matter is firstly calculated by utilizing the corresponding layer thickness of the MRI image of each region with bone marrow gray matter, so that the total volume of the bone marrow gray matter is obtained. The embodiment of the invention has high automation and intelligence degree, can effectively and accurately calculate the volume of the gray matter of the bone marrow, has high resolution and is particularly suitable for diagnosing the vertebra.
In a specific embodiment, the method further comprises the following steps:
and the training image acquisition module is used for inputting the acquired MRI sample images of the preset number into a preset initial U-Net model to obtain various output polio training images after the MRI sample images are segmented by the initial U-Net model.
And the deviation calculation module is used for calculating the deviation of the bone marrow gray matter training image and the bone marrow gray matter standard image of the corresponding MRI sample image.
And the learning module is used for updating the weight coefficient of the initial U-Net model by utilizing a back propagation algorithm based on the deviation until the updated initial U-Net model meets a preset condition, and taking the finally updated initial U-Net model as the trained U-Net model.
The bone marrow gray matter detection device provided by the embodiment of the invention can train the U-Net model capable of accurately segmenting the bone marrow gray matter image from the MRI image in an artificial intelligence mode, has high intelligence degree, can realize automatic detection of the bone marrow gray matter area, and can improve the diagnosis efficiency.
In a particular embodiment, the region growing module includes:
and the seed point selecting unit is used for selecting initial pixel points with the same position information from the MRI image to be detected and taking the initial pixel points as seed points.
And the merging unit is used for gradually merging the adjacent pixel points meeting the preset growth rule into the current image area by taking the seed points as growth starting points.
And the extraction unit is used for extracting the current image area which is combined to be used as the segmentation image.
The poliomy detection device provided by the embodiment of the invention can identify the connected regions with similar characteristics in the MRI image to be detected through the attribute characteristics of the seed points, is convenient for extracting the poliomy regions meeting the conditions, has high efficiency, is beneficial to obtaining complete images, and ensures the continuity of the images. The embodiment of the invention particularly improves the detection precision of the spinal poliosis.
Referring to fig. 5, in one embodiment, there is also provided a bone marrow ash detection apparatus, including a memory storing a computer program and a processor implementing the bone marrow ash detection method when the processor executes the computer program.
The gray matter detecting apparatus may be a terminal, and its internal structure diagram may be as shown in fig. 5. The polio detecting equipment comprises a processor, a memory, a network interface, a display screen and an input device which are connected through a system bus. Wherein the processor of the gray matter detection device is configured to provide computational and control capabilities. The memory of the polio detecting device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the polio detecting device is used for communicating with an external terminal through network connection. The computer program is executed by a processor to implement a method of gray matter detection. The display screen of the poliosis detecting equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the poliosis detecting equipment can be a touch layer covered on the display screen, a button, a track ball or a touch pad arranged on the casing of the poliosis detecting equipment, an external keyboard, a touch pad or a mouse and the like.
In an embodiment, there is further provided a computer-readable storage medium having a computer program stored thereon, which when executed by a processor, implements any one of the above embodiments of the gray bone detection method.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative and, for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, each functional module or unit in each embodiment of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention or a part of the technical solution that contributes to the prior art in essence can be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a smart phone, a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
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, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A method for detecting polio, comprising the steps of:
sequentially inputting each MRI image of the vertebra into a pre-trained U-Net model for bone marrow gray segmentation to obtain at least one initial bone marrow gray image;
determining a pixel point from the selected initial bone marrow gray image corresponding to the MRI image belonging to the middle part of the vertebra, and collecting the position information of the pixel point;
respectively selecting seed points from the MRI images to be detected according to the position information, and respectively extracting segmentation images containing the seed points from the MRI images to be detected by using a region growing algorithm, wherein the distance between the pixel point coordinates of the positions of the seed points and the pixel point coordinates in the position information is smaller than or equal to a preset distance, and each MRI image to be detected is an MRI image corresponding to each initial grey bone marrow image;
and when the error between the extracted area of the segmented image and the area of the corresponding initial bone marrow grey matter image is smaller than or equal to a preset error, determining that the MRI image to be detected corresponding to the segmented image has a bone marrow grey matter region.
2. The method according to claim 1, wherein after determining that the MRI image to be measured corresponding to the segmented image has a gray matter region when an error between the extracted area of the segmented image and the area of the corresponding initial gray matter image is smaller than or equal to a preset error, the method further comprises:
taking the area of the initial bone marrow gray matter image corresponding to each MRI image with the bone marrow gray matter area as a first area, and taking the area of the corresponding segmentation image as a second area;
and taking the average value of the first area and the corresponding second area as the final area of the polio area of the MRI image of the area with the polio.
3. The method according to claim 2, wherein the step of determining an average value of the first area and the corresponding second area as a final area of the gray matter region of the MRI image of the gray matter region further comprises:
acquiring the corresponding layer thickness of the MRI image of each area with the bone marrow gray matter;
calculating the product of each said layer thickness and the corresponding area of said final polio area;
the sum of the products is calculated and taken as the total volume of the gray matter of the bone marrow of the vertebra.
4. The method of claim 1, wherein the training process of the trained U-Net model comprises:
inputting the obtained MRI sample images of a preset number into a preset initial U-Net model to obtain various gray matter training images output after the MRI sample images are segmented by the initial U-Net model;
calculating a deviation of the gray matter training image from a gray matter standard image of the corresponding MRI sample image;
and updating the weight coefficient of the initial U-Net model by using a back propagation algorithm based on the deviation until the updated initial U-Net model meets a preset condition, and taking the finally updated initial U-Net model as the trained U-Net model.
5. The method for detecting gray matter of bone marrow according to claim 1, wherein the "selecting seed points from each MRI image to be measured according to the position information and extracting segmented images including the seed points from each MRI image to be measured by using a region growing algorithm" includes:
selecting initial pixel points which are the same as the position information from the MRI image to be detected, and taking the initial pixel points as seed points;
taking the seed points as growth starting points, and gradually merging adjacent pixel points meeting a preset growth rule into a current image area;
and extracting the current image area which is combined completely as the segmentation image.
6. A polio detection apparatus, comprising:
the initial image acquisition module is used for sequentially inputting each MRI image of the vertebra into a pre-trained U-Net model for bone marrow gray matter segmentation to acquire at least one initial bone marrow gray matter image;
the selection module is used for determining a pixel point from the selected initial gray bone matter image corresponding to the MRI image belonging to the middle part of the vertebra, and collecting the position information of the pixel point;
the region growing module is used for respectively selecting seed points from the MRI images to be detected according to the position information and respectively extracting segmentation images containing the seed points from the MRI images to be detected by using a region growing algorithm, wherein the distance between the pixel point coordinates of the positions of the seed points and the pixel point coordinates in the position information is smaller than or equal to a preset distance, and each MRI image to be detected is an MRI image corresponding to each initial bone marrow gray matter image;
and the bone marrow grey matter determining module is used for determining that the MRI image to be detected corresponding to the segmentation image has a bone marrow grey matter area when the error between the extracted area of the segmentation image and the area of the corresponding initial bone marrow grey matter image is less than or equal to a preset error.
7. The polio detection apparatus of claim 6, further comprising:
the area determination module is used for taking the area of the initial bone marrow gray matter image corresponding to each MRI image with the bone marrow gray matter area as a first area and taking the area of the corresponding segmentation image as a second area;
and the calculation module is used for taking the average value of the first area and the corresponding second area as the final area of the polio area of the MRI image of the area with the polio.
8. The polio detection apparatus of claim 7, further comprising:
the layer thickness information acquisition module is used for acquiring the layer thickness corresponding to the MRI image of each area with the bone marrow gray matter;
a layer volume calculation module for calculating the product of each layer thickness and the corresponding final polio area;
and the gray matter volume calculation module is used for calculating the sum of the products and taking the sum as the total volume of the bone marrow and the gray matter of the vertebra.
9. A bone marrow ash detection apparatus comprising a memory storing a computer program and a processor implementing the bone marrow ash detection method according to any one of claims 1 to 5 when the processor executes the computer program.
10. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, implements the gray bone detection method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910237817.3A CN109978861B (en) | 2019-03-27 | 2019-03-27 | Polio detection method, apparatus, device and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910237817.3A CN109978861B (en) | 2019-03-27 | 2019-03-27 | Polio detection method, apparatus, device and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109978861A CN109978861A (en) | 2019-07-05 |
| CN109978861B true CN109978861B (en) | 2021-03-26 |
Family
ID=67080929
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910237817.3A Active CN109978861B (en) | 2019-03-27 | 2019-03-27 | Polio detection method, apparatus, device and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109978861B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599505A (en) * | 2019-09-17 | 2019-12-20 | 上海微创医疗器械（集团）有限公司 | Organ image segmentation method and device, electronic equipment and storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102006041619A1 (en) * | 2005-09-07 | 2007-04-19 | Siemens Medical Solutions Usa, Inc. | Systems and methods for computer-assisted detection of spinal curvature using images and angle measurements |
| CN102727200A (en) * | 2011-03-31 | 2012-10-17 | 深圳迈瑞生物医疗电子股份有限公司 | Method and device for dividing spine centrum and intervertebral disk, and magnetic resonance imaging system |
| CN104346799A (en) * | 2013-08-01 | 2015-02-11 | 上海联影医疗科技有限公司 | Method for extracting spinal marrow in CT (Computed Tomography) image |
| CN104751178A (en) * | 2015-03-31 | 2015-07-01 | 上海理工大学 | Pulmonary nodule detection device and method based on shape template matching and combining classifier |
| CN105719276A (en) * | 2016-01-07 | 2016-06-29 | 于翠妮 | Liver parenchymal segmentation method based on CT images |
| CN107977971A (en) * | 2017-11-09 | 2018-05-01 | 哈尔滨理工大学 | The method of vertebra positioning based on convolutional neural networks |
| CN108040496A (en) * | 2015-06-01 | 2018-05-15 | 尤尼伐控股有限公司 | The computer implemented method of distance of the detection object away from imaging sensor |
| CN108197526A (en) * | 2017-11-23 | 2018-06-22 | 西安艾润物联网技术服务有限责任公司 | Detection method, system and computer readable storage medium |
| CN108537784A (en) * | 2018-03-30 | 2018-09-14 | 四川元匠科技有限公司 | A kind of CT figure pulmonary nodule detection methods based on deep learning |
| CN109447969A (en) * | 2018-10-29 | 2019-03-08 | 北京青燕祥云科技有限公司 | Hepatic space occupying lesion recognition methods, device and realization device |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101756710A (en) * | 2008-12-19 | 2010-06-30 | 曹淑兰 | Measuring method of volume of intracranial hematoma |
| CN109146885B (en) * | 2018-08-17 | 2021-08-17 | 深圳蓝胖子机器智能有限公司 | Image segmentation method, apparatus, and computer-readable storage medium |
| CN109409503B (en) * | 2018-09-27 | 2020-07-24 | 深圳市铱硙医疗科技有限公司 | Neural network training method, image conversion method, device, equipment and medium |
-
2019
- 2019-03-27 CN CN201910237817.3A patent/CN109978861B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102006041619A1 (en) * | 2005-09-07 | 2007-04-19 | Siemens Medical Solutions Usa, Inc. | Systems and methods for computer-assisted detection of spinal curvature using images and angle measurements |
| CN102727200A (en) * | 2011-03-31 | 2012-10-17 | 深圳迈瑞生物医疗电子股份有限公司 | Method and device for dividing spine centrum and intervertebral disk, and magnetic resonance imaging system |
| CN104346799A (en) * | 2013-08-01 | 2015-02-11 | 上海联影医疗科技有限公司 | Method for extracting spinal marrow in CT (Computed Tomography) image |
| CN104751178A (en) * | 2015-03-31 | 2015-07-01 | 上海理工大学 | Pulmonary nodule detection device and method based on shape template matching and combining classifier |
| CN108040496A (en) * | 2015-06-01 | 2018-05-15 | 尤尼伐控股有限公司 | The computer implemented method of distance of the detection object away from imaging sensor |
| CN105719276A (en) * | 2016-01-07 | 2016-06-29 | 于翠妮 | Liver parenchymal segmentation method based on CT images |
| CN107977971A (en) * | 2017-11-09 | 2018-05-01 | 哈尔滨理工大学 | The method of vertebra positioning based on convolutional neural networks |
| CN108197526A (en) * | 2017-11-23 | 2018-06-22 | 西安艾润物联网技术服务有限责任公司 | Detection method, system and computer readable storage medium |
| CN108537784A (en) * | 2018-03-30 | 2018-09-14 | 四川元匠科技有限公司 | A kind of CT figure pulmonary nodule detection methods based on deep learning |
| CN109447969A (en) * | 2018-10-29 | 2019-03-08 | 北京青燕祥云科技有限公司 | Hepatic space occupying lesion recognition methods, device and realization device |
Non-Patent Citations (2)
| Title |
|---|
| Learning-Based Vertebra Detection and Iterative Normalized-Cut Segmentation for Spinal MRI;Szu-Hao Huang等;《IEEE Transactions on Medical Imaging》;20090526;第1595-1605页 * |
| 脊髓瘢痕分割算法研究;解威;《中国优秀硕士学位论文全文数据库 信息科技辑》;20160815;第I138-1246页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109978861A (en) | 2019-07-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3879485B1 (en) | Tissue nodule detection and model training method and apparatus thereof, device and system | |
| CN108765369B (en) | Method, apparatus, computer device and storage medium for detecting lung nodule | |
| CN111539944B (en) | Method, device, electronic equipment and storage medium for acquiring statistical attribute of lung focus | |
| CN112184617B (en) | Spine MRI image key point detection method based on deep learning | |
| CN112734757B (en) | Spine X-ray image cobb angle measuring method | |
| CN103249358B (en) | Medical image-processing apparatus | |
| CN111047572A (en) | Automatic spine positioning method in medical image based on Mask RCNN | |
| CN112102237A (en) | Brain tumor recognition model training method and device based on semi-supervised learning | |
| CN110570407B (en) | Image processing method, storage medium, and computer device | |
| CN108062749B (en) | Identification method and device for levator ani fissure hole and electronic equipment | |
| CN111932492B (en) | Medical image processing method and device and computer readable storage medium | |
| CN110738643B (en) | Analysis method for cerebral hemorrhage, computer device and storage medium | |
| CN107680134B (en) | Spine calibration method, device and equipment in medical image | |
| CN113989407B (en) | Training method and system for limb part recognition model in CT image | |
| CN114972255B (en) | Image detection method and device for cerebral micro-bleeding, computer equipment and storage medium | |
| CN115311309A (en) | Method and system for identifying and extracting focus of nuclear magnetic resonance image | |
| CN114332132A (en) | Image segmentation method and device and computer equipment | |
| CN111524188A (en) | Lumbar positioning point acquisition method, equipment and medium | |
| CN109978861B (en) | Polio detection method, apparatus, device and computer readable storage medium | |
| CN110992312B (en) | Medical image processing method, medical image processing device, storage medium and computer equipment | |
| US20220301177A1 (en) | Updating boundary segmentations | |
| CN115249279A (en) | Medical image processing method, medical image processing device, computer equipment and storage medium | |
| CN112184623A (en) | Intervertebral space analysis method, equipment and storage medium for vertebral bodies of spine | |
| CN112529918B (en) | Method, device and equipment for segmenting brain room area in brain CT image | |
| CN115439453A (en) | Vertebral body positioning method and device, electronic equipment 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 |