CN116977351A - Interactive hematoma segmentation and analysis method and system based on brain CT image - Google Patents
Interactive hematoma segmentation and analysis method and system based on brain CT image Download PDFInfo
- Publication number
- CN116977351A CN116977351A CN202310928052.4A CN202310928052A CN116977351A CN 116977351 A CN116977351 A CN 116977351A CN 202310928052 A CN202310928052 A CN 202310928052A CN 116977351 A CN116977351 A CN 116977351A
- Authority
- CN
- China
- Prior art keywords
- hematoma
- segmentation
- area
- dimensional
- brain
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 206010018852 Haematoma Diseases 0.000 title claims abstract description 269
- 230000011218 segmentation Effects 0.000 title claims abstract description 136
- 210000004556 brain Anatomy 0.000 title claims abstract description 56
- 238000004458 analytical method Methods 0.000 title claims abstract description 46
- 230000002452 interceptive effect Effects 0.000 title claims abstract description 26
- 238000000034 method Methods 0.000 claims abstract description 38
- 238000001914 filtration Methods 0.000 claims abstract description 19
- 238000004364 calculation method Methods 0.000 claims abstract description 10
- 230000003993 interaction Effects 0.000 claims abstract description 10
- 238000010276 construction Methods 0.000 claims description 3
- 230000010354 integration Effects 0.000 abstract description 4
- 238000000513 principal component analysis Methods 0.000 description 10
- 238000001514 detection method Methods 0.000 description 6
- 238000013135 deep learning Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 210000001519 tissue Anatomy 0.000 description 4
- 206010053942 Cerebral haematoma Diseases 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000003780 insertion Methods 0.000 description 2
- 230000037431 insertion Effects 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 235000002566 Capsicum Nutrition 0.000 description 1
- 208000018152 Cerebral disease Diseases 0.000 description 1
- 206010008111 Cerebral haemorrhage Diseases 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 239000006002 Pepper Substances 0.000 description 1
- 235000016761 Piper aduncum Nutrition 0.000 description 1
- 235000017804 Piper guineense Nutrition 0.000 description 1
- 244000203593 Piper nigrum Species 0.000 description 1
- 235000008184 Piper nigrum Nutrition 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 206010008118 cerebral infarction Diseases 0.000 description 1
- 208000026106 cerebrovascular disease Diseases 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 208000024200 hematopoietic and lymphoid system neoplasm Diseases 0.000 description 1
- 208000003906 hydrocephalus Diseases 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000012958 reprocessing Methods 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 210000003625 skull Anatomy 0.000 description 1
- 210000004872 soft tissue Anatomy 0.000 description 1
Classifications
-
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
- A61B6/5211—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20092—Interactive image processing based on input by user
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Radiology & Medical Imaging (AREA)
- Geometry (AREA)
- Optics & Photonics (AREA)
- Heart & Thoracic Surgery (AREA)
- Biophysics (AREA)
- High Energy & Nuclear Physics (AREA)
- Epidemiology (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Primary Health Care (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The application discloses an interactive hematoma segmentation and analysis method and system based on a brain CT image, wherein the method comprises the following steps: acquiring a brain medical image, and setting a window width and a window level for the medical image; determining a first hematoma area according to the human-computer interaction frame selection area; performing three-dimensional median filtering; performing three-dimensional Ojin threshold segmentation; performing three-dimensional open operation; selecting the largest three-dimensional connected domain of the binary segmentation image to obtain a segmentation result of a second hematoma region; acquiring hematoma centroid and hematoma volume; and the long axis direction of hematoma was calculated using PCA. The method occupies small memory, is easy to integrate into software and deploy on various devices, provides relatively perfect hematoma information, effectively improves calculation speed and hematoma segmentation accuracy, can assist doctors or surgical robots in brain hematoma positioning analysis and puncture path planning, and has the characteristics of light weight, easiness in integration, easiness in operation, information diversity, accuracy and rapidity.
Description
Technical Field
The application relates to the technical field of brain hematoma detection and segmentation, in particular to an interactive hematoma segmentation and analysis method and system based on brain CT images.
Background
Brain hematoma detection and segmentation is an important medical image processing task aimed at automatically or semi-automatically identifying and segmenting brain hematoma areas.
At present, most of traditional hematoma detection segmentation techniques are based on a deep learning method, larger model files are needed to be loaded, the requirement of light weight is not met, at least one model file is needed to be used for calculation reasoning in the hematoma detection segmentation method based on deep learning, but the model file occupies too much memory, and the method is not suitable for light weight deployment of brain hematoma segmentation and puncture analysis software; deep learning based methods also require configuration of the virtual environment at deployment time, which also adds unnecessary effort to software deployment. In addition, many steps are required when using the conventional medical image processing software 3D sler for hematoma segmentation. In 3D slicers, it is first necessary to delineate the hematoma interior region in three views, respectively; then, respectively outlining the external areas of the hematoma in the three views; finally, the segmentation is carried out to obtain a result. The above-mentioned hematoma segmentation operation is too troublesome and not easy to handle.
In addition, in the conventional technology, the hematoma is segmented and the volume is calculated, and other auxiliary information such as the mass center of the hematoma, the concentrated direction of the hematoma and the like are not provided in a displaying way. If a doctor or a surgical robot needs to perform operations such as positioning hematoma, the hematoma segmentation result needs to be reprocessed. This approach can increase the complexity of the operation.
More importantly, the traditional technology basically focuses on detection and segmentation of cerebral hematoma only, for example, in the traditional technology, a deep learning method is utilized to segment and calculate the volume of hematoma, and no possible guidance is provided for a needle insertion route according to the existing brain region image; in clinic, when cerebral diseases such as tumor, hydrocephalus, cerebral hematoma and the like are encountered, brain puncture operation is needed to relieve the illness or treat the disease; if a doctor wants to comprehensively utilize the two methods, the result of image analysis needs to be exported to a puncture path planning algorithm, so that the operation is time-consuming and the requirement of the operation on timeliness cannot be met;
it is therefore a need for a skilled artisan to devise an integrated, lightweight, easy to operate, interactive hematoma segmentation and analysis method and system.
Disclosure of Invention
In view of the above, the application provides an interactive hematoma segmentation and analysis method and system based on brain CT images, which can effectively assist in brain hematoma segmentation and analysis and puncture path planning while meeting the requirement of light weight.
In order to achieve the above purpose, the present application adopts the following technical scheme:
in a first aspect, the present application provides an interactive hematoma segmentation and analysis method based on brain CT images, comprising the specific steps of:
s1, acquiring a brain medical image, and setting a window width and a window position for the medical image;
s2, determining a first hematoma area according to the human-computer interaction frame selection area;
s3, performing three-dimensional median filtering on the first hematoma area;
s4, performing three-dimensional Ojin threshold segmentation on the first hematoma region after three-dimensional median filtering to obtain a binary segmentation image;
s5, performing three-dimensional open operation on the binary segmentation image;
s6, selecting the largest three-dimensional connected domain of the binary segmentation image from the three-dimensional open operation result to obtain a segmentation result of a second hematoma region;
s7, acquiring a hematoma centroid and a hematoma volume according to the segmentation result of the second hematoma region; and the long axis direction of hematoma was calculated using PCA.
Preferably, the method further comprises:
s8, outputting a ray based on the mass center of hematoma and the long axis direction, wherein the ray is a recommended puncture direction;
preferably, the step S2 includes:
s21, identifying a cuboid region of the envelope hematoma drawn in the first view;
s22, determining a first hematoma area according to the cuboid area and generating cuboid projections in a second view and three views;
and S23, when the cuboid area does not completely cover the hematoma area, fine adjustment is carried out on the cuboid area, so that the cuboid area is ensured to completely cover the hematoma area.
Preferably, the step S4 includes:
s41, performing binary segmentation on the first hematoma area according to a three-dimensional Ojin threshold method, wherein the binary segmentation comprises the following steps: calculating a segmentation threshold value by using the maximized inter-class variance and segmenting the first hematoma region;
s42, adjusting the window width and the window level, and analyzing the first hematoma area by utilizing the segmentation threshold value based on the difference of gray values.
Preferably, the step S7 of obtaining a hematoma centroid includes:
s701, constructing a dense three-dimensional point cloud by utilizing a hematoma region based on the segmentation result of the second hematoma region, and acquiring a pixel coordinate value of each point of the hematoma region;
s702, respectively calculating average values of pixel values of three dimensions in the three-dimensional point cloud to obtain hematoma centroid.
Preferably, the step S7 of acquiring a hematoma volume includes:
s711, based on the segmentation result of the second hematoma area, counting the number of voxels occupied by the second hematoma area;
s712, acquiring the length, width and height real data corresponding to each voxel in the medical image file, and calculating the real volume of each voxel;
s713, multiplying the number of the voxels and the real volume of the voxels to obtain the hematoma volume.
In a second aspect, the present application further provides an interactive hematoma segmentation and analysis system based on brain CT images, specifically including:
the acquisition module is used for acquiring brain medical images and setting window width and window positions for the medical images;
the determining module is used for determining a first hematoma area according to the human-computer interaction frame selection area;
the filtering module is used for carrying out three-dimensional median filtering on the first hematoma area;
the segmentation module is used for carrying out three-dimensional Ojin threshold segmentation on the first hematoma region after three-dimensional median filtering to obtain a binary segmentation image;
the operation module is used for carrying out three-dimensional open operation on the binary segmentation image;
the selecting module is used for selecting the maximum three-dimensional connected domain of the binary segmentation image from the three-dimensional open operation result to obtain a segmentation result of the second hematoma region;
the analysis module is used for acquiring a hematoma centroid and a hematoma volume according to the segmentation result of the second hematoma region; and the long axis direction of hematoma was calculated using PCA.
Preferably, the system further comprises:
the output module is used for outputting a ray based on the mass center of hematoma and the long axis direction, wherein the ray is a recommended puncture direction;
preferably, the determining module includes:
an identification unit for identifying a rectangular parallelepiped region of the envelope hematoma drawn in the first view;
a generation unit that determines a first hematoma area and generates a cuboid projection in a second view and three views, according to the cuboid area;
and the adjusting unit is used for judging that when the cuboid area does not completely cover the hematoma area, the cuboid area is finely adjusted, so that the cuboid area is ensured to completely cover the hematoma area.
Preferably, the dividing module includes:
the first calculating unit is configured to perform binary segmentation on the first hematoma area according to a three-dimensional oxford thresholding method, and includes: calculating a segmentation threshold value by using the maximized inter-class variance and segmenting the first hematoma region;
and the analysis unit is used for adjusting the window width and the window level, and analyzing the first hematoma area by utilizing the segmentation threshold value based on the difference of the gray values.
Preferably, the analysis module obtains a hematoma centroid, including:
a construction unit, which is used for constructing a dense three-dimensional point cloud by utilizing the hematoma region based on the segmentation result of the second hematoma region and acquiring the pixel coordinate value of each point of the hematoma region;
and the second calculation unit calculates average values of the pixel values of the three dimensions in the three-dimensional point cloud to obtain hematoma centroid.
Preferably, the analysis module acquires a hematoma volume, including:
a statistics unit for counting the number of voxels occupied by the second hematoma region based on the segmentation result of the second hematoma region;
the acquisition unit is used for acquiring the length, width and height real data corresponding to each voxel in the medical image file and calculating the real volume of each voxel;
and the operation unit is used for carrying out multiplication operation by using the number of the voxels and the real volume of the voxels to obtain the hematoma volume.
Compared with the prior art, the application has the following beneficial effects:
1. the method overcomes the defect that the traditional deep learning-based model file needs to be loaded to occupy excessive memory, and meanwhile, does not need to be configured with a virtual environment, so that the method is suitable for light deployment of brain hematoma segmentation and puncture analysis;
2. the application can provide other auxiliary information such as mass center of hematoma, direction of hematoma and the like; redundant operation steps for reprocessing hematoma segmentation results are simplified; the brain hematoma detection and segmentation are combined with needle insertion route recommendation guidance, so that a doctor or an operation robot can be effectively assisted in brain hematoma positioning analysis and puncture path planning.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present application, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of an interactive hematoma segmentation and analysis method based on brain CT images provided by the application;
FIG. 2 is a schematic view of medical images introduced into the brain according to the present application;
FIG. 3 is a schematic view of a window width and level setting according to the present application;
FIG. 4 is a schematic diagram of a human-computer interaction frame for selecting hematoma areas provided by the application;
fig. 5 is a schematic diagram showing the completion of hematoma segmentation and the recommended puncture path according to the present application.
FIG. 6 is a block diagram of an interactive hematoma segmentation and analysis system based on brain CT images provided by the application;
Detailed Description
The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
Example 1:
as shown in fig. 1, an interactive hematoma segmentation and analysis method based on brain CT images specifically includes the following steps:
s1, acquiring a brain medical image, and setting a window width and a window position for the medical image;
s2, determining a first hematoma area according to the human-computer interaction frame selection area;
s3, performing three-dimensional median filtering on the first hematoma area;
s4, performing three-dimensional Ojin threshold segmentation on the first hematoma region after three-dimensional median filtering to obtain a binary segmentation image;
s5, performing three-dimensional open operation on the binary segmentation image;
s6, selecting the largest three-dimensional connected domain of the binary segmentation image from the three-dimensional open operation result to obtain a segmentation result of a second hematoma region;
s7, acquiring a hematoma centroid and a hematoma volume according to the segmentation result of the second hematoma region; and the long axis direction of hematoma was calculated using PCA.
In the embodiment, the method occupies smaller memory, is easy to integrate into software and be deployed on various devices, provides more perfect hematoma information, and effectively improves the calculation speed and the hematoma segmentation accuracy; in addition, based on the mass center and the long axis direction of hematoma, a ray can be output, and the ray is a recommended puncture direction; the brain hematoma positioning analysis and puncture path planning device can assist doctors or surgical robots to carry out brain hematoma positioning analysis and puncture path planning, and has the characteristics of light weight, easiness in integration, easiness in operation, information diversity, accuracy and rapidity.
The following describes each of the above steps in detail:
in step S1, as shown in fig. 2, a brain medical image to be analyzed is acquired, and input of nii files and dcm files is supported; after the file is imported, the medical image is visualized from three angles, namely, coronal, sagittal and axial.
As shown in fig. 3, the window width is set for better display, and the window width for brain has been preset for brain medical images, and for CT effect, the window width for observing brain region is usually selected to be 100,50. That is, a voxel value having a HU value between [0,100] is mapped to a gray scale of [0,255], a voxel value having a HU value <0 is mapped to a gray scale value=0, and a voxel value having a HU value >100 is mapped to a gray scale value=255. If the user has special requirements, the window width and window level parameters can be set by himself, for example, a certain user is interested in observing cerebral hemorrhage and needs to observe whether cerebral infarction occurs or not, and at the moment, the window level of the HU value can be reduced from 100HU to 60HU so as to be observed better.
In step S2, as shown in fig. 4, the method includes:
s21, acquiring a cuboid region of the envelope hematoma drawn in the first view;
s22, determining a first hematoma area according to the cuboid area and generating cuboid projections in a second view and three views;
and S23, when the cuboid area does not completely cover the hematoma area, fine adjustment is carried out on the cuboid area, so that the cuboid area is ensured to completely cover the hematoma area.
This step requires interaction with the user, who needs to find the slice with the largest hematoma area approximately in the cross section, and select a rectangle on the slice that can cover the hematoma area by means of a mouse frame.
After the user completes the rectangular drawing in the first view, the algorithm participates in the interactive determination of the blood tumor approximate area and generates cuboid projections in the second and third views. If the selected cuboid does not cover the hematoma area completely, fine adjustments can be made in the three views to ensure that the cuboid covers the hematoma area completely. In the step, expert priori knowledge is introduced before hematoma segmentation, the region of interest is reduced from the whole image to a small cuboid region, and the effect of improving the precision and the speed can be achieved by less human intervention. The method can effectively reduce the calculation power consumption, improve the calculation speed, improve the accuracy of hematoma segmentation, and is less prone to segmenting contents such as skull, noise and other tissues.
Machine interaction in this step: when the rectangle is drawn, the hematoma on the slice is segmented by first using (two-dimensional) Ojin threshold segmentation for the region. Next, a two-dimensional center of the hematoma is approximately calculated from the divided mask, and an approximately three-dimensional center of the hematoma can be obtained by combining the coordinates of the slice. Further, the center of the rectangle is moved to the approximate center of the hematoma. Then, the rectangle is expanded in the axial direction by default with the short side (width) of the planar rectangle as the length, so that the rectangle becomes a cuboid. Finally, on the hematoma axial surface slice containing the approximate three-dimensional center, carrying out (two-dimensional) Ojin threshold segmentation on hematoma in a rectangle on the slice, and if the segmented hematoma area has a boundary contacting the rectangle, expanding the size of the rectangle in the direction by 1.5 times. The intelligent adjustment aims to obtain a cuboid which completely covers the hematoma area through one-time frame selection.
After intelligent adjustment, the user can observe the cuboid in slices of all directions, ensuring that it can encompass all hematoma areas. If not covered, manual adjustments may be made on the corresponding slices. After manual fine tuning, a cuboid is obtained which covers the hematoma area in all three views.
After the above steps S21 to S23 are completed, two diagonal corner coordinates (x 1 ,y 1 ,z 1 ,x 2 ,y 2 ,z 2 ) Input into the algorithm. The algorithm will intercept the cuboid voxel block from the complete brain medical image and perform a fully automatic analysis in this region.
In the step S3 of the process, the filter size selected is 3 x 3 voxels. The median filter is a nonlinear digital filter that can effectively filter out salt and pepper noise on the basis of preserving the main information of the image. Median filtering traverses over the image using a filter kernel of a given size. Within the filter kernel, all pixels are ordered, the median of all pixels is taken and assigned to the center pixel of the filter kernel. And outputting the filtered image after the traversing is completed.
In step S4, it includes:
s41, performing binary segmentation on the first hematoma area according to a three-dimensional Ojin threshold method, wherein the binary segmentation comprises the following steps: calculating a segmentation threshold value by using the maximized inter-class variance and segmenting the first hematoma region;
s42, adjusting the window width and the window level, and analyzing the first hematoma area by utilizing the segmentation threshold value based on the difference of gray values.
After the median filtering is completed, a three-dimensional Ojin threshold method is selected to carry out binary segmentation on the first hematoma region. The Ojin thresholding method does not require manual input of a threshold, and automatically computes a segmentation threshold and segments selected regions using the principle of maximizing inter-class variance. After adjusting the window width level for the brain region, the gray level of the hematoma is significantly higher than the surrounding soft tissue area, so the area above the threshold may be the hematoma area.
The method calculates the threshold value of the image in a self-adaptive way by maximizing the inter-class variance, and carries out binary segmentation on the image by using the threshold value. The three-dimensional Ojin threshold segmentation method calculates the inter-class variance corresponding to each threshold in a given three-dimensional voxel region, and takes the threshold corresponding to the maximum inter-class variance as the final segmentation threshold. The calculation formula of the inter-class variance is as follows:
wherein,,inter-class variance for two classes of voxel values that are thresholded; is w 0 And w 1 The number of voxels of the first type and the second type respectively; mu (mu) 0 、μ 1 Sum mu T Is the mean of the first class, the second class, and the population of voxels. For example, after converting the three-dimensional image to a gray value of 0-255 in this embodiment, the threshold is traversed in the gray interval and the inter-class variance of each threshold is calculated. And dividing the image by applying a threshold value corresponding to the maximum inter-class variance, and outputting a binary image after division.
In step S5, after the division of the oxford threshold, there are cases where there are many noise points and minute connections in the divided binary image, the kernel size selected is 5 x 5 voxels, by which the imperfections of the binary image are eliminated and the connection resulting from the small noise is broken.
The open operation belongs to morphological operation, and the expansion operation is performed after the corrosion operation is performed on the image. The method can remove isolated noise, burrs and tiny connection in the image under the condition that main information of the image is unchanged.
In step S6, small connected regions may exist in the binary image, and these small connected regions may be the segmentation result of other tissues or noise. Since the rough hematoma region (first hematoma region) has been framed, the hematoma region can be considered as the largest connected region in the region division result. Therefore, after the maximum connected domain is obtained, the hematoma area is reserved, and the segmentation results of other tissues or noise are removed.
After the open operation, the connected domain in the binary image mainly comprises a first hematoma area, other tissue areas and a large block noise area. After the hematoma area is selected by the frame, the largest communicating area in the cuboid is the first hematoma area. Therefore, firstly, calculating the number of voxels occupied by each connected domain in the cuboid by using a simpleITK library, obtaining the connected domain with the most voxels through sequencing, and removing the rest connected domains. And outputting a binary image of the maximum connected domain, namely a second hematoma area. Segmentation of hematoma is thus completed.
Acquiring a hematoma centroid in step S7 includes:
s701, constructing a dense three-dimensional point cloud based on the hematoma region based on the segmentation result of the second hematoma region;
s702, acquiring pixel coordinate values of each point of the hematoma area;
s703, respectively calculating average values of the pixel values of the three dimensions to obtain hematoma mass centers.
The density of hematoma areas is uniform. Therefore, for the hematoma three-dimensional point cloud, the pixel values of three dimensions are respectively averaged, and the mass center of the hematoma can be obtained.
Acquiring a hematoma volume in step S7, comprising:
s711, based on the segmentation result of the second hematoma area, counting the number of voxels occupied by the second hematoma area;
s712, acquiring the length, width and height real data corresponding to each voxel in the medical image file, and calculating the real volume of each voxel;
s713, multiplying the number of the voxels and the real volume of the voxels to obtain the hematoma volume.
Step S7 further includes calculating a major axis direction of hematoma using PCA;
for a three-dimensional point cloud of hematoma, the long axis refers to the direction or axis of maximum variance of the point cloud, which intuitively is the direction in which the point cloud extends longest. Taking a two-dimensional rectangle as an example, the major axis direction is the long axis direction. After the hematoma segmentation result is obtained, the major axis direction of the hematoma is calculated by PCA (principal component analysis). According to the method, hematoma in the binary image is regarded as three-dimensional point cloud, and the direction in which the hematoma extends to the longest is rapidly and efficiently calculated by using a statistical analysis method PCA. If the "rotating carpule method" is used to calculate the smallest circumscribed cuboid of hematoma, the time consumption is greatly increased.
PCA (principal component analysis) is a statistical method that can reduce the dimension of data and output principal component directions orthogonal to each other. The method treats the hematoma as a three-dimensional point cloud, and the three-dimensional coordinates of each voxel point are regarded as the characteristics of the voxel. Assuming that there are N voxels in the hematoma region, a dimension of [ N,3 can be obtained]Is a sample matrix X of (a); dividing the sample matrix minus the mean value by the standard deviation, and performing standardization treatment to obtain a standardized sample matrixCalculate->Covariance matrix of (a) and eigenvalues and eigenvectors of the covariance matrix; and selecting a feature vector corresponding to the maximum feature value, wherein the feature vector is the long axis direction of the hematoma. The cloud of hematoma points has the greatest variance in the long axis direction, and the hematoma extends the longest in this direction as seen visually.
As shown in fig. 5; based on the mass center and the long axis direction of hematoma, a ray is output, and the ray is the recommended puncture direction, so that a suggestion can be provided for puncture operation planning. Combining the mass center of hematoma and the long axis direction to obtain the recommended puncture path. This path can refer directly to the mass center of the hematoma along the long axis direction of the hematoma, and the hematoma area is more easily reached.
Example 2:
based on the same inventive concept, the embodiment of the application also provides an interactive hematoma segmentation and analysis system based on the brain CT image, and the principle of the system for solving the problem is similar to that of an interactive hematoma segmentation and analysis method based on the brain CT image, so that the implementation of the system can be referred to the implementation of the method, and the repetition is omitted.
The embodiment of the application provides an interactive hematoma segmentation and analysis system based on brain CT images, as shown in fig. 6, which specifically comprises the following steps:
the application also provides an interactive hematoma segmentation and analysis system based on the brain CT image, which specifically comprises the following steps:
the acquisition module is used for acquiring brain medical images and setting window width and window positions for the medical images;
the determining module is used for determining a first hematoma area according to the human-computer interaction frame selection area;
the filtering module is used for carrying out three-dimensional median filtering on the first hematoma area;
the segmentation module is used for carrying out three-dimensional Ojin threshold segmentation on the first hematoma region after three-dimensional median filtering to obtain a binary segmentation image;
the operation module is used for carrying out three-dimensional open operation on the binary segmentation image;
the selecting module is used for selecting the maximum three-dimensional connected domain of the binary segmentation image from the three-dimensional open operation result to obtain a segmentation result of the second hematoma region;
the analysis module is used for acquiring a hematoma centroid and a hematoma volume according to the segmentation result of the second hematoma region; and the long axis direction of hematoma was calculated using PCA.
In one embodiment, the determining module includes:
an identification unit for identifying a rectangular parallelepiped region of the envelope hematoma drawn in the first view;
a generation unit that determines a first hematoma area and generates a cuboid projection in a second view and three views, according to the cuboid area;
and the adjusting unit is used for judging that when the cuboid area does not completely cover the hematoma area, the cuboid area is finely adjusted, so that the cuboid area is ensured to completely cover the hematoma area.
In one embodiment, the segmentation module includes:
the first calculating unit is configured to perform binary segmentation on the first hematoma area according to a three-dimensional oxford thresholding method, and includes: calculating a segmentation threshold value by using the maximized inter-class variance and segmenting the first hematoma region;
and the analysis unit is used for adjusting the window width and the window level, and analyzing the first hematoma area by utilizing the segmentation threshold value based on the difference of the gray values.
In one embodiment, the analysis module obtains a hematoma centroid comprising:
a construction unit, which is used for constructing a dense three-dimensional point cloud by utilizing the hematoma region based on the segmentation result of the second hematoma region and acquiring the pixel coordinate value of each point of the hematoma region;
and the second calculation unit calculates average values of the pixel values of the three dimensions in the three-dimensional point cloud to obtain hematoma centroid.
In one embodiment, the analysis module obtains a hematoma volume comprising:
a statistics unit for counting the number of voxels occupied by the second hematoma region based on the segmentation result of the second hematoma region;
the acquisition unit is used for acquiring the length, width and height real data corresponding to each voxel in the medical image file and calculating the real volume of each voxel;
and the operation unit is used for carrying out multiplication operation by using the number of the voxels and the real volume of the voxels to obtain the hematoma volume.
Compared with the traditional brain CT image processing, in the above embodiment, the general areas of the hematoma selected by the brain medical image and the human are taken as input, and after a series of traditional image processing operations (including median filtering, three-dimensional Ojin threshold segmentation and three-dimensional open operation) are performed, auxiliary information such as the segmentation result of the hematoma areas, the position of the mass center of the hematoma, the direction of the main axis of the hematoma (recommended puncture direction), the volume of the hematoma and the like is output. The brain hematoma positioning analysis and puncture path planning can be effectively assisted by a doctor or a surgical robot, and the brain hematoma positioning analysis and puncture path planning device has the characteristics of light weight, easiness in integration, easiness in operation, information diversity, accuracy and rapidity.
Lightweight and easy integration: the application does not need to attach model files with more occupied memory, and the occupied memory is smaller and lighter and is easy to integrate into software and be deployed on various devices.
Easy operability: for a brain medical image, the application can perform quick and accurate hematoma segmentation only by a few steps of intuitive frame selection operation.
Information diversity: for cerebral hematoma puncture operation, the application can provide perfect hematoma information.
Accuracy and rapidity: using the human frame selected hematoma area as a priori, only focusing on the cuboid area containing hematoma; the calculation speed and hematoma segmentation accuracy can be effectively improved through the operation.
In this specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, so that the same or similar parts between the embodiments are referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. An interactive hematoma segmentation and analysis method based on brain CT images is characterized by comprising the following specific steps:
s1, acquiring a brain medical image, and setting a window width and a window position for the medical image;
s2, determining a first hematoma area according to the human-computer interaction frame selection area;
s3, performing three-dimensional median filtering on the first hematoma area;
s4, performing three-dimensional Ojin threshold segmentation on the first hematoma region after three-dimensional median filtering to obtain a binary segmentation image;
s5, performing three-dimensional open operation on the binary segmentation image;
s6, selecting the largest three-dimensional connected domain of the binary segmentation image from the three-dimensional open operation result to obtain a segmentation result of a second hematoma region;
s7, acquiring a hematoma centroid and a hematoma volume according to the segmentation result of the second hematoma region; and the long axis direction of hematoma was calculated using PCA.
2. The method for interactive hematoma segmentation and analysis based on brain CT images according to claim 1, wherein said step S2 comprises:
s21, identifying a cuboid region of the envelope hematoma drawn in the first view;
s22, determining a first hematoma area according to the cuboid area and generating cuboid projections in a second view and three views;
and S23, when the cuboid area does not completely cover the hematoma area, fine adjustment is carried out on the cuboid area, so that the cuboid area is ensured to completely cover the hematoma area.
3. The method for interactive hematoma segmentation and analysis based on brain CT images according to claim 1, wherein said step S4 comprises:
s41, performing binary segmentation on the first hematoma area according to a three-dimensional Ojin threshold method, wherein the binary segmentation comprises the following steps: calculating a segmentation threshold value by using the maximized inter-class variance and segmenting the first hematoma region;
s42, adjusting the window width and the window level, and analyzing the first hematoma area by utilizing the segmentation threshold value based on the difference of gray values.
4. The method for interactive hematoma segmentation and analysis based on brain CT image according to claim 1, wherein the step S7 of obtaining the mass center of hematoma comprises:
s701, constructing a dense three-dimensional point cloud by utilizing a hematoma region based on the segmentation result of the second hematoma region, and acquiring a pixel coordinate value of each point of the hematoma region;
s702, respectively calculating average values of pixel values of three dimensions in the three-dimensional point cloud to obtain hematoma centroid.
5. The method for interactive hematoma segmentation and analysis based on brain CT images according to claim 1, wherein the step S7 of acquiring the hematoma volume comprises:
s711, based on the segmentation result of the second hematoma area, counting the number of voxels occupied by the second hematoma area;
s712, acquiring the length, width and height real data corresponding to each voxel in the medical image file, and calculating the real volume of each voxel;
s713, multiplying the number of the voxels and the real volume of the voxels to obtain the hematoma volume.
6. An interactive hematoma segmentation and analysis system based on brain CT images is characterized by comprising the following specific components:
the acquisition module is used for acquiring brain medical images and setting window width and window positions for the medical images;
the determining module is used for determining a first hematoma area according to the human-computer interaction frame selection area;
the filtering module is used for carrying out three-dimensional median filtering on the first hematoma area;
the segmentation module is used for carrying out three-dimensional Ojin threshold segmentation on the first hematoma region after three-dimensional median filtering to obtain a binary segmentation image;
the operation module is used for carrying out three-dimensional open operation on the binary segmentation image;
the selecting module is used for selecting the maximum three-dimensional connected domain of the binary segmentation image from the three-dimensional open operation result to obtain a segmentation result of the second hematoma region;
the analysis module is used for acquiring a hematoma centroid and a hematoma volume according to the segmentation result of the second hematoma region; and the long axis direction of hematoma was calculated using PCA.
7. An interactive hematoma segmentation and analysis system based on brain CT images according to claim 1 wherein the determination module comprises:
an identification unit for identifying a rectangular parallelepiped region of the envelope hematoma drawn in the first view;
a generation unit that determines a first hematoma area and generates a cuboid projection in a second view and three views, according to the cuboid area;
and the adjusting unit is used for judging that when the cuboid area does not completely cover the hematoma area, the cuboid area is finely adjusted, so that the cuboid area is ensured to completely cover the hematoma area.
8. An interactive hematoma segmentation and analysis system based on brain CT images according to claim 1 wherein the segmentation module comprises:
the first calculating unit is configured to perform binary segmentation on the first hematoma area according to a three-dimensional oxford thresholding method, and includes: calculating a segmentation threshold value by using the maximized inter-class variance and segmenting the first hematoma region;
and the analysis unit is used for adjusting the window width and the window level, and analyzing the first hematoma area by utilizing the segmentation threshold value based on the difference of the gray values.
9. The interactive hematoma segmentation and analysis system based on brain CT images of claim 1, wherein the analysis module obtains a hematoma centroid comprising:
a construction unit, which is used for constructing a dense three-dimensional point cloud by utilizing the hematoma region based on the segmentation result of the second hematoma region and acquiring the pixel coordinate value of each point of the hematoma region;
and the second calculation unit calculates average values of the pixel values of the three dimensions in the three-dimensional point cloud to obtain hematoma centroid.
10. The interactive hematoma segmentation and analysis system based on brain CT images of claim 1, wherein the analysis module acquires a hematoma volume comprising:
a statistics unit for counting the number of voxels occupied by the second hematoma region based on the segmentation result of the second hematoma region;
the acquisition unit is used for acquiring the length, width and height real data corresponding to each voxel in the medical image file and calculating the real volume of each voxel;
and the operation unit is used for carrying out multiplication operation by using the number of the voxels and the real volume of the voxels to obtain the hematoma volume.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310928052.4A CN116977351A (en) | 2023-07-26 | 2023-07-26 | Interactive hematoma segmentation and analysis method and system based on brain CT image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310928052.4A CN116977351A (en) | 2023-07-26 | 2023-07-26 | Interactive hematoma segmentation and analysis method and system based on brain CT image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116977351A true CN116977351A (en) | 2023-10-31 |
Family
ID=88472534
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310928052.4A Pending CN116977351A (en) | 2023-07-26 | 2023-07-26 | Interactive hematoma segmentation and analysis method and system based on brain CT image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116977351A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117976225A (en) * | 2024-03-05 | 2024-05-03 | 齐鲁工业大学(山东省科学院) | Method, system, storage medium and device for predicting hematoma change probability |
-
2023
- 2023-07-26 CN CN202310928052.4A patent/CN116977351A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117976225A (en) * | 2024-03-05 | 2024-05-03 | 齐鲁工业大学(山东省科学院) | Method, system, storage medium and device for predicting hematoma change probability |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108010021B (en) | Medical image processing system and method | |
| Gu et al. | Automated delineation of lung tumors from CT images using a single click ensemble segmentation approach | |
| Shattuck et al. | BrainSuite: an automated cortical surface identification tool | |
| Dubey et al. | Evaluation of three methods for MRI brain tumor segmentation | |
| CN109035261B (en) | Medical image processing method and device, electronic device and storage medium | |
| JP6539303B2 (en) | Transforming 3D objects to segment objects in 3D medical images | |
| US20110002523A1 (en) | Method and System of Segmenting CT Scan Data | |
| US20070031020A1 (en) | Method and apparatus for intracerebral hemorrhage lesion segmentation | |
| CN102419864B (en) | Method and device for extracting skeletons of brain CT (computerized tomography) image | |
| Balan et al. | Smart histogram analysis applied to the skull-stripping problem in T1-weighted MRI | |
| CN107481254A (en) | Processing method, device, medium and the electronic equipment of medical image | |
| CN115359257B (en) | Spine image segmentation method and operation navigation positioning system based on deep learning | |
| CN116977351A (en) | Interactive hematoma segmentation and analysis method and system based on brain CT image | |
| CN110428488A (en) | One kind being based on area of computer aided Hepatic CT diagnostic system and method | |
| AU2019430369A1 (en) | VRDS 4D medical image-based vein Ai endoscopic analysis method and product | |
| WO2007058631A1 (en) | Method and device for correction of magnetic resonance images | |
| CN105976394A (en) | Cutting method capable of realizing self-adaptive adjustment of fetus three-dimensional ultrasonic data | |
| WO2011050454A1 (en) | Bone imagery segmentation method and apparatus | |
| CN114365188A (en) | Analysis method and product based on VRDS AI inferior vena cava image | |
| CN111612762B (en) | MRI brain tumor image generation method and system | |
| Joshi et al. | A novel methodology for brain tumor detection based on two stage segmentation of MRI images | |
| EP4016470A1 (en) | 3d morhphological or anatomical landmark detection method and device using deep reinforcement learning | |
| WO2006121410A1 (en) | Method, apparatus and computer software for segmenting the brain from mr data | |
| AU2019430258B2 (en) | VRDS 4D medical image-based tumor and blood vessel ai processing method and product | |
| Bandyopadhyay | Pre-processing and segmentation of brain image for tumor detection |
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 |