Nasal root detection method and apparatus, storage medium, and electronic device

Abstract

The present disclosure relates to a nasal root detection method, device, storage medium and electronic device. The method comprises: acquiring an NCCT image sequence of a head of a to-be-detected object; determining a skull region according to the NCCT image sequence; determining a nasal root candidate region from the skull region based on a position feature of a nasal root; and determining the position of the nasal root from the nasal root candidate region according to a bone shape feature at the nasal root. The present disclosure is more accurate and safe in determining the position of the nasal root.

Description

Methods, devices, storage media and electronic equipment for nasal root detection Technical Field

[0001] This disclosure relates to the field of image processing technology, and more specifically, to a method, apparatus, storage medium, and electronic device for detecting the root of the nose. Background Technology

[0002] The nasal root is located at the upper end of the bridge of the nose between the two orbits. The point where the nasal root is located (referred to as the nasal root point) is situated on the upper part of the nose, at the intersection of the frontonasal suture and the midsagittal plane. The nasal root point is the point where the upper end of the nasal bridge connects to the forehead. Therefore, the nasal root is an important bony landmark. Because the nasal root is a crucial bony landmark of the face, and its plane forms the watershed between the face and the brain, it is an important reference point in clinical neurosurgery for calculating the anatomical positions of the head, such as the hairline, coronal suture, and Kocher's point. For example, in clinical surgery, a measuring tape is used to measure a certain distance from the nasal root towards the top of the head to determine the coronal suture. One specific implementation method is to determine the intersection of the coronal and sagittal sutures by pushing a curved distance of 13 cm from the nasal root towards the top of the head.

[0003] In related techniques, medical staff determine the location of the nasal root based on their experience, using visual observation combined with manual palpation, and then mark it with a marker. However, manually marking the location of the nasal root requires moving the patient's head, which is unsafe for the patient and can easily lead to secondary head injury. Furthermore, the pulling force from moving the patient's head can cause scalp displacement, resulting in errors in the marking of the nasal root location. Summary of the Invention

[0004] To address the problems existing in related technologies, this disclosure provides a method, apparatus, storage medium, and electronic device for detecting the root of the nose.

[0005] To achieve the above objectives, a first aspect of this disclosure provides a method for detecting the root of the nose, the method comprising:

[0006] Obtain the NCCT image sequence of the head of the object to be examined;

[0007] The skull region was determined based on the NCCT image sequence;

[0008] Candidate regions for the nasal root are determined from the skull region based on the location characteristics of the nasal root.

[0009] The location of the nasal root is determined from the candidate region of the nasal root based on the bone shape characteristics at the nasal root.

[0010] Optionally, determining the candidate nasal root region from the skull region based on the location features of the nasal root includes:

[0011] Determine the midsagittal and midcoronal planes corresponding to the skull region;

[0012] From the skull region, sagittal tomographic images of interest whose distance from the midsagittal plane is less than a preset threshold are determined, and a sequence of sagittal tomographic images of interest is obtained.

[0013] The candidate region of the nasal root is segmented from the bone region corresponding to the sagittal tomographic image sequence of interest based on the midline coronal plane.

[0014] Optionally, determining the location of the nasal root from the candidate nasal root region based on the bone shape characteristics at the nasal root includes:

[0015] Edge detection is performed on the candidate nasal root region to obtain the bone edge region, which represents at least one bone region in a sagittal tomographic image of the bone edge.

[0016] A filter is constructed based on the bone shape features at the root of the nose;

[0017] For each of the aforementioned sagittal tomographic images of the bone margin, candidate nasal root coordinates are determined from the sagittal tomographic image of the bone margin according to the filter;

[0018] The position of the nasal root is obtained by filtering from all the candidate nasal root coordinates based on the bone shape characteristics at the nasal root.

[0019] Optionally, the bone shape features at the nasal root include a first feature indicating that the angle between the first line connecting the nasal root and the forehead and the second line connecting the nasal root and the nasal dorsum is an obtuse angle, and the step of constructing a filter based on the bone shape features at the nasal root includes:

[0020] Based on the first feature, an isosceles obtuse triangle with an obtuse angle of a preset angle and a short side length of a preset length is constructed;

[0021] The filter is obtained by digitally converting the isosceles obtuse triangle.

[0022] Optionally, determining the candidate nasal root coordinates from the sagittal tomographic image of the bone margin based on the filter includes:

[0023] The sagittal tomographic image of the bone edge is subjected to matched filtering based on the filter to obtain candidate sagittal plane coordinates, which include sagittal axis coordinates and vertical axis coordinates.

[0024] The three-dimensional coordinates of the candidate nasal root are determined based on the candidate sagittal plane coordinates and the coronal axis coordinates corresponding to the sagittal tomographic image of the bone edge.

[0025] Optionally, the bone shape feature at the nasal root includes a second feature characterizing the location of the nasal root as a concave point, and the step of filtering the position of the nasal root from all the candidate nasal root coordinates based on the bone shape feature at the nasal root includes:

[0026] Clustering is performed on all the candidate nasal root coordinates to obtain the target nasal root coordinates;

[0027] For each target nasal root coordinate, if the target nasal root coordinate meets the second feature based on the positional features between the target nasal root coordinate and the neighboring bone coordinates of the target nasal root coordinate, the target nasal root coordinate is determined as the position of the nasal root.

[0028] Optionally, the step of clustering all the candidate nasal root coordinates to obtain the target nasal root coordinates includes:

[0029] All the candidate nasal root coordinates are clustered to obtain multiple clusters, each cluster including at least one of the candidate nasal root coordinates;

[0030] Based on the number of candidate nasal root coordinates in each cluster, the multiple clusters are sorted from most to least to obtain a cluster sequence;

[0031] The first N clusters in the cluster sequence are identified as the target clusters;

[0032] The cluster center coordinates of each target cluster are determined as the target nasal root coordinates.

[0033] A second aspect of this disclosure provides a nasal root detection device, the device comprising:

[0034] The acquisition module is used to acquire NCCT image sequences of the head of the object to be inspected;

[0035] The first determining module is used to determine the skull region based on the NCCT image sequence;

[0036] The second determining module is used to determine a candidate region for the nasal root from the skull region based on the location features of the nasal root.

[0037] The third determining module is used to determine the position of the nasal root from the candidate region of the nasal root based on the bone shape characteristics at the nasal root.

[0038] Optionally, the second determining module includes:

[0039] The first determining submodule is used to determine the median sagittal plane and the median coronal plane corresponding to the skull region;

[0040] The second determining submodule is used to determine the sagittal tomographic images of interest from the skull region whose distance from the midsagittal plane is less than a preset threshold, and to obtain a sequence of sagittal tomographic images of interest.

[0041] The segmentation submodule is used to segment the candidate nasal root region from the bone region corresponding to the sagittal tomographic image sequence of interest based on the midline coronal plane.

[0042] Optionally, the third determining module includes:

[0043] The detection submodule is used to perform edge detection on the candidate nasal root region to obtain the bone edge region, wherein the bone edge region represents at least one bone region in a sagittal tomographic image of the bone edge.

[0044] A submodule is constructed to build a filter based on the bone shape features at the root of the nose.

[0045] The third determining submodule is used to determine the candidate nasal root coordinates from each of the sagittal tomographic images of the bone edge according to the filter.

[0046] The filtering submodule is used to filter the position of the nasal root from all the candidate nasal root coordinates based on the bone shape features at the nasal root.

[0047] Optionally, the bone shape feature at the nasal root includes a first feature indicating that the angle between the first line connecting the nasal root and the forehead and the second line connecting the nasal root and the nasal dorsum is an obtuse angle. The construction submodule includes:

[0048] The first construction submodule is used to construct an isosceles obtuse triangle with a preset obtuse angle and a preset length of the short side based on the first feature.

[0049] The conversion submodule is used to perform digital conversion on the isosceles obtuse triangle to obtain the filter.

[0050] Optionally, the third determining submodule includes:

[0051] The filtering submodule is used to perform matched filtering on the sagittal tomographic image of the bone edge according to the filter to obtain candidate sagittal plane coordinates, which include sagittal axis coordinates and vertical axis coordinates;

[0052] The first execution submodule is used to determine the three-dimensional coordinates of the candidate nasal root based on the candidate sagittal plane coordinates and the coronal axis coordinates corresponding to the sagittal tomographic image of the bone edge.

[0053] Optionally, the bone shape feature at the nasal root includes a second feature characterizing the location of the nasal root as a concave point, and the filtering submodule includes:

[0054] The clustering submodule is used to perform clustering processing based on all the candidate nasal root coordinates to obtain the target nasal root coordinates;

[0055] The second execution submodule is used to determine the target nasal root coordinate as the position of the nasal root for each target nasal root coordinate, provided that the target nasal root coordinate conforms to the second feature based on the positional features between the target nasal root coordinate and the neighboring bone coordinates of the target nasal root coordinate.

[0056] Optionally, the clustering submodule is used for:

[0057] Clustering is performed on all the candidate nasal root coordinates to obtain multiple clusters, each cluster including at least one candidate nasal root coordinate; the multiple clusters are sorted from most to least according to the number of candidate nasal root coordinates in each cluster to obtain a cluster sequence; the first N clusters in the cluster sequence are determined as target clusters; the cluster center coordinates of each target cluster are determined as the target nasal root coordinates.

[0058] A third aspect of this disclosure provides a non-transitory computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of the nasal root detection method described in the first aspect.

[0059] A fourth aspect of this disclosure provides an electronic device, including:

[0060] A memory on which computer programs are stored;

[0061] A processor for executing the computer program in the memory to implement the steps of the nasal root detection method described in the first aspect.

[0062] By adopting the above technical solution, at least the following beneficial technical effects can be achieved:

[0063] This method involves acquiring NCCT image sequences of the head of the subject and determining the skull region based on these sequences. A candidate region for the nasal root is identified from the skull region based on the location features of the nasal root. The location of the nasal root is then determined from this candidate region based on the bone shape features at the nasal root. This method, which reconstructs the skull region from NCCT image sequences and determines the location of the nasal root based on its location and bone shape features, is more accurate and safer than related techniques that involve moving the patient's head and relying on manual observation and palpation to determine the location of the nasal root, as it does not require moving the patient's head or relying on human experience.

[0064] Other features and advantages of this disclosure will be described in detail in the following detailed description section. Attached Figure Description

[0065] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the following detailed description to explain the present disclosure, but do not constitute a limitation thereof. In the drawings:

[0066] Figure 1 illustrates a human coordinate system according to an exemplary embodiment of the present disclosure.

[0067] Figure 2 is a flowchart illustrating a nasal root detection method according to an exemplary embodiment of the present disclosure.

[0068] Figure 3 illustrates another human coordinate system according to an exemplary embodiment of the present disclosure.

[0069] Figure 4 is a schematic diagram of the raphe in the XOY plane of the brain according to an exemplary embodiment of the present disclosure.

[0070] Figure 5 is a three-dimensional image corresponding to a candidate region according to an exemplary embodiment of the present disclosure.

[0071] Figure 6 is a schematic diagram illustrating a segmentation according to an exemplary embodiment of the present disclosure.

[0072] Figure 7 is a schematic diagram illustrating a first feature according to an exemplary embodiment of the present disclosure.

[0073] Figure 8 is a schematic diagram of an isosceles obtuse triangle according to an exemplary embodiment of the present disclosure.

[0074] Figure 9 is a schematic diagram illustrating a process of performing matched filtering on a sagittal tomographic image of a bone edge according to a filter, based on an exemplary embodiment of the present disclosure.

[0075] Figure 10 is a schematic diagram of a head contour region according to an exemplary embodiment of the present disclosure.

[0076] Figure 11 is a schematic diagram of a binary image of a head contour region according to an exemplary embodiment of the present disclosure.

[0077] Figure 12 is a block diagram illustrating a nasal root detection device according to an exemplary embodiment of the present disclosure.

[0078] Figure 13 is a block diagram illustrating an electronic device according to an exemplary embodiment of the present disclosure. Detailed Implementation

[0079] The specific embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit this disclosure.

[0080] It should be noted that all actions involving the acquisition of signals, information, or data in this disclosure are carried out in compliance with the relevant data protection laws and policies of the country where the location is situated, and with authorization from the owner of the relevant device.

[0081] To facilitate a better understanding of the technical solutions of this disclosure by those skilled in the art, the basic human anatomy sections involved in the embodiments of this disclosure will be described below.

[0082] Sagittal plane: A plane perpendicular to the ground along the front and back of the body is called the sagittal plane. The sagittal plane divides the human body into left and right parts. The sagittal plane drawn along the midline of the body is called the median plane or median sagittal plane. For example, the sagittal plane shown in Figure 1.

[0083] Coronal plane: A plane perpendicular to the ground along the sides of the body is called the coronal plane, also known as the frontal plane. The coronal plane is used to divide the human body into anterior and posterior parts. For example, the (central) coronal plane shown in Figure 1.

[0084] Horizontal plane: A plane perpendicular to the longitudinal axis of the human body and parallel to the ground is called a horizontal plane. The horizontal plane divides the human body into upper and lower parts. For example, the horizontal plane shown in Figure 1.

[0085] The basic axes in human anatomy involved in the embodiments of this disclosure will be further explained below based on Figure 1.

[0086] Coronal axis: The axis that passes perpendicularly through the sagittal plane in the left-right direction is called the coronal axis or frontal axis. This axis can be characterized as the x-axis, such as the x-axis shown in Figure 1.

[0087] Sagittal axis: The axis that passes perpendicularly through the coronal plane in the anteroposterior direction is called the sagittal axis. This axis can be characterized as the y-axis, such as the y-axis shown in Figure 1.

[0088] Vertical axis: An axis that passes vertically through the horizontal plane in the up-down direction is called a vertical axis. This axis can be characterized as the z-axis, such as the z-axis shown in Figure 1.

[0089] The following provides a detailed description of the embodiments of the technical solution disclosed herein.

[0090] Figure 2 is a flowchart illustrating a nasal root detection method according to an exemplary embodiment of the present disclosure. As shown in Figure 2, the nasal root detection method includes the following steps.

[0091] S21. Obtain the NCCT image sequence of the head of the object to be tested.

[0092] CT is short for Computer Tomography. Noncontrast CT (NCCT) is an imaging technique used in the field of acute stroke, typically to check for hyperacute cerebral infarction lesions, rule out cerebral hemorrhage, or hemorrhage transformation caused by arterial reperfusion injury. In this disclosure, the NCCT image sequence refers to the image sequence obtained by performing a CT scan of the head of the subject (e.g., a patient or test subject) using the noncontrast CT method. This image sequence includes multiple slices, each of which is a two-dimensional medical image.

[0093] In other words, the NCCT image sequence of the head of the subject to be examined includes multiple slices, each of which is a two-dimensional image. Since the combination of multiple slices can represent a three-dimensional image of the head, this NCCT image sequence can represent a three-dimensional image of the head of the subject to be examined.

[0094] It should be noted that the NCCT image sequence of the subject's head in this disclosure can be an image sequence obtained by performing a plain CT scan of the subject's head at any time and in any scene. In other words, it is not necessary to specifically perform a plain CT scan on the subject to detect the location of the nasal root; any NCCT image sequence from the subject's medical record can be used. This does not affect the accuracy of determining the location of the nasal root using the nasal root detection method of this disclosure.

[0095] S22. Determine the skull region based on the NCCT image sequence.

[0096] The skull region is the area occupied by the skull in three-dimensional space.

[0097] Since NCCT image sequences are obtained from plain CT scans of the head of the subject being examined, they include imaging information of the head. The head comprises the skull and scalp; therefore, the skull region of the subject can be determined / reconstructed based on the NCCT image sequences.

[0098] S23. Determine candidate regions for the nasal root from the skull region based on the location features of the nasal root.

[0099] In human biology, the root of the nose has a biological location feature on the face, located near the midline of the brain. Therefore, in order to reduce the amount of data processing, a candidate region for the root of the nose, including the midline of the brain, can be identified from the skull region to more quickly determine the location of the root of the nose.

[0100] S24. Determine the position of the nasal root from the candidate nasal root region based on the bone shape characteristics at the nasal root.

[0101] In human biology, the location of the nasal root also has a concave feature relative to the neighboring bone points. Based on this bone shape feature at the nasal root, the specific location of the nasal root can be determined from the candidate region of the nasal root.

[0102] Using the above method, NCCT image sequences of the head of the subject to be tested are acquired, and the skull region is determined based on the NCCT image sequences. A candidate region for the nasal root is determined from the skull region based on the positional features of the nasal root, and the position of the nasal root is determined from the candidate region based on the bone shape features at the nasal root. This method of reconstructing the skull region based on NCCT image sequences and determining the position of the nasal root based on its positional and bone shape features is more accurate and safer than related techniques that first move the patient's head and then rely on manual observation and palpation to determine the position of the nasal root, because it does not require moving the patient's head or relying on human experience.

[0103] Optionally, determining the candidate nasal root region from the skull region based on the location features of the nasal root includes:

[0104] The midsagittal and midcoronal planes corresponding to the skull region are determined; sagittal tomographic images of interest with a distance less than a preset threshold from the midsagittal plane are determined from the skull region to obtain a sagittal tomographic image sequence; the nasal root candidate region is segmented from the bone region corresponding to the sagittal tomographic image sequence based on the midcoronal plane.

[0105] The preset threshold can be an empirical value such as 1.5 cm or 2 cm. The following embodiments of this disclosure are illustrated using a preset threshold of 1.5 cm as an example.

[0106] The definition of the standard coordinate system for medical imaging is shown in Figure 3, where the x-axis represents the coronal axis, the y-axis represents the sagittal axis, and the z-axis represents the vertical axis.

[0107] For example, since the root of the nose has a biological location characteristic of the face near the midline of the brain in human biology, in the process of determining the candidate region of the root of the nose, the candidate region including the midline of the brain can be determined first, and then the candidate region of the root of the nose including the face can be determined from the candidate region.

[0108] In determining the candidate region including the midline of the brain, it is necessary to first determine the midsagittal plane where the midline of the brain is located, for example, the YOZ sagittal plane where the midline of the brain is located in the XOY horizontal view shown in Figure 4. Then, sagittal tomographic images of interest that are less than 1.5 cm away from the midsagittal plane are determined from the skull region, for example, the YOZ sagittal plane within the range of the two white lines shown in Figure 4. All sagittal tomographic images of interest constitute the above-mentioned candidate region, for example, the candidate region corresponds to the three-dimensional image shown in Figure 5.

[0109] It should be noted here that this disclosure does not limit the number of slices in the sagittal tomographic image of interest. For example, assuming the spatial resolution of the coronal axis X is normalized to 1 mm, then the number of slices in the sagittal tomographic image of interest within a 3 cm range can be 30.

[0110] After identifying candidate regions including the raphe ridge of the brain, it is necessary to further identify candidate regions for the nasal root, including the face, from these candidate regions. For example, since the median coronal plane divides the human body into anterior and posterior parts, specifically the skull region, it can be used to segment each sagittal tomographic image to obtain candidate regions for the nasal root, including the facial bones. For instance, segmenting the sagittal tomographic image shown in Figure 6, the candidate nasal root region includes the bone region in the left-hand sagittal tomographic sub-image of Figure 6.

[0111] Optionally, determining the location of the nasal root from the candidate nasal root region based on the bone shape characteristics at the nasal root includes:

[0112] Edge detection is performed on the candidate nasal root region to obtain a bone edge region, wherein the bone edge region represents a bone region in at least one bone edge sagittal tomographic image; a filter is constructed based on the bone shape features at the nasal root; for each bone edge sagittal tomographic image, candidate nasal root coordinates are determined from the bone edge sagittal tomographic image based on the filter; the position of the nasal root is obtained by filtering from all the candidate nasal root coordinates based on the bone shape features at the nasal root.

[0113] Since the candidate region of the nasal root is segmented from the bone region corresponding to the sagittal tomographic image sequence of interest based on the midline coronal plane, the bone edge region obtained by edge detection of the candidate region of the nasal root can be characterized as the bone region in the sagittal tomographic image sequence corresponding to the number of images in the sagittal tomographic image sequence of interest.

[0114] For example, a filter is constructed based on the bone shape features at the nasal root. Each sagittal tomographic image of a bone edge is traversed, and candidate nasal root coordinates, all conforming to the bone shape features at the nasal root, are filtered out from these images using the filter.

[0115] Because the process of filtering candidate nasal root coordinates from sagittal tomographic images of the bone margin does not distinguish which side of the bone margin is medial and which side is lateral, the filtered candidate nasal root coordinates may be concave or convex points on the bone margin. Therefore, it is necessary to further refine the candidate nasal root coordinates based on the bone shape characteristics at the nasal root to obtain a more accurate nasal root location.

[0116] Optionally, the bone shape features at the nasal root include a first feature indicating that the angle between the first line connecting the nasal root and the forehead and the second line connecting the nasal root and the nasal dorsum is an obtuse angle, and the step of constructing a filter based on the bone shape features at the nasal root includes:

[0117] Based on the first feature, an isosceles obtuse triangle with a preset obtuse angle and a preset short side length is constructed; the isosceles obtuse triangle is digitally converted to obtain the filter.

[0118] For example, referring to Figure 7, based on the first characteristic that the angle between the first line connecting the root of the nose and the forehead and the second line connecting the root of the nose and the bridge of the nose is an obtuse angle, and the statistical characteristic that this obtuse angle is approximately 120° to 130°, an isosceles obtuse triangle with a preset obtuse angle such as 120° and a preset side length such as 2 cm can be constructed, as shown in Figure 8. m Furthermore, the isosceles obtuse triangle is digitally transformed, such as into a 12*12, 9*9, or 9*12 digital matrix, to obtain the filter.

[0119] Optionally, determining the candidate nasal root coordinates from the sagittal tomographic image of the bone margin based on the filter includes:

[0120] The sagittal tomographic image of the bone edge is subjected to matched filtering based on the filter to obtain candidate sagittal plane coordinates, which include sagittal axis coordinates and vertical axis coordinates. The three-dimensional candidate nasal root coordinates are determined based on the candidate sagittal plane coordinates and the coronal axis coordinates corresponding to the sagittal tomographic image of the bone edge.

[0121] For example, matched filtering is performed on the sagittal tomographic image of the bone margin according to the filter to obtain candidate sagittal plane coordinates, which include the sagittal axis coordinate y and the vertical axis coordinate z. The three-dimensional candidate nasal root coordinates (x, y, z) are determined based on the candidate sagittal plane coordinates and the corresponding coronal axis coordinate x of the sagittal tomographic image of the bone margin.

[0122] The following uses Figure 9 as an example to illustrate the process of matched filtering on a sagittal tomographic image of the bone margin. Using a sliding filter, points p1 and p2 of the isosceles obtuse triangle (i.e., the visual representation of the filter) are controlled to slide on the position of the bone margin voxel (white area) in the sagittal tomographic image of the bone margin shown in Figure 9. During the sliding process, points p1 and p2 remain on the bone margin voxel. During the movement, if the p1 and p2 of the isosceles obtuse triangle... m If the point is also located on the bone edge voxel, then the isosceles obtuse triangle at this time is determined to match the bone edge, and the coordinates of the Pm point in the sagittal tomographic image of the bone edge at this time are determined as candidate sagittal plane coordinates.

[0123] Optionally, the bone shape feature at the nasal root includes a second feature characterizing the location of the nasal root as a concave point, and the step of filtering the position of the nasal root from all the candidate nasal root coordinates based on the bone shape feature at the nasal root includes:

[0124] Clustering is performed on all the candidate nasal root coordinates to obtain the target nasal root coordinates; for each target nasal root coordinate, if the target nasal root coordinate meets the second feature based on the positional features between the target nasal root coordinate and the neighboring bone coordinates of the target nasal root coordinate, the target nasal root coordinate is determined as the position of the nasal root.

[0125] One implementation method for obtaining the target nasal root coordinates by clustering all candidate nasal root coordinates is as follows: cluster all candidate nasal root coordinates to obtain multiple clusters, each cluster including at least one candidate nasal root coordinate; sort the multiple clusters from most to least according to the number of candidate nasal root coordinates in each cluster to obtain a cluster sequence; determine the first N clusters in the cluster sequence as the target clusters; and determine the cluster center coordinates of each target cluster as the target nasal root coordinates.

[0126] Due to various errors, including the acquisition accuracy and imaging errors of NCCT images, segmentation errors of sagittal tomographic images of bone margins, and errors introduced by other factors, the candidate nasal root coordinates P obtained by sliding filtering are... x1,y1,z1 ,P x2,y2,z2 ...P xm,ym,zm 'm' represents the number of candidate nasal root coordinates. These may belong to multiple categories, and before clustering, it's uncertain how many categories the candidate nasal root coordinates will be divided into, or their location. Therefore, in this embodiment, the DBSCAN clustering algorithm is used to cluster the candidate nasal root coordinates, resulting in multiple classes (clusters). Furthermore, during the DBSCAN clustering process, the clustering radius, the input value of the DBSCAN clustering algorithm, is used...

[0127] m

[0128] Set the height to 5cm and the number of items in each category to 2.

[0129] Furthermore, based on the number of candidate nasal root coordinates in each cluster, the multiple clusters are sorted from most to least, resulting in a cluster sequence. The first N clusters in the cluster sequence are determined as target clusters. The cluster center coordinates of each target cluster are determined as target nasal root coordinates. It should be noted that, besides the aforementioned sorting method based on most to least, other methods can be used to select the top N target clusters with the most candidate nasal root coordinates from multiple clusters, such as sorting from least to most, or using a bubble sort algorithm. This disclosure does not impose specific limitations on this. The value of N can be 1, 2, 3, etc.

[0130] Furthermore, for each target nasal root coordinate, if the target nasal root coordinate meets the second feature based on the positional characteristics between the target nasal root coordinate and the neighboring bone coordinates of the target nasal root coordinate, that is, if the target nasal root coordinate is a concave point coordinate, the target nasal root coordinate is determined as the position of the nasal root.

[0131] Optionally, determining the skull region based on the NCCT image sequence includes:

[0132] The NCCT image sequence is preprocessed to obtain a preprocessed NCCT image sequence; the head contour region is determined based on the preprocessed NCCT image sequence; the skull region is determined from the head contour region based on the CT value range of the bony structure.

[0133] Preprocessing of NCCT image sequences includes denoising and / or image correction.

[0134] In one scenario, during a plain CT scan, auxiliary devices are often used to stabilize the patient's head to prevent image retention due to head movement. Therefore, the resulting NCCT image sequence may include not only the patient's head imaging information but also the imaging information of the auxiliary devices. This auxiliary device imaging information is interference and can interfere with subsequent image processing. In view of this, in this embodiment, the NCCT image sequence undergoes preprocessing, such as noise reduction, to remove noise interference, resulting in a noise-reduced preprocessed NCCT image sequence. For example, for each image in the NCCT image sequence, the imaging information of the auxiliary devices, such as the bed board, is removed, resulting in a first NCCT image sequence after removing the bed board imaging information.

[0135] In another scenario, during a plain CT scan, the patient's head may tilt due to the lack of an auxiliary device to stabilize it, or the head may tilt even when an auxiliary device is used to stabilize it. Therefore, the resulting NCCT image sequence may show a tilted head image, which can interfere with subsequent image processing. In view of this, in this embodiment, the NCCT image sequence undergoes preprocessing, such as image correction, to correct the tilted head image, resulting in a corrected preprocessed NCCT image sequence.

[0136] In addition, there is another situation where the NCCT image sequence may contain noise interference from auxiliary devices, or the head imaging image may be tilted. Therefore, in some embodiments, the preprocessing of the NCCT image sequence to obtain a preprocessed NCCT image sequence includes:

[0137] For each image in the NCCT image sequence, the imaging information of the auxiliary device on the image is removed to obtain a first NCCT image sequence; according to the standard NCCT image sequence, the first NCCT image sequence is subjected to rigid image registration processing to obtain the registered preprocessed NCCT image sequence.

[0138] For example, for each image in an NCCT image sequence, the imaging information of auxiliary devices is removed from that image to obtain a first NCCT image sequence. Each image in this first NCCT image sequence is a clean image after removing the imaging information of auxiliary devices. Furthermore, since in the medical field, the two images for rigid registration can be medical images from different subjects, an NCCT image sequence with no lesions in the head, no obvious abnormalities in the bony features of the head, and no angular displacement of the patient's head during plain CT scan can be selected as the standard NCCT image sequence. Then, rigid registration processing is performed on the first NCCT image sequence based on this standard NCCT image sequence to obtain a pre-processed NCCT image sequence after registration.

[0139] Since the NCCT image sequence represents a three-dimensional image of the head, the first NCCT image sequence can be rigidly registered according to the standard NCCT image sequence to obtain the registered preprocessed NCCT image sequence. This can be achieved by using the ITK (Insight Toolkit) three-dimensional rigid registration tool in related technologies to rigidly register the first NCCT image sequence according to the standard NCCT image sequence to obtain the registered preprocessed NCCT image sequence.

[0140] The auxiliary device may be a bed board, headrest, head restraint, or other similar device. One method for removing the imaging information of the auxiliary device from an image is to remove it based on its morphological characteristics. For example, the HU value of the voxel representing the auxiliary device can be set to 0. Another method is to manually remove the imaging information of the auxiliary device.

[0141] Optionally, determining the head contour region based on the preprocessed NCCT image sequence includes:

[0142] For each preprocessed image in the preprocessed NCCT image sequence, a first target voxel with a CT value greater than a first preset threshold is identified in the preprocessed image; the region composed of all the first target voxels in the preprocessed NCCT image sequence is identified as the head contour region.

[0143] In one implementation, if the ITK 3D rigid registration tool from related technologies is used to perform rigid registration processing on a first NCCT image sequence based on a standard NCCT image sequence, a preprocessed NCCT image sequence is obtained. Since the ITK 3D rigid registration tool's algorithm normalizes the CT values ​​(HU values) of voxels near the head to 0, and the HU value of air is approximately -1000, while the HU value of the head is greater than 0, a first preset threshold can be set to 0. For each preprocessed image in the preprocessed NCCT image sequence, a first target voxel with a CT value greater than 0 is identified. The region comprised of all first target voxels in the preprocessed NCCT image sequence is defined as the head contour region.

[0144] To facilitate a more intuitive understanding by those skilled in the art of how this disclosure determines the head contour region, Figures 10 and 11 are used as examples below. For each preprocessed image in the preprocessed NCCT image sequence shown in Figure 10, a first target voxel with a CT value greater than 0 is identified. The HU value of the first target voxel is set to 1, and the remaining voxels are set to 0, resulting in a binary image as shown in Figure 11. In the image neighborhood, 1 represents white and 0 represents black. The white area in each image in Figure 11 represents the head region. The three-dimensional image obtained by combining the head regions in all images is the head contour region.

[0145] Optionally, the CT value range based on bony structure, which determines the skull region from the head contour region, includes:

[0146] Determine a second target voxel whose CT value falls within the CT value range from the head contour region; define the region composed of all the second target voxels as the skull region.

[0147] Since the CT value (HU value) of bony structures is 150 to 1000, the CT value range of bony structures can be [150, 1000].

[0148] For example, since the skull is located within the head contour region, all voxels representing the skull can be determined from within the head contour region. Voxels with CT values ​​in the range [150, 1000] within the head contour region are identified as the second target voxels representing the skull. Because the second target voxels represent the skull, the three-dimensional region comprised of all the second target voxels constitutes the skull region.

[0149] The skull region can be represented as the white area in the three-dimensional binary image corresponding to the NCCT binary image sequence obtained by setting the second target voxel in the NCCT image sequence to 1 and the other voxels to 0.

[0150] Figure 12 is a block diagram illustrating a nasal root detection device according to an exemplary embodiment of the present disclosure.

[0151] As shown in Figure 12, the nasal root detection device 1200 includes:

[0152] The acquisition module 1201 is used to acquire the NCCT image sequence of the head of the object to be detected;

[0153] The first determining module 1202 is used to determine the skull region based on the NCCT image sequence;

[0154] The second determining module 1203 is used to determine a candidate region for the nasal root from the skull region based on the positional features of the nasal root.

[0155] The third determining module 1204 is used to determine the position of the nasal root from the candidate region of the nasal root based on the bone shape characteristics at the nasal root.

[0156] Using the aforementioned device 1200, an NCCT image sequence of the head of the subject to be examined is acquired, and the skull region is determined based on the NCCT image sequence. A candidate region for the nasal root is determined from the skull region based on the positional characteristics of the nasal root, and the position of the nasal root is determined from the candidate region based on the bone shape characteristics at the nasal root. This method of reconstructing the skull region based on the NCCT image sequence of the head and determining the position of the nasal root based on its positional and bone shape characteristics is more accurate and safer than related techniques that first move the patient's head and then rely on manual observation and palpation to determine the position of the nasal root, because it does not require moving the patient's head or relying on human experience.

[0157] Optionally, the second determining module 1203 includes:

[0158] The first determining submodule is used to determine the median sagittal plane and the median coronal plane corresponding to the skull region;

[0159] The second determining submodule is used to determine the sagittal tomographic images of interest from the skull region whose distance from the midsagittal plane is less than a preset threshold, and to obtain a sequence of sagittal tomographic images of interest.

[0160] The segmentation submodule is used to segment the candidate nasal root region from the bone region corresponding to the sagittal tomographic image sequence of interest based on the midline coronal plane.

[0161] Optionally, the third determining module 1204 includes:

[0162] The detection submodule is used to perform edge detection on the candidate nasal root region to obtain the bone edge region, wherein the bone edge region represents at least one bone region in a sagittal tomographic image of the bone edge.

[0163] A submodule is constructed to build a filter based on the bone shape features at the root of the nose.

[0164] The third determining submodule is used to determine the candidate nasal root coordinates from each of the sagittal tomographic images of the bone edge according to the filter.

[0165] The filtering submodule is used to filter the position of the nasal root from all the candidate nasal root coordinates based on the bone shape features at the nasal root.

[0166] Optionally, the bone shape feature at the nasal root includes a first feature indicating that the angle between the first line connecting the nasal root and the forehead and the second line connecting the nasal root and the nasal dorsum is an obtuse angle. The construction submodule includes:

[0167] The first construction submodule is used to construct an isosceles obtuse triangle with a preset obtuse angle and a preset length of the short side based on the first feature.

[0168] The conversion submodule is used to perform digital conversion on the isosceles obtuse triangle to obtain the filter.

[0169] Optionally, the third determining submodule includes:

[0170] The filtering submodule is used to perform matched filtering on the sagittal tomographic image of the bone edge according to the filter to obtain candidate sagittal plane coordinates, which include sagittal axis coordinates and vertical axis coordinates;

[0171] The first execution submodule is used to determine the three-dimensional coordinates of the candidate nasal root based on the candidate sagittal plane coordinates and the coronal axis coordinates corresponding to the sagittal tomographic image of the bone edge.

[0172] Optionally, the bone shape feature at the nasal root includes a second feature characterizing the location of the nasal root as a concave point, and the filtering submodule includes:

[0173] The clustering submodule is used to perform clustering processing based on all the candidate nasal root coordinates to obtain the target nasal root coordinates;

[0174] The second execution submodule is used to determine the target nasal root coordinate as the position of the nasal root for each target nasal root coordinate, provided that the target nasal root coordinate conforms to the second feature based on the positional features between the target nasal root coordinate and the neighboring bone coordinates of the target nasal root coordinate.

[0175] Optionally, the clustering submodule is used for:

[0176] Clustering is performed on all the candidate nasal root coordinates to obtain multiple clusters, each cluster including at least one candidate nasal root coordinate; the multiple clusters are sorted from most to least according to the number of candidate nasal root coordinates in each cluster to obtain a cluster sequence; the first N clusters in the cluster sequence are determined as target clusters; the cluster center coordinates of each target cluster are determined as the target nasal root coordinates.

[0177] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0178] Figure 13 is a block diagram illustrating an electronic device 700 according to an exemplary embodiment of the present disclosure. As shown in Figure 13, the electronic device 700 may include a processor 701 and a memory 702.

[0179] The processor 701 controls the overall operation of the electronic device 700 to complete all or part of the steps in the aforementioned nasal root detection method. The memory 702 stores various types of data to support the operation of the electronic device 700. This data may include, for example, instructions for any application or method operating on the electronic device 700, as well as application-related data, such as image data.

[0180] In one exemplary embodiment, a computer-readable storage medium including program instructions is also provided, which, when executed by a processor, implement the steps of the nasal root detection method described above. For example, the computer-readable storage medium may be the memory 702 including the program instructions, which may be executed by the processor 701 of the electronic device 700 to complete the nasal root detection method described above.

[0181] In another exemplary embodiment, a computer program product is also provided, the computer program product comprising a computer program executable by a programmable device, the computer program having a code portion for performing the above-described nasal root detection method when executed by the programmable device.

[0182] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure.

[0183] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0184] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

Claims

1. A method for detecting the root of the nose, characterized in that, The method includes: acquiring an NCCT image sequence of the head of the subject to be detected; determining a skull region based on the NCCT image sequence; determining a candidate nasal root region from the skull region based on the positional features of the nasal root; determining the position of the nasal root from the candidate nasal root region based on the bone shape features at the nasal root; the step of determining the position of the nasal root from the candidate nasal root region based on the bone shape features at the nasal root includes: performing edge detection on the candidate nasal root region to obtain a bone edge region, the bone edge region representing at least one bone region in a sagittal tomographic image of a bone edge; constructing a filter based on the bone shape features at the nasal root; and for each... The sagittal tomographic image of the bone edge is used to determine candidate nasal root coordinates based on the filter. The position of the nasal root is obtained by filtering all the candidate nasal root coordinates based on the bone shape features at the nasal root. The bone shape features at the nasal root include a first feature indicating that the angle between the first line connecting the nasal root and the forehead and the second line connecting the nasal root and the nasal dorsum is obtuse. The step of constructing a filter based on the bone shape features at the nasal root includes: constructing an isosceles obtuse triangle with a preset obtuse angle and a preset side length based on the first feature; and performing digital conversion on the isosceles obtuse triangle to obtain the filter.

2. The method according to claim 1, characterized in that, The step of determining the candidate nasal root region from the skull region based on the positional features of the nasal root includes: determining the midsagittal plane and the midcoronal plane corresponding to the skull region; determining sagittal tomographic images of interest from the skull region whose distance from the midsagittal plane is less than a preset threshold, thereby obtaining a sagittal tomographic image sequence; and segmenting the candidate nasal root region from the bone region corresponding to the sagittal tomographic image sequence of interest according to the midcoronal plane.

3. The method according to claim 1, characterized in that, The step of determining candidate nasal root coordinates from the sagittal tomographic image of the bone edge according to the filter includes: performing matched filtering on the sagittal tomographic image of the bone edge according to the filter to obtain candidate sagittal plane coordinates, the candidate sagittal plane coordinates including sagittal axis coordinates and vertical axis coordinates; and determining the three-dimensional candidate nasal root coordinates according to the candidate sagittal plane coordinates and the coronal axis coordinates corresponding to the sagittal tomographic image of the bone edge.

4. The method according to claim 1, characterized in that, The bone shape feature at the root of the nose includes a second feature that characterizes the location of the root of the nose as a concave point. The step of filtering the position of the root of the nose from all the candidate root of the nose coordinates based on the bone shape feature at the root of the nose includes: performing clustering processing on all the candidate root of the nose coordinates to obtain target root of the nose coordinates; for each target root of the nose coordinates, if the target root of the nose coordinates meets the second feature based on the positional features between the target root of the nose coordinates and the neighboring bone coordinates of the target root of the nose coordinates, the target root of the nose coordinates is determined as the position of the root of the nose.

5. The method according to claim 4, characterized in that, The step of clustering all the candidate nasal root coordinates to obtain the target nasal root coordinates includes: clustering all the candidate nasal root coordinates to obtain multiple clusters, each cluster including at least one candidate nasal root coordinate; sorting the multiple clusters from most to least according to the number of candidate nasal root coordinates in each cluster to obtain a cluster sequence; determining the first N clusters in the cluster sequence as the target clusters; and determining the cluster center coordinates of each target cluster as the target nasal root coordinates.

6. A nasal root detection device, characterized in that, The device includes: an acquisition module for acquiring NCCT image sequences of the head of a subject to be tested; a first determination module for determining a skull region based on the NCCT image sequences; a second determination module for determining a candidate nasal root region from the skull region based on the positional features of the nasal root; and a third determination module for determining the position of the nasal root from the candidate nasal root region based on the bone shape features at the nasal root. The third determination module includes: a detection submodule for performing edge detection on the candidate nasal root region to obtain a bone edge region, wherein the bone edge region represents at least one bone region in a sagittal tomographic image of a bone edge; and a construction submodule for constructing a filter based on the bone shape features at the nasal root. The stator module is used to determine candidate nasal root coordinates from each sagittal tomographic image of the bone edge according to the filter; the filter module is used to filter the position of the nasal root from all the candidate nasal root coordinates according to the bone shape features at the nasal root; the bone shape features at the nasal root include a first feature indicating that the angle between the first line connecting the nasal root and the forehead and the second line connecting the nasal root and the nasal dorsum is obtuse; the construction module includes: a first construction module, used to construct an isosceles obtuse triangle with a preset obtuse angle and a preset side length based on the first feature; and a conversion module, used to perform digital conversion on the isosceles obtuse triangle to obtain the filter.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method described in any one of claims 1-5.

8. An electronic device, characterized in that, include: A memory on which computer programs are stored; A processor for executing the computer program in the memory to implement the steps of the method according to any one of claims 1-5.

Images