Image position adjusting method and device, electronic equipment and storage medium

Abstract

The invention discloses an image position adjusting method and device, electronic equipment and a storage medium, and the method comprises the steps: determining an attenuation image based on head data obtained through scanning, and enabling the head offset in the head data to be a first offset; determining image adjustment data based on the attenuated image; based on the image adjustment data and the head data, a target PET image is obtained, the head offset in the target PET image is a second offset, and the second offset is smaller than the first offset.

Description

Technical Field

[0001] The present invention relates to the technical fields such as medical imaging, and in particular, to an image position adjustment method, device, electronic device and storage medium. Background Art

[0002] PET / CT (Positron Emission Tomography / Computed Tomography) is an advanced imaging examination device. Among them, PET mainly reflects the metabolic information of tissues, and CT mainly reflects the anatomical information. The combination of the two is of great significance in brain examinations. However, when the patient's head is slightly tilted, how to ensure that each layer of the image obtained by the PET / CT device is symmetric has become the focus of research.

[0003] Related technologies obtain the facial image of the scanned object, extract a number of feature points throughout the facial image based on the facial image, calculate the rotation matrix and tilt direction category of the face, and adjust the scanning device or the output image to make the output image frontal. This method has a slow scanning speed for conventional PET devices (without an independent facial acquisition module such as a camera); another related technology extracts the sagittal chain codes of the upper and lower edges of the brain tissue in the medical image, and then detects and determines the target corner points on the upper and lower sections of the chain codes respectively. The line connecting these target corner points is the head symmetry axis. This method has no clear usage scenario or application value. Summary of the Invention

[0004] Embodiments of the present invention aim to solve at least one of the technical problems in the related technologies to some extent. For this reason, an object of the present invention is to provide an image position adjustment method, device, electronic device and storage medium, which reduce the offset degree of the user's head in the image, improve the readability of the image, and facilitate doctors to use the image as a diagnostic reference.

[0005] Embodiments of the present invention provide an image position adjustment method, and the image position adjustment method includes: determining an attenuation image based on the scanned head data, where the head offset in the head data is a first offset; determining image adjustment data based on the attenuation image; obtaining a target PET image based on the image adjustment data and the head data, where the head offset in the target PET image is a second offset, and the second offset is less than the first offset.

[0006] Exemplarily, determining image adjustment data based on the attenuation image includes: obtaining preliminary direction information of the head and reference direction information of the head based on the attenuation image; adjusting the attenuation image based on the preliminary direction information of the head and the reference direction information of the head to obtain image adjustment data.

[0007] Exemplarily, based on the attenuation image, preliminary direction information of the head and reference direction information of the head are obtained, including: detecting the cranial vertex region and the cranial vertex in the attenuation image; obtaining the preliminary direction information of the head based on the normal vector direction of the cranial vertex region and the cranial vertex; and obtaining the reference direction information of the head based on the preliminary direction information of the head.

[0008] Exemplarily, detecting the cranial vertex region and the cranial vertex in the attenuation image includes: performing image segmentation on the attenuation image to obtain the skull region; detecting the frontal sinus position in the skull region; determining the cranial vertex region from the skull region based on the frontal sinus position; and obtaining the cranial vertex based on the central position of the cranial vertex region.

[0009] Exemplarily, determining the cranial vertex region from the skull region based on the frontal sinus position includes: determining the outer surface curve of the skull from the skull region based on the frontal sinus position; determining a plurality of local curves from the outer surface curve of the skull; respectively determining the overall curvature of each local curve in the plurality of local curves; and determining the local curve whose overall curvature satisfies a preset curvature condition as the cranial vertex region from the plurality of local curves.

[0010] Exemplarily, obtaining the preliminary direction information of the head based on the normal vector direction of the cranial vertex region and the cranial vertex includes: starting from the cranial vertex and extending along the normal vector direction of the cranial vertex region to obtain the preliminary direction information of the head.

[0011] Exemplarily, obtaining the reference direction information of the head based on the preliminary direction information of the head includes: rotating the attenuation image based on the preliminary direction information of the head according to the minimum moment of inertia method to obtain the reference direction information of the head.

[0012] Exemplarily, the image adjustment data includes a rotation matrix. Based on the preliminary direction information of the head and the reference direction information of the head, the attenuation image is adjusted to obtain the image adjustment data, including: rotating the attenuation image based on the preliminary direction information of the head and the reference direction information of the head to obtain the rotation matrix.

[0013] Exemplarily, rotating the attenuation image based on the preliminary direction information of the head and the reference direction information of the head to obtain the rotation matrix includes: rotating the attenuation image to align the preliminary direction information of the head and the reference direction information of the head to obtain the rotation matrix.

[0014] Exemplarily, the preliminary direction information of the head includes a preliminary Z-direction symmetry axis, and the reference direction information of the head includes a reference Z-direction symmetry axis. In the image coordinate system, the preliminary Z direction is the Z axis of the image coordinate system, and the image coordinate system further includes an X axis and a Y axis; rotating the attenuation image to align the preliminary direction information of the head and the reference direction information of the head, obtaining a rotation matrix, including: rotating the attenuation image to align the preliminary Z-direction symmetry axis with the reference Z-direction symmetry axis, obtaining at least one of the X-axis rotation angle component, the Y-axis rotation angle component, and the Z-axis rotation angle component, where rotating the attenuation image includes rotating the attenuation image around at least one of the X axis, the Y axis, and the Z axis; obtaining the rotation matrix based on at least one of the X-axis rotation angle component, the Y-axis rotation angle component, and the Z-axis rotation angle component. Exemplarily, obtaining a target PET image based on the image adjustment data and the head data includes: performing image reconstruction on the head data based on the image adjustment data to obtain the target PET image.

[0015] Exemplarily, performing image reconstruction on the head data based on the image adjustment data to obtain a target PET image includes: performing image reconstruction on the head data, the image adjustment data, the scanning system matrix, and the scanning system noise to obtain the target PET image.

[0016] Exemplarily, obtaining a target PET image based on the image adjustment data and the head data includes: obtaining an initial PET image based on the head data; adjusting the initial PET image based on the image adjustment data to obtain the target PET image.

[0017] Another embodiment of the present invention provides an image position adjustment device, which includes: a first determination module, configured to determine an attenuation image based on the head data obtained by scanning, where the head offset in the head data is a first offset; a second determination module, configured to determine image adjustment data based on the attenuation image; an obtaining module, configured to obtain a target PET image based on the image adjustment data and the head data, where the head offset in the target PET image is a second offset, and the second offset is less than the first offset.

[0018] An embodiment of the present invention provides an electronic device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the steps of the method in any of the above embodiments are implemented.

[0019] An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method in any of the above embodiments are implemented.

[0020] In the above-described embodiments, the image position adjustment method includes: determining an attenuation image based on the scanned head data, wherein the head offset in the head data is a first offset; determining image adjustment data based on the attenuation image; and obtaining a target PET image based on the image adjustment data and the head data, wherein the head offset in the target PET image is a second offset, and the second offset is smaller than the first offset. The above method determines the image adjustment data based on the attenuation image and obtains the target PET image based on the image adjustment data and the head data, thereby realizing the positioning correction of the head PET image. Each layer of the obtained target PET image is axisymmetric, improving the image readability and enhancing the diagnostic accuracy.

[0021] Additional aspects and advantages of the present invention will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present invention. Description of the Drawings

[0022] Figure 1 Flowchart of the image scanning method provided for the related art;

[0023] Figure 2 Flowchart of the head symmetry axis recognition method provided for another related art;

[0024] Figure 3 Flowchart of the image position adjustment method provided for the embodiments of the present invention;

[0025] Figure 4 Skull distribution diagram of the top of the head provided for the embodiments of the present invention;

[0026] Figure 5 Flowchart of the reference Z-axis symmetry axis determination method provided for the embodiments of the present invention;

[0027] Figure 6 Schematic diagram of rotation based on the X-axis and Y-axis provided for the embodiments of the present invention;

[0028] Figure 7 Schematic diagram of rotation based on the Z-axis provided for the embodiments of the present invention;

[0029] Figure 8 Flowchart of the detailed implementation of the image position adjustment method provided for the embodiments of the present invention;

[0030] Figure 9 Block diagram of the image position adjustment device provided for the embodiments of the present invention;

[0031] Figure 10 Block diagram of the electronic device provided for the embodiments of the present invention. Detailed Embodiments

[0032] Embodiments of the present invention will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where like or similar reference numerals denote like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are intended to explain the present invention, but should not be construed as limiting the present invention.

[0033] PET / CT (Positron Emission Tomography / Computed Tomography) is an advanced imaging examination device. Among them, PET mainly reflects the metabolic information of tissues, and CT mainly reflects the anatomical information. PET / CT can reflect indicators such as cerebral glucose metabolism and oxygen metabolism. Because tumor cells are metabolically active and have a strong ability to uptake tracers (radionuclides), the location, size, grade, and malignancy of tumors can be judged based on this, providing a basis for subsequent treatment; by comparing the cerebral metabolic differences in different regions, doctors can detect degenerative diseases such as Alzheimer's disease at an early stage for timely treatment and detection; by comparing the metabolic differences between the seizure period and the interictal period of epilepsy, doctors can find the brain tissue with abnormal discharges, providing an important reference for the diagnosis and surgical treatment of epilepsy. The combination of the two is of great significance in brain examinations. However, when the patient's head is tilted slightly, how to ensure that each layer of the image obtained by the PET / CT device is symmetric has become the focus of research.

[0034] Figure 1 It is a flowchart of an image scanning method provided by the related art.

[0035] As Figure 1 shown, the image scanning method 100 includes step S110 to step S130.

[0036] Step S110, obtaining a facial image of the target scanning object.

[0037] Step S120, obtaining the tilt direction category of the face based on the facial image.

[0038] Step S130, based on the tilt direction category, adopting the corresponding processing method for processing.

[0039] This technology obtains the facial image of the scanned object, extracts a number of feature points based on this facial image, calculates the rotation matrix and tilt direction category of the face, and adjusts the scanning device or the output image to make the output image frontal. This method is applicable to scenarios such as cameras and CT that can quickly complete and conveniently re-acquire facial features. For example, in the case of the pitch tilt direction category mentioned in this invention, it is necessary to first obtain the facial image. If it is determined that the tilt type is pitch tilt, it is necessary to adjust the scanning device and re-obtain the facial image. This cycle continues until the pitch tilt determination condition is no longer met, and then the formal scanning begins. For conventional PET devices (without an independent facial acquisition module such as a camera), the scanning speed is slow, and the clarity of facial features is strongly related to the type of radionuclide-labeled drug, so this method is not applicable.

[0040] Figure 2 Flowchart of the head symmetry axis recognition method provided for another related technology.

[0041] As Figure 2 shown, the head symmetry axis recognition method 200 includes steps S210 to S240.

[0042] Step S210, identify the brain tissue edge chain code and the chain code cutting point corresponding to the brain tissue region in the target head medical image, and extract the anterior superior sagittal sinus chain code and the posterior superior sagittal sinus chain code in the brain tissue edge chain code according to the positional relationship between the brain tissue edge chain code and the chain code cutting point.

[0043] Exemplarily, the superior sagittal sinus is an important venous structure in the brain. The sagittal chain code is often used to describe object contour or shape information and has certain applications in the fields of image analysis and computer vision. The sagittal chain code is essentially a way to encode the contour direction with a digital sequence. For the upper and lower edges of an object, along the edge contour, starting from a specific starting point, according to the preset direction rule, a corresponding digital code is assigned to each step segment of the contour. For example, a common rule is to use numbers such as 0-7 to represent 8 different directions for the line segment direction between adjacent contour points in the clockwise or counterclockwise direction.

[0044] First, accurately segment the target object from the image. Then, use the edge detection algorithm (Canny edge detection) to identify all the edge pixels of the object, and then screen and locate the upper and lower edge pixel sets from them. This step is extremely crucial for generating the exclusive chain code subsequently. Incorrect edge extraction will lead to a huge deviation in the chain code description. Starting from one endpoint of the upper and lower edges, in accordance with the established direction sequence, check the displacement directions between adjacent pixels one by one, and replace them with corresponding digital encodings. Finally, form a string of digital chain codes representing the edge shape. In this way, this digital chain code can compactly and effectively record the geometric characteristics of the upper and lower edge contours of the object, and can be used for subsequent tasks such as shape matching, target recognition, and pose estimation, and realize the comparative analysis with the existing model by virtue of the chain code data. For example, in medical image analysis, generate sagittal chain codes for the upper and lower edges of human bones to assist in diagnosing whether the bone morphology is abnormal.

[0045] Step S220, determine the front-segment candidate corner points according to the front-segment curvature values of the points in the front-segment superior sagittal sinus chain code, and determine the rear-segment candidate corner points according to the rear-segment curvature values of the points in the rear-segment superior sagittal sinus chain code.

[0046] Step S230, detect the first position feature of the front-segment candidate corner points in the front-segment superior sagittal sinus chain code, and determine the front-segment target corner points based on the first position feature, and detect the second position feature of the rear-segment candidate corner points in the rear-segment superior sagittal sinus chain code, and determine the rear-segment target corner points based on the second position feature.

[0047] Step S240, determine the head symmetry axis corresponding to the brain tissue region according to the connection line between the front-segment target corner points and the rear-segment target corner points.

[0048] This technology extracts the sagittal chain codes of the upper and lower edges of the brain tissue in the medical image, and then detects and determines the target corner points on the upper and lower segments of the chain code respectively. The connection line of these target corner points is the head symmetry axis. This method only uses the edge feature points of the brain tissue to identify the symmetry axis, does not use this symmetry axis to complete related tasks, and has no clear usage scenario or application value.

[0049] Therefore, an embodiment of the present invention provides an image position adjustment method, which can realize the automatic adjustment of the head scan image positioning without using additional equipment, and will not affect the image quality, improving the diagnostic accuracy.

[0050] Figure 3 Flowchart of the image position adjustment method provided by the embodiment of the present invention.

[0051] As Figure 3 shown, the image position adjustment method 300 includes steps S310 to S330.

[0052] Step S310: Based on the scanned head data, determine an attenuation image, where the head offset in the head data is the first offset.

[0053] Exemplarily, the head data is acquired from a PET detector by executing a head scanning protocol; the attenuation image includes a linear attenuation coefficient image, and there are multiple ways to estimate the attenuation image: one requires a CT image to generate an attenuation map; another does not require a CT image and uses iterative reconstruction algorithms such as MLAA (Morphological Anti-Aliasing) to simultaneously estimate the activity distribution map (PET image with incorrect positioning) and the attenuation map based on PET data; it can also come from other modality data, such as CT, MR (Magnetic Resonance Imaging), etc.

[0054] Step S320: Based on the attenuation image, determine image adjustment data.

[0055] Exemplarily, based on the obtained attenuation image μ, rotate it along the X-axis, Y-axis, and Z-axis to obtain a set of Euler angles, convert them into a rotation relationship, and based on this rotation relationship, obtain the adjustment data of the attenuation image.

[0056] The attenuation map μ quantitatively describes the ability of a certain substance per unit thickness to attenuate rays of a specific energy, and the unit is usually per centimeter (cm-1). Different human tissues, such as bones, muscles, and fats, have different attenuation abilities for rays due to differences in density, atomic composition, etc., corresponding to different attenuations. The PET image reflects tissue metabolic information, but the rays received by the PET detector have been attenuated when penetrating the human body. Using the attenuation map μ to correct the deviation caused by this attenuation can make the PET image more accurately present the true distribution of the radioactive tracer in the body and improve the diagnostic accuracy.

[0057] Step S330: Based on the image adjustment data and the head data, obtain a target PET image, where the head offset in the target PET image is the second offset, and the second offset is less than the first offset.

[0058] Exemplarily, based on the obtained image adjustment data and the scanned head data, establish a model according to the rotation relationship and use methods such as EM (Expectation-Maximization) to solve and obtain the target PET image.

[0059] When a PET device scans the head, data is acquired from different depth positions, as if the head is "cut" into many thin slices along a specific direction. The planar region where each thin slice is located is what is referred to as a "layer". In commonly used clinical PET devices, the layer thickness is generally between 2 and 10 millimeters. This layering can cover a certain range of the head, facilitating doctors to observe the head tissue conditions at different depth levels one by one.

[0060] In the above embodiments, based on the attenuation image, image adjustment data is determined. Based on the image adjustment data and the head data, a target PET image is obtained, realizing the positioning correction of the head PET image. When processing the attenuation image, in fact, the three-dimensional head image is cut into multiple layers (with a specific thickness) to obtain the image adjustment data for each layer. Finally, based on the image adjustment data and the head data, PET images of multiple layers are obtained and stitched together to obtain the target PET image. Through the above method, each layer of the obtained target PET image is axisymmetric, improving the image readability and enhancing the diagnostic accuracy.

[0061] Based on the attenuation image, determining the image adjustment data includes: based on the attenuation image, obtaining the preliminary direction information of the head and the reference direction information of the head; based on the preliminary direction information of the head and the reference direction information of the head, adjusting the attenuation image to obtain the image adjustment data.

[0062] The preliminary direction information of the head, for example, indicates that the attenuation image is not axisymmetric about the initial direction in three-dimensional space and has an offset, and the reference direction information of the head, for example, indicates that the attenuation image is axisymmetric about the initial direction and has no offset, etc.

[0063] Specifically, based on the attenuation image, obtaining the preliminary direction information of the head and the reference direction information of the head includes: detecting the cranial vertex region and the cranial vertex in the attenuation image; based on the normal vector direction of the cranial vertex region and the cranial vertex, obtaining the preliminary direction information of the head; based on the preliminary direction information of the head, obtaining the reference direction information of the head.

[0064] The shape near the cranial vertex region is relatively flat, and for the attenuation image, it is relatively easy to determine. Therefore, the present invention first determines the cranial vertex region based on the attenuation image and determines the cranial vertex based on the center of the cranial vertex region.

[0065] Exemplarily, detecting the cranial vertex region and the cranial vertex in the attenuation image includes: performing image segmentation on the attenuation image to obtain the skull region; detecting the frontal sinus position in the skull region; based on the frontal sinus position, determining the cranial vertex region from the skull region; based on the central position of the cranial vertex region, obtaining the cranial vertex.

[0066] The skull region can be segmented from the attenuation image using: threshold segmentation method, region growing method, edge detection-based algorithm, or machine learning algorithm. The threshold segmentation method is based on the difference in the gray values of the image pixels, and a suitable threshold is set. Since the attenuation of the skull is significantly different from that of the surrounding soft tissues, brain tissues, etc., when the pixel gray value is higher (or lower) than a specific threshold, it is determined to belong to the skull region; the region growing method first selects several seed points within the skull region. These seed points can be manually selected or automatically located based on some prior knowledge. For example, the points with known high attenuation coefficients in the image are selected. Then, according to the pre-set growth criteria, such as pixel gray similarity, texture similarity, etc., the surrounding eligible pixels are continuously incorporated into the skull region, gradually expanding the segmentation range like "snowballing" until all the connected pixels belonging to the skull are included. The edge detection-based algorithm utilizes the gray mutation between the skull and the surrounding tissues. Through edge detection operators such as Sobel and Canny, the edge information in the image is extracted. Then, morphological processing, contour tracking, etc. are performed on these edges to outline the contour of the skull, thus completing the segmentation. Machine learning algorithms, such as using convolutional neural networks (CNNs), first collect a large number of labeled attenuation images and corresponding skull segmentation labels as training data, and let the network learn the characteristics of the skull region. The trained model can then predict new unlabeled images and output the segmented skull region. U-Net is a commonly used CNN architecture for medical image segmentation. It has an encoder and a decoder structure. The encoder extracts image features, and the decoder restores the image resolution, thus accurately locating the skull.

[0067] To detect the position of the frontal sinus in the skull region, it can be based on traditional image processing algorithms such as: morphological operation and template matching. First, the skull image is preprocessed using morphological erosion, dilation, etc. operations to strengthen the boundary features between the frontal sinus region and the surrounding tissues. For example, the erosion operation can remove small noise points in the image and thin the boundary, while dilation does the opposite, filling holes and connecting discontinuous boundaries. Then, the template matching technique is adopted. Standard templates of the frontal sinus at different angles and sizes are prepared in advance, and these templates are slid and compared on the preprocessed image to calculate the matching degree. When the matching degree exceeds the set threshold, it is determined that the position of the frontal sinus has been found. Or based on edge detection and Hough transform. First, edge detection operators such as Sobel and Canny are used to extract the edge information of the skull image. Due to its unique anatomical structure, the boundary of the frontal sinus region often shows specific gray value changes, and corresponding contours will appear on the edge map. Then, the Hough transform is applied. This method is good at converting geometric shapes such as straight lines and curves in the image space to the parameter space and detecting the curves that conform to the shape characteristics of the frontal sinus boundary, thereby locating the frontal sinus. For example, the edge part of the frontal sinus may be approximately elliptical, and the Hough transform can detect the elliptical contour in the image to lock the position of the frontal sinus.

[0068] Among them, based on the position of the frontal sinus, the cranial vertex region is determined from the cranial bone region, including: determining the outer surface curve of the cranial bone from the cranial bone region based on the position of the frontal sinus; determining a plurality of local curves from the outer surface curve of the cranial bone; respectively determining the overall curvature of each local curve in the plurality of local curves; and determining, from the plurality of local curves, the local curve whose overall curvature satisfies a preset curvature condition as the cranial vertex region.

[0069] For each local curve, the principal curvature at each position of each local curve is calculated respectively, and the average value of the larger values among all the principal curvatures is used as the overall curvature of the local curve. Among the overall curvatures of all the local curves, the smallest one is the cranial vertex region, and the center of the curve is the vertex of the brain.

[0070] Figure 4 This is the distribution map of the cranial vertex bone of the embodiment of the present invention.

[0071] For example, as Figure 4 shown, the position pointed by the arrow is the position of the frontal sinus, the red curve is the outer surface curve Ω(x, y, z) of the cranial bone, and the blue dot represents the vertex of the brain, that is, the center of the cranial vertex region. The cranial bone is segmented from the attenuation map μ, and the curve Ω(x, y, z) of the outer surface of the cranial bone is generated. The range of this curve is restricted. By detecting the position of the frontal sinus of the cranial bone, only all the layers above the frontal sinus are considered to form the outer surface curve Ω(x, y, z) of the cranial bone, that is, the middle layer part between the cranial vertex and the frontal sinus ( Figure 4 the two dashed lines in). A sliding window W is set. This window is a circle on the XOY plane, and the local curve formed on the outer surface curve Ω(x, y, z) of the cranial bone is: Ω′ = {p(x, y, z)|(x, y) ∈ W, p ∈ Ω}. The principal curvature at each position in each small curve formed by the sliding window is calculated, and the average value of the larger values among all the principal curvatures is used as the overall curvature of the local curve. Among all the local curves, the one with the smallest overall curvature is considered to be the cranial vertex region. The center of this curve is the vertex of the brain.

[0072] Exemplarily, based on the normal vector direction of the cranial vertex region and the vertex of the brain, preliminary direction information of the head is obtained, including: starting from the vertex of the brain and extending along the normal vector direction of the cranial vertex region to obtain the preliminary direction information of the head. Based on the preliminary direction information of the head, reference direction information of the head is obtained, including: according to the minimum moment of inertia method, rotating the attenuation image based on the preliminary direction information of the head to obtain the reference direction information of the head.

[0073] Among them, since the axis of symmetry with the minimum moment of inertia needs to pass through the center of mass of the object, rotating the attenuation image based on the preliminary head direction information to obtain the reference direction information of the head may involve translation. For example, if the preliminary head direction information does not pass through the center of mass of the head, before rotating the attenuation image, the attenuation image can be moved (such as translated) so that the preliminary head direction information passes through the center of mass of the head.

[0074] For example, starting from the vertex of the skull, along the direction of the normal vector of the cranial vertex region, a range of the top of the skull is selected. This normal vector is the preliminary head direction information (such as the preliminary Z-direction information of the head). Then, based on the skull distribution within this range, the minimum moment of inertia method is used to calculate the reference direction information of the head (such as the reference Z-direction information of the head).

[0075] For the minimum moment of inertia method, the moment of inertia is a physical quantity that measures the inertia of a rigid body rotating about an axis, and its value depends on the mass distribution of the rigid body and the position of the axis of rotation. For a given object, when it rotates about different axes, the moment of inertia will be different. According to relevant theories such as the parallel axis theorem, there is a specific axis of rotation that makes the moment of inertia of the object about this axis reach the minimum value. The calculation of the moment of inertia I usually involves the integral of the mass element dm of the object and the square of its distance r from the axis of rotation, that is, i = ∫r 2 dm. By changing the assumed position of the axis of rotation, the axis of rotation that makes the integral value minimum is solved. This process uses mathematical means of finding extreme values, such as taking the derivative and setting the derivative to zero.

[0076] In the above embodiments, through the shape characteristics of the skull, the cranial vertex region, the vertex of the skull, and the position of the frontal sinus of the skull are determined, and the preliminary head direction information is obtained by extending along the normal vector direction of the cranial vertex region; according to the minimum moment of inertia method, the attenuation image is rotated based on the preliminary head direction information to obtain the reference direction information of the head. The above method is relatively easier to determine because the shape near the center of the cranial vertex is relatively flat, and it is more convenient; and the minimum moment of inertia method has a low calculation complexity and higher calculation efficiency.

[0077] Figure 5 This is a flowchart of the method for determining the reference Z-direction axis of symmetry provided by the embodiments of the present invention.

[0078] As Figure 5 shown, the method 500 for determining the reference Z-direction axis of symmetry includes steps S510 to S550.

[0079] Step S510, segment the skull.

[0080] Step S520, detect the position of the frontal sinus and form the cranial vertex surface.

[0081] Step S530: Set a sliding window determined by XY coordinates and calculate the average value of the maximum principal curvature within the sliding window.

[0082] Step S540: The center position of the ROI (Region of Interest) with the minimum principal curvature is considered as the cranial vertex position. Along the direction of the normal vector of the cranial vertex position, select a section of the upper skull.

[0083] Step S550: Calculate the reference Z axis of symmetry (reference direction information of the head).

[0084] Figure 6 Schematic diagram of rotation based on the X axis and Y axis provided by the embodiment of the present invention Figure 7 Schematic diagram of rotation based on the Z axis provided by the embodiment of the present invention.

[0085] The image adjustment data includes a rotation matrix. Based on the preliminary direction information of the head and the reference direction information of the head, the attenuation image is adjusted to obtain the image adjustment data, including: rotating the attenuation image based on the preliminary direction information of the head and the reference direction information of the head to obtain the rotation matrix.

[0086] Among them, before rotation, translation may be involved. Because the axis of symmetry with the minimum moment of inertia needs to pass through the centroid of the object. For example, if the preliminary direction information of the head does not pass through the centroid of the head, before rotating the attenuation image, the attenuation image can be moved (such as translated) so that the preliminary direction information of the head passes through the centroid of the head.

[0087] The rotation matrix is a three-dimensional rotation matrix (3x3), which rotates around the X axis, Y axis, and Z axis respectively to obtain the corresponding rotation angles, and a three-dimensional rotation matrix is obtained through specific transformation based on the rotation angles. For example, the three rotation matrices obtained by rotating around the X axis, Y axis, and Z axis are all in the 3x3 form, respectively representing the transformation of coordinate points in three-dimensional space when rotating around a single coordinate axis.

[0088] Exemplarily, rotating the attenuation image based on the preliminary direction information of the head and the reference direction information of the head to obtain the rotation matrix includes: rotating the attenuation image to align the preliminary direction information of the head and the reference direction information of the head to obtain the rotation matrix.

[0089] Rotating the attenuation image to align the preliminary direction information of the head and the reference direction information of the head, and the rotation includes rotating around the X axis, Y axis, or Z axis to align the preliminary Z direction axis of symmetry with the reference Z direction axis of symmetry.

[0090] Specifically, the preliminary direction information of the head includes a preliminary Z-direction symmetry axis, and the reference direction information of the head includes a reference Z-direction symmetry axis. In the image coordinate system, the preliminary Z-direction is the Z-axis of the image coordinate system, and the image coordinate system also includes an X-axis and a Y-axis; rotate the attenuation image so that the preliminary direction information of the head is aligned with the reference direction information of the head, and obtain a rotation matrix, including: rotating the attenuation image so that the preliminary Z-direction symmetry axis is aligned with the reference Z-direction symmetry axis, and obtaining at least one of an X-axis rotation angle component, a Y-axis rotation angle component, and a Z-axis rotation angle component, where rotating the attenuation image includes rotating the attenuation image around at least one of the X-axis, Y-axis, and Z-axis; based on at least one of the X-axis rotation angle component, the Y-axis rotation angle component, and the Z-axis rotation angle component, obtain the rotation matrix.

[0091] Wherein, the symmetry axis mentioned in the present invention is a straight line passing through a certain point in space, not a vector that can be arbitrarily moved.

[0092] The rotation matrix can be determined in various ways, such as: Euler angle method, quaternion method, axis-angle representation method and Rodriguez formula or rotation vector method. In this article, it is explained in the way of Euler angles.

[0093] In three-dimensional space, there are usually 12 different rotation sequences for Euler angles. Usually, a fixed coordinate system (inertial coordinate system) and a coordinate system fixed to the rigid body are defined first. Euler angles make the rigid body coordinate system coincide with the inertial coordinate system through three consecutive basic rotations. Each rotation corresponds to an Euler angle. The most commonly used is to rotate in the order of "Z-Y-X", that is, the first rotation is around the Z-axis, controlling the left and right rotation of the rigid body around the vertical axis, also called the yaw angle; the second rotation is around the Y-axis, used to adjust the attitude of the rigid body to look up and down, called the pitch angle; the third rotation is around the X-axis, and the rigid body rolls around the axis in its own forward direction, also called the roll angle.

[0094] In this article, when implementing, the rotation matrix is obtained by rotating in the order of "X-Y-Z". Let the Euler angles be α (rotation angle around the X-axis), β (rotation angle around the Y-axis), and γ (rotation angle around the Z-axis) respectively. Rotate around the X-axis, Y-axis, and Z-axis respectively to obtain three rotation matrices M x 、M y 、M z , and multiply the three rotation matrices from right to left according to the rotation order to obtain the final rotation matrix M. Of course, in some methods, if the alignment can be achieved by rotating around the X-axis, there is no need to rotate around the Y-axis and Z-axis anymore. If the alignment can be achieved by rotating around the X-axis and Y-axis, there is no need to rotate around the Z-axis anymore.

[0095] For example, as Figure 6As shown, the X-axis, Y-axis, and Z-axis are preliminary direction information. X′, Y′, and Z′ are the information obtained after the first rotation of the attenuation image (rotation around the X-axis), and X″, Y″, and Z″ are the information obtained after the second rotation of the attenuation image (rotation around the Y-axis). The two rotations can respectively obtain the reference Z direction A z The rotation angle components between the attenuation image μ are respectively the X-axis rotation angle components Y-axis rotation angle components Rotate the attenuation map μ around the image center (the image center is the coordinate origin) to obtain the attenuation map μ′ that is symmetric in the Z direction.

[0096] As Figure 7 shown, based on the attenuation map μ′ that is symmetric in the Z direction, select the middle layer part between the cranial vertex and the frontal sinus ( Figure 4 the two dotted lines in the figure), with the X-axis or Y-axis as the initial value, rotate around the reference Z direction A z Perform rotation. Based on the minimum moment of inertia method, the reference X direction, reference Y direction, and Z-axis rotation angle components can be calculated

[0097] Based on the obtained set of rotation angle components Convert it into a rotation matrix M, that is, obtain the image adjustment data.

[0098] In the above embodiment, the method of first calculating the symmetry axis in the Z direction of the object and then calculating the symmetry axes in the XY directions of the object determines a new coordinate system. This method makes the PET image axisymmetric by establishing a new coordinate system, does not require additional equipment, and reduces costs.

[0099] In another example, use other tissue organs outside the skull to calculate the symmetry axes in three directions; or use MR (Magnetic Resonance Imaging) or other modality devices to provide human body structure information and calculate the symmetry axes in three directions.

[0100] Based on the image adjustment data and the head data, obtain the target PET image, including: based on the image adjustment data, perform image reconstruction on the head data to obtain the target PET image.

[0101] Exemplarily, based on the image adjustment data, perform image reconstruction on the head data to obtain the target PET image, including: based on the image adjustment data, head data, scanning system matrix, and scanning system noise, perform image reconstruction to obtain the target PET image.

[0102] For example, for PET reconstruction, there is usually a relationship as shown in formula (1):

[0103] Y = AX + B (1)

[0104] Where Y is the PET data, X is the PET image, A is the system matrix of PET, including geometric factors, detector efficiency, attenuation correction coefficient, decay coefficient, count loss coefficient, etc.; B is the noise in the system, including random coincidence events, scattered coincidence events, system noise, etc.

[0105] In this embodiment, the rotation matrix M is incorporated into the reconstruction model, and a geometric rotation relationship is added to the original system matrix, and the relationship shown in formula (2) can be obtained:

[0106] Y = AMX + B (2)

[0107] Based on the relationship shown in formula (2), methods such as EM (Expectation-Maximization) are used to solve for X to obtain the target PET image.

[0108] In the above embodiment, the rotation matrix is incorporated into the reconstruction model, and a geometric rotation relationship is added to the original system matrix, avoiding the problem of image clarity degradation caused by rotating interpolation of the PET image.

[0109] Figure 8 It is a flowchart for implementing the detailed image position adjustment method provided by the embodiment of the present invention.

[0110] As Figure 8 shown, the detailed image position adjustment method 800 includes steps S810 to S860.

[0111] Step S810, collect PET data.

[0112] Step S820, obtain the attenuation map μ.

[0113] Exemplarily, based on the collected PET data, the radionuclide distribution map Y0 and the attenuation map μ are estimated. Where Y0 is the PET image, reflecting the distribution of radionuclides in the object to be scanned; μ is the attenuation map, indicating the attenuation corresponding to each pixel, and the higher the density, the greater the attenuation.

[0114] Step S830, calculate the reference Z-direction symmetry axis (reference direction information of the head) A according to the distribution of the skull on the head Z .

[0115] Step S840, with the image center as the rotation center, align the Z-axis of the attenuation map μ with A Z aligned.

[0116] Step S850, calculate the symmetry axis A in the rotated XOY plane xy (reference direction information of the head) can determine a set of Euler angles.

[0117] Step S860: Calculate the rotation matrix M through Euler angles and incorporate M into the PET image reconstruction to obtain a correctly positioned PET image.

[0118] Based on the image adjustment data and the head data, obtain the target PET image, including: obtaining an initial PET image based on the head data; adjusting the initial PET image based on the image adjustment data to obtain the target PET image.

[0119] For example, estimate the activity distribution map (the initial PET image with incorrect positioning) and the attenuation map μ based on the PET data, calculate the rotation matrix according to the above method based on the attenuation map μ, and directly adjust the initial PET image based on the rotation matrix to obtain the target PET image.

[0120] The image position adjustment method proposed by the present invention can correct the positioning of the head PET image in a PET device without relying on external image acquisition devices such as CT / MR / cameras, etc. Under the condition that the patient's head is slightly tilted, automatically correct the central axis of the image so that each layer of the PET image is axisymmetric, improving the image readability and enhancing the doctor's confidence in diagnosis.

[0121] Figure 9 It is a block diagram of an image position adjustment device provided in another embodiment of the present invention.

[0122] An embodiment of the present invention provides an image position adjustment device 900. Please refer to Figure 9 , the image position adjustment device 900 includes: a first determination module 910, a second determination module 920, and an acquisition module 930.

[0123] Exemplarily, the first determination module 910 is used to determine the attenuation image based on the scanned head data, where the head offset in the head data is the first offset.

[0124] Exemplarily, the second determination module 920 is used to determine the image adjustment data based on the attenuation image.

[0125] Exemplarily, the acquisition module 930 is used to obtain the target PET image based on the image adjustment data and the head data, where the head offset in the target PET image is the second offset, and the second offset is less than the first offset.

[0126] It can be understood that for the specific description of the image position adjustment device 900, reference can be made to the description of the image position adjustment method in the above text, which will not be elaborated here.

[0127] Figure 10 It is a block diagram of an electronic device provided in an embodiment of the present invention.

[0128] An embodiment of the present application provides an electronic device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above-mentioned method is implemented.

[0129] As Figure 10 shown, for ease of understanding, an embodiment of the present application shows a specific electronic device 1000.

[0130] The electronic device 1000 is intended to represent various forms of digital computers, such as, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, personal digital processors, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0131] As Figure 10 shown, the device 1000 includes a computing unit 1001, which can execute various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 802 or the computer program loaded from the storage unit 1008 into the random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation of the electronic device 1000 can also be stored. The computing unit 1001, the ROM 1002, and the RAM 1003 are connected to each other through a bus 1004. The input / output (I / O) interface 1005 is also connected to the bus 1004.

[0132] Multiple components in the electronic device 1000 are connected to the I / O interface 1005. The multiple components include: an input unit 1006, such as a keyboard, a mouse, etc.; an output unit 1007, such as various types of displays, speakers, etc.; a storage unit 1008, such as a magnetic disk, an optical disc, etc.; and a communication unit 1009, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1009 allows the electronic device 1000 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0133] The computing unit 1001 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1001 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1001 executes the various methods described above. For example, in some embodiments, any one or more of the above-described methods may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 1008. In some embodiments, part or all of the computer program may be loaded and / or installed onto the electronic device 1000 via the ROM 1002 and / or the communication unit 1009. When the computer program is loaded into the RAM 1003 and executed by the computing unit 1001, one or more steps of any one or more of the above-described methods may be executed. Alternatively, in other embodiments, the computing unit 1001 may be configured to execute any one or more of the above-described methods in any other suitable manner (e.g., by means of firmware).

[0134] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method according to any one of the above embodiments are implemented.

[0135] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in combination with these instruction execution systems, apparatuses, or devices. For the purposes of the present invention, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in combination with an instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of the computer-readable medium include the following: an electrical connection part (electronic device) having one or more wirings, a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.

[0136] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), and the like.

[0137] In the description of the present invention, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In the present invention, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0138] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present invention.

[0139] In addition, the terms "first", "second", etc. used in the embodiments of the present invention are only for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated in this embodiment. Thus, the features defined with the terms "first", "second", etc. in the embodiments of the present invention may explicitly or implicitly indicate that at least one such feature is included in this embodiment. In the description of the present invention, the meaning of the word "plurality" is at least two or more, such as two, three, four, etc., unless otherwise specifically defined in the embodiment.

[0140] In the present invention, unless otherwise clearly specified or limited in the embodiments, the terms "installation", "connection", "connection" and "fixation" and the like appearing in the embodiments should be understood in a broad sense. For example, the connection can be a fixed connection, a detachable connection, or integrated. It can be understood that it can also be a mechanical connection, an electrical connection, etc.; of course, it can also be directly connected, or indirectly connected through an intermediate medium, or it can be the communication inside two elements, or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific implementation situations.

[0141] In the present invention, unless otherwise clearly specified and limited, the first feature being "on" or "under" the second feature may be that the first and second features are in direct contact, or the first and second features are indirectly in contact through an intermediate medium. Moreover, the first feature being "above", "over" and "on top of" the second feature may be that the first feature is directly above or obliquely above the second feature, or merely indicates that the first feature has a higher horizontal height than the second feature. The first feature being "under", "below" and "beneath" the second feature may be that the first feature is directly below or obliquely below the second feature, or merely indicates that the first feature has a lower horizontal height than the second feature.

[0142] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for adjusting image position, characterized in that: The method comprises: Determine an attenuation image based on the head data obtained by scanning, wherein the head offset in the head data is a first offset; Based on the attenuation image, determining image adjustment data; A target PET image is obtained based on the image adjustment data and the head data, wherein the head offset in the target PET image is a second offset, and the second offset is smaller than the first offset.

2. The method according to claim 1, characterized in that: The step of determining image adjustment data based on the attenuation image comprises: Based on the attenuation image, obtaining preliminary direction information of the head and reference direction information of the head; The attenuated image is adjusted based on the preliminary direction information of the head and the reference direction information of the head to obtain the image adjustment data.

3. The method according to claim 2, characterized in that The obtaining of preliminary direction information of the head and reference direction information of the head based on the attenuation image includes: Detecting the skull top area and the skull vertex in the attenuation image; Based on the normal vector direction of the cranial vertex area and the cranial vertex, obtaining preliminary direction information of the head; Based on the preliminary direction information of the head, reference direction information of the head is obtained.

4. The method according to claim 3, characterized in that The detecting of the skull top area and the cranial vertex in the attenuation image comprises: Performing image segmentation on the attenuation image to obtain a skull region; Detecting the skull region to obtain the position of the frontal sinus; determining the skull vertex region from the skull region based on the frontal sinus position; Based on the center position of the skull top area, the skull top point is obtained.

5. The method according to claim 4, characterized in that The step of determining the skull top region from the skull region based on the frontal sinus position comprises: determining a skull outer surface curve from the skull region based on the frontal sinus position; Determine a plurality of local surfaces from the outer surface of the skull; respectively determining the overall curvature of each of the plurality of local curved surfaces; From the multiple local curved surfaces, a local curved surface whose overall curvature satisfies a preset curvature condition is determined as the skull top region.

6. The method according to claim 3, characterized in that The obtaining of preliminary direction information of the head based on the normal vector direction of the skull vertex region and the skull vertex includes: Starting from the cranial vertex, the method extends along the normal vector direction of the cranial vertex region to obtain preliminary direction information of the head.

7. The method according to claim 3, characterized in that The obtaining the reference direction information of the head based on the preliminary direction information of the head includes: According to the minimum moment of inertia method, the attenuation image is rotated based on preliminary direction information of the head to obtain reference direction information of the head.

8. The method according to claim 2, characterized in that: The image adjustment data includes a rotation matrix, and the attenuation image is adjusted based on the preliminary direction information of the head and the reference direction information of the head to obtain the image adjustment data, including: The attenuation image is rotated based on the preliminary direction information of the head and the reference direction information of the head to obtain the rotation matrix.

9. The method according to claim 8, characterized in that The step of rotating the attenuated image based on the preliminary direction information of the head and the reference direction information of the head to obtain the rotation matrix includes: The attenuation image is rotated to align the preliminary direction information of the head with the reference direction information of the head, thereby obtaining the rotation matrix.

10. The method according to claim 9, characterized in that The preliminary direction information of the head includes a preliminary Z-direction symmetry axis, the reference direction information of the head includes a reference Z-direction symmetry axis, in an image coordinate system, the preliminary Z direction is the Z axis of the image coordinate system, and the image coordinate system also includes an X axis and a Y axis; the attenuation image is rotated so that the preliminary direction information of the head is aligned with the reference direction information of the head to obtain the rotation matrix, including: The attenuation image is rotated so that the preliminary Z-direction symmetry axis is aligned with the reference Z-direction symmetry axis to obtain at least one of an X-axis rotation angle component, a Y-axis rotation angle component, and a Z-axis rotation angle component, Wherein, rotating the attenuation image comprises rotating the attenuation image around at least one of an X-axis, a Y-axis, and a Z-axis; The rotation matrix is ​​obtained based on at least one of the X-axis rotation angle component, the Y-axis rotation angle component, and the Z-axis rotation angle component.

11. The method according to any one of claims 1 to 10, characterized in that: The step of obtaining a target PET image based on the image adjustment data and the head data includes: Based on the image adjustment data, image reconstruction is performed on the head data to obtain the target PET image.

12. The method according to claim 11, characterized in that The step of reconstructing the head data based on the image adjustment data to obtain the target PET image includes: Image reconstruction is performed based on the image adjustment data, the head data, the scanning system matrix and the scanning system noise to obtain a target PET image.

13. The method according to any one of claims 1 to 10, characterized in that: The step of obtaining a target PET image based on the image adjustment data and the head data includes: Obtaining an initial PET image based on the head data; The initial PET image is adjusted based on the image adjustment data to obtain the target PET image.

14. An image position adjustment device, characterized in that: The device comprises: A first determination module is used to determine an attenuation image based on head data obtained by scanning, wherein a head offset in the head data is a first offset; A second determination module, configured to determine image adjustment data based on the attenuation image; An acquisition module is used to obtain a target PET image based on the image adjustment data and the head data, wherein the head offset in the target PET image is a second offset, and the second offset is smaller than the first offset.

15. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 13 are implemented.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 13 is implemented.

Images