Method and apparatus for positioning slice image reconstruction, and computed tomography apparatus

Abstract

The application provides a positioning film image reconstruction method and device and a computer tomography equipment. The method comprises the following steps: acquiring a first local projection image and a second local projection image of a positioning film, the first local projection image comprising an overlapping area overlapping with the second local projection image; determining a plurality of first frequency band images of the first local projection image and a plurality of second frequency band images of the second local projection image, the number of the plurality of first frequency band images being the same as the number of the plurality of second frequency band images; determining a plurality of weighted mask frequency band images of the overlapping area; fusing the plurality of first frequency band images and the plurality of second frequency band images based on the plurality of weighted mask frequency band images to obtain a plurality of fused frequency band images; and determining a positioning film image based on the plurality of fused frequency band images. The application realizes the matching of the image size and the image scale feature, avoids the occurrence of ghosting and truncation, and improves the image quality of the determined positioning film image.

Description

Technical Field

[0001] This invention relates to the field of medical imaging, specifically to a method, apparatus, and computed tomography (CT) scanner for reconstructing localization images. Background Technology

[0002] Micro Computed Tomography (MicroCT), also known as miniature CT or microscopic CT, is a non-destructive 3D imaging technique that can clearly reveal the internal microstructure of a sample without damaging it, offering extremely high resolution. Before performing a computed tomographic scan in MicroCT, a localization scan is performed to obtain localization images. Since the length of the object being scanned often exceeds the detector's coverage area, current MicroCT techniques typically perform localization scans of long objects on multiple beds, obtaining multiple local projection images of the localization image. There is an overlap between adjacent local projection images, allowing multiple local projection images to be fused (reconstructed) based on the overlapping areas to obtain the localization image.

[0003] The existing technology for fusing multiple local projection images uses a pixel-level fusion method, which directly fuses the pixels in the overlapping area. This existing technology has the following technical problems: because the local projection images of a positioning patch contain features of multiple different scales with a large spectral span, using a pixel-level fusion method leads to a mismatch between the size of the overlapping area and the feature scale in the image. This results in ghosting or truncation in the fused positioning patch image, leading to poor image quality. Summary of the Invention

[0004] In view of this, it is necessary to provide a method, apparatus and computed tomography (CT) scanning device for reconstructing localization images, in order to solve the technical problem that the use of pixel-level fusion in the prior art results in ghosting or truncation in the fused localization images, thus leading to poor image quality.

[0005] On one hand, the present invention provides a method for reconstructing a localization patch image, comprising:

[0006] A first local projection image and a second local projection image of the positioning patch are obtained, wherein the first local projection image includes an overlapping region that overlaps with the second local projection image;

[0007] A plurality of first frequency band images of the first local projection image and a plurality of second frequency band images of the second local projection image are determined, wherein the number of the plurality of first frequency band images is the same as the number of the plurality of second frequency band images;

[0008] Determine multiple weighted mask frequency band images of the overlapping region;

[0009] Based on the multiple weighted mask frequency band images, the multiple first frequency band images and the multiple second frequency band images are fused to obtain multiple fused frequency band images;

[0010] The positioning patch image is determined based on the multiple fused frequency band images.

[0011] In some possible implementations, determining the plurality of first frequency band images of the first local projection image and the plurality of second frequency band images of the second local projection image includes:

[0012] Construct a first Laplacian pyramid of the first local projection image and a second Laplacian pyramid of the second local projection image to obtain the plurality of first frequency band images and the plurality of second frequency band images accordingly;

[0013] or,

[0014] The first local projection image and the second local projection image are filtered by a high-pass filter respectively to obtain the plurality of first frequency band images and the plurality of second frequency band images.

[0015] or,

[0016] Principal component analysis is used to decompose the first local projection image and the second local projection image to obtain the plurality of first frequency band images and the plurality of second frequency band images.

[0017] or,

[0018] The first local projection image and the second local projection image are decomposed based on the wavelet transform method to obtain the plurality of first frequency band images and the plurality of second frequency band images.

[0019] In some possible implementations, constructing the first Laplacian pyramid of the first local projected image includes:

[0020] The first local projection image is decomposed into a Gaussian pyramid to obtain the first Gaussian pyramid.

[0021] The first Laplace pyramid is determined based on the first Gaussian pyramid;

[0022] The construction of the first Laplacian pyramid of the first local projection image includes:

[0023] The second local projection image is decomposed into a Gaussian pyramid to obtain the second Gaussian pyramid.

[0024] The second Laplace pyramid was determined based on the second Gaussian pyramid.

[0025] In some possible implementations, the first Gaussian pyramid includes a first layer of Gaussian images, a second layer of Gaussian images, and a third layer of Gaussian images;

[0026] The step of performing Gaussian pyramid decomposition on the first local projection image to obtain the first Gaussian pyramid includes:

[0027] The first local projection image is used as the first layer Gaussian image;

[0028] The first local projection image is subjected to Gaussian filtering based on preset Gaussian filtering parameters to obtain the first preprocessed projection image.

[0029] The first preprocessed projection image is downsampled based on preset downsampling parameters to obtain the second layer Gaussian image;

[0030] The second layer Gaussian image is subjected to Gaussian filtering based on preset Gaussian filtering parameters to obtain the second preprocessed projection image.

[0031] The second preprocessed projection image is downsampled based on preset downsampling parameters to obtain the third layer Gaussian image.

[0032] In some possible implementations, the first Laplacian pyramid includes a first layer of Laplacian images, a second layer of Laplacian images, and a third layer of Laplacian images;

[0033] The determination of the first Laplace pyramid based on the first Gaussian pyramid includes:

[0034] Upsample the second-layer Gaussian image and the third-layer Gaussian image to obtain the second-layer upsampled image and the third-layer upsampled image respectively;

[0035] Gaussian convolution is performed on the second-layer upsampled image and the third-layer upsampled image to obtain the second-layer Gaussian convolution image and the third-layer Gaussian convolution image respectively;

[0036] The difference between the first Gaussian image and the second Gaussian convolution image is used as the first Laplacian image.

[0037] The difference between the second Gaussian image and the third Gaussian convolution image is used as the second Laplacian image.

[0038] The third Gaussian image is used as the third Laplacian image.

[0039] In some possible implementations, determining the multiple weighted mask frequency band images of the overlapping region includes:

[0040] Determine the initial weighted mask image;

[0041] Gaussian pyramid decomposition is performed on the initial weighted mask image to obtain the multiple weighted mask frequency band images;

[0042] or,

[0043] The initial weighted mask is filtered using a high-pass filter to obtain the frequency band images of the multiple weighted masks;

[0044] or,

[0045] The initial weighted mask is decomposed based on principal component analysis to obtain the frequency band images of the multiple weighted masks;

[0046] or,

[0047] The initial weighted mask is decomposed using wavelet transform to obtain the frequency band images of the multiple weighted masks.

[0048] In some possible implementations, the plurality of weighted mask frequency band images include a first weighted mask frequency band image, a second weighted mask frequency band image, and a third weighted mask frequency band image; the second Laplacian pyramid includes a first adjacent layer Laplacian image, a second adjacent layer Laplacian image, and a third adjacent layer Laplacian image, respectively corresponding to the first layer Laplacian image, the second layer Laplacian image, and the third layer Laplacian image;

[0049] The process of fusing the multiple first frequency band images and the multiple second frequency band images based on the multiple weighted mask frequency band images to obtain multiple fused frequency band images includes:

[0050] The first layer fused Laplacian image is obtained based on the first weighted mask frequency band image, the first layer Laplacian image, the first adjacent layer Laplacian image, and the preset fusion model.

[0051] The second-layer fused Laplacian image is obtained based on the second weighted mask frequency band image, the second-layer Laplacian image, the second adjacent-layer Laplacian image, and the preset fusion model;

[0052] The third-layer fused Laplacian image is obtained based on the third weighted mask frequency band image, the third-layer Laplacian image, the third adjacent layer Laplacian image, and the preset fusion model.

[0053] The fused Laplacian pyramid is obtained based on the first layer fused Laplacian image, the second layer fused Laplacian image, and the third layer fused Laplacian image.

[0054] In some possible implementations, determining the location patch image based on the multiple fused frequency band images includes:

[0055] The localization patch image is determined based on the first layer fused Laplacian image, the second layer fused Laplacian image, and the third layer fused Laplacian image;

[0056] or,

[0057] Determine the first weight of the first layer fused Laplacian image, the second weight of the second layer fused Laplacian image, and the third weight of the third layer fused Laplacian image;

[0058] The localization patch image is determined based on the first weight, the second weight, the third weight, the first layer fused Laplacian image, the second layer fused Laplacian image, and the third layer fused Laplacian image.

[0059] On the other hand, the present invention also provides a positioning patch image reconstruction apparatus, comprising:

[0060] A local projection image acquisition unit is used to acquire a first local projection image and a second local projection image of the positioning patch, wherein the first local projection image includes an overlapping region that overlaps with the second local projection image;

[0061] A local projection image frequency division unit is used to determine a plurality of first frequency band images of the first local projection image and a plurality of second frequency band images of the second local projection image, wherein the number of the plurality of first frequency band images is the same as the number of the plurality of second frequency band images;

[0062] The overlapping region frequency division unit is used to determine multiple weighted mask frequency band images of the overlapping region;

[0063] A frequency band fusion unit is used to fuse the multiple first frequency band images and the multiple second frequency band images based on the multiple weighted mask frequency band images to obtain multiple fused frequency band images;

[0064] The positioning image determination unit is used to determine the positioning image based on the multiple fused frequency band images.

[0065] On the other hand, the present invention also provides a computed tomography (CT) scanning device, including a memory and a processor, wherein,

[0066] The memory is used to store programs;

[0067] The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the localization patch image reconstruction method described in any of the above possible implementations.

[0068] The beneficial effects of the above embodiments are as follows: The positioning patch image reconstruction method provided by the present invention, after obtaining the first local projection image and the second local projection image of the positioning patch, determines multiple first frequency band images of the first local projection image and multiple second frequency band images of the second local projection image. Since different frequency band images represent features of different scales of the local projection image, and the image size of each layer is also different, the matching of image size and image scale features is realized. That is, large-scale features are matched with larger image sizes to avoid highlighting detailed features in larger image sizes, thereby preventing ghosting. Small-scale features are matched with smaller image sizes to avoid loss of detailed features, thereby preventing truncation. This improves the image quality of the determined positioning patch image. Attached Figure Description

[0069] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0070] Figure 1 A schematic flowchart of an embodiment of the positioning patch image reconstruction method provided by the present invention;

[0071] Figure 2 For the present invention Figure 1 A schematic diagram of an embodiment of S102;

[0072] Figure 3 For the present invention Figure 2 A schematic diagram of an embodiment of S201;

[0073] Figure 4 For the present invention Figure 2 Another embodiment of S202 is illustrated in the flowchart.

[0074] Figure 5 For the present invention Figure 1 A schematic diagram of an embodiment of S103;

[0075] Figure 6 For the present invention Figure 1 A schematic diagram of an embodiment of S104;

[0076] Figure 7 For the present invention Figure 1A schematic diagram of an embodiment of S105;

[0077] Figure 8 For the present invention Figure 1 Another embodiment of the process diagram of S105;

[0078] Figure 9 For the present invention Figure 8 A schematic diagram of an embodiment of S802;

[0079] Figure 10 A schematic diagram of an embodiment of the positioning patch image reconstruction device provided by the present invention;

[0080] Figure 11 This is a schematic diagram of an embodiment of the computed tomography (CT) scanning device provided by the present invention. Detailed Implementation

[0081] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0082] It should be understood that the illustrative drawings are not drawn to scale. The flowcharts used in this invention illustrate operations implemented according to some embodiments of the invention. It should be understood that the operations in the flowcharts may be implemented out of order, and steps without logical contextual relationships may be reversed or performed simultaneously. Furthermore, those skilled in the art, guided by the content of this invention, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor systems and / or microcontroller systems.

[0083] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0084] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0085] This invention provides a method, apparatus, and computed tomography (CT) scanning device for reconstructing localization images, which will be described below.

[0086] Figure 1 This is a schematic flowchart of an embodiment of the positioning patch image reconstruction method provided by the present invention, as shown below. Figure 1 As shown, the localization patch image reconstruction method includes:

[0087] S101. Obtain a first partial projection image and a second partial projection image of the positioning patch, wherein the first partial projection image includes an overlapping area that overlaps with the second partial projection image.

[0088] S102. Determine multiple first frequency band images of the first local projection image and multiple second frequency band images of the second local projection image, wherein the number of multiple first frequency band images is the same as the number of multiple second frequency band images;

[0089] S103. Determine multiple weighted mask frequency band images of the overlapping region;

[0090] S104. Based on multiple weighted mask frequency band images, fuse multiple first frequency band images and multiple second frequency band images to obtain multiple fused frequency band images;

[0091] S105. Determine the location patch image based on multiple fused frequency band images.

[0092] Compared with the prior art, the positioning patch image reconstruction method provided in this embodiment of the invention, after acquiring the first local projection image and the second local projection image of the positioning patch, determines multiple first frequency band images of the first local projection image and multiple second frequency band images of the second local projection image. Since different frequency band images represent features of different scales of the local projection image, and the image size of each layer is also different, the matching of image size and image scale features is achieved. That is, large-scale features are matched with larger image sizes to avoid highlighting detailed features in larger image sizes, thereby preventing ghosting. Small-scale features are matched with smaller image sizes to avoid loss of detailed features, thereby preventing truncation. This improves the image quality of the positioning patch image determined based on the fused Laplacian pyramid.

[0093] It should be understood that the method for obtaining the first and second local projection images of the positioning piece in step S101 can be either by obtaining the first and second local projection images through real-time scanning by the scanning system, or by retrieving them from a storage medium containing the first and second local projection images. The scanning system can be a Micro CT scanning device.

[0094] It should be noted that in practical applications, the positioning image is often formed by fusing multiple local projection images, and not just by fusing the first local projection image and the second local projection image. In order to maximize the image quality of the determined positioning image, when the positioning image is formed by fusing multiple local projection images, any two adjacent local projection images in the multiple local projection images are fused according to the method of steps S101-S105, which will not be elaborated here.

[0095] In some embodiments of the present invention, such as Figure 2 As shown, step S102 can be specifically described as follows:

[0096] Construct a first Laplacian pyramid of the first local projection image and a second Laplacian pyramid of the second local projection image to obtain multiple first frequency band images and multiple second frequency band images accordingly;

[0097] or,

[0098] The first local projection image and the second local projection image are filtered by a high-pass filter respectively to obtain multiple first frequency band images and multiple second frequency band images.

[0099] or,

[0100] Principal Component Analysis (PCA) is used to decompose the first local projection image and the second local projection image to obtain multiple first frequency band images and multiple second frequency band images.

[0101] or,

[0102] The first local projection image and the second local projection image are decomposed based on the wavelet transform method to obtain multiple first frequency band images and multiple second frequency band images.

[0103] The embodiments of the present invention provide the above four methods for frequency division of the first local projection image and the second local projection image, which can improve the diversity of frequency division methods.

[0104] It should be noted that the frequency division of the first local projection image and the second local projection image can also be based on contour wave transformation, curvature transformation, etc., which will not be elaborated here.

[0105] In a specific embodiment of the present invention, filtering the first local projection image based on a high-pass filter to obtain multiple first frequency band images specifically involves filtering the first local projection image based on a high-pass filter to obtain high-frequency band images and low-frequency band images.

[0106] In a specific embodiment of the present invention, the decomposition of the first local projection image based on principal component analysis to obtain multiple first frequency band images specifically involves: First, calculating the covariance matrix of the first local projection image and obtaining the eigenvectors and eigenvalues ​​of the covariance matrix. Then, sorting the eigenvalues ​​in descending order, the eigenvector corresponding to the largest eigenvalue is the first principal component, the eigenvector corresponding to the second largest eigenvalue is the second principal component, and so on. Finally, multiplying the first local projection image by each eigenvector yields the multiple first frequency band images.

[0107] In a specific embodiment of the present invention, the method of decomposing the first local projection image based on wavelet transform to obtain multiple first frequency band images is as follows: select a set of wavelets of different scales (such as Haar wavelets), first perform wavelet transform on the horizontal direction of the image respectively, and then perform wavelet transform on the transformed image in the vertical direction respectively, thereby obtaining different frequency components of the image.

[0108] Because obtaining multiple first-band images based on a high-pass filter only yields low-frequency and high-frequency images of the first local projection image, the number of frequency band images is limited, sometimes resulting in the loss of important details. Obtaining multiple first-band images based on principal component analysis concentrates the main information in the first-band images, with subsequent band images decreasing in number, leading to significant variations in the brightness of the projected images of different frequency bands. Obtaining multiple first-band images based on wavelet transform requires selecting wavelets of different scales, and the transformation process is complex. Therefore, in a preferred embodiment of the present invention, the method for obtaining multiple first-band images is to construct a first Laplacian pyramid of the first local projection image to obtain multiple first-band images. Each layer of the first Laplacian pyramid represents a different first-band image, characterizing features of the local projection image at different scales.

[0109] It should be understood that the method for obtaining the second frequency band image is the same as the method for obtaining the first frequency band image, and will not be elaborated here.

[0110] In some embodiments of the present invention, such as Figure 2 As shown, constructing the first Laplacian pyramid of the first local projection image includes:

[0111] S201. Perform Gaussian pyramid decomposition on the first local projection image to obtain the first Gaussian pyramid;

[0112] S202. Determine the first Laplace pyramid based on the first Gaussian pyramid.

[0113] In this embodiment of the invention, the first local projection image is first decomposed into a Gaussian pyramid to obtain images of different scales. Then, a first Laplacian pyramid is determined based on the first Gaussian pyramid. Each layer of the first Laplacian pyramid represents an image of a first frequency band, which is used to characterize the different scale features of images of different scales, thereby realizing the decomposition of different scale features of the first local projection image.

[0114] It should be understood that when obtaining the second frequency band image, it is necessary to construct a second Laplacian pyramid. The construction method of the second Laplacian pyramid is the same as that of the first Laplacian pyramid, and will not be elaborated here.

[0115] It should be noted that, in order to facilitate image fusion, the number of layers in the first Laplacian pyramid is the same as that in the second Laplacian pyramid, and the Laplacian images located in the same layer have the same resolution. The number of layers in the first and second Laplacian pyramids should be preset or adjusted according to the actual application scenario, such as image resolution requirements, and no specific limitation is made here.

[0116] In a specific embodiment of the present invention, taking an example where both the first and second Laplace pyramids have three layers, the first Gaussian pyramid also has three layers. That is, the first Gaussian pyramid includes a first layer Gaussian image, a second layer Gaussian image, and a third layer Gaussian image; therefore, as shown... Figure 3 As shown, step S201 includes:

[0117] S301. Use the first local projection image as the first layer Gaussian image;

[0118] S302. Perform Gaussian filtering on the first local projection image based on preset Gaussian filtering parameters to obtain the first preprocessed projection image.

[0119] S303. The first preprocessed projection image is downsampled based on preset downsampling parameters to obtain a second Gaussian image.

[0120] S304. Perform Gaussian filtering on the second layer Gaussian image based on preset Gaussian filtering parameters to obtain the second preprocessed projection image.

[0121] S305. The second preprocessed projection image is downsampled based on preset downsampling parameters to obtain the third layer Gaussian image.

[0122] This invention, by performing Gaussian filtering on the preprocessed projection image before downsampling, reduces image noise and further improves the image quality of the final positioning image.

[0123] The preset Gaussian filter parameters include the window radius of the Gaussian density function, and the preset downsampling parameter is the sampling interval.

[0124] In a specific embodiment of the present invention, the window radius of the Gaussian density function is 2, and the sampling interval is 1 pixel, that is, 1 pixel is sampled for each 1 pixel, which means that the number of pixels in the downsampled image is 1 / 2 of the original image.

[0125] The method for determining each Gaussian image layer can be expressed by the following formula:

[0126]

[0127] In the formula, G k (x,y) is the k-th Gaussian image, G k-1 (2x+m,2y+n) is the (k-1)th layer Gaussian image; h(m,n) is the Gaussian density function; w is the window radius of the Gaussian density function; (x,y) are the pixel coordinates; (m,n) are the indices of the surrounding neighboring pixels of pixel (x,y).

[0128] In some embodiments of the present invention, the first Laplacian pyramid includes a first layer of Laplacian images, a second layer of Laplacian images, and a third layer of Laplacian images, wherein each layer of Laplacian images is a first frequency band image; then as follows Figure 4 As shown, step S202 includes:

[0129] S401. Upsample the second-layer Gaussian image and the third-layer Gaussian image to obtain the second-layer upsampled image and the third-layer upsampled image respectively;

[0130] S402. Perform Gaussian convolution processing on the second-layer upsampled image and the third-layer upsampled image to obtain the second-layer Gaussian convolution image and the third-layer Gaussian convolution image respectively.

[0131] S403. The difference between the first Gaussian image and the second Gaussian convolution image is used as the first Laplacian image.

[0132] S404. The difference between the second Gaussian image and the third Gaussian convolution image is used as the second Laplacian image.

[0133] S405. Use the third-layer Gaussian image as the third-layer Laplacian image.

[0134] In a specific embodiment of the present invention, the method for determining each layer of the Laplacian image can be expressed by the following formula:

[0135]

[0136]

[0137] In the formula, L k The image is the Laplacian image of the k-th layer; Up() is the upsampling function; G k+1 This is the Laplacian image of the (k+1)th layer; Gaussian convolution; is the convolution operator; h(m,n) is the Gaussian density function; N is the total number of layers in the first Laplace pyramid.

[0138] In some embodiments of the present invention, such as Figure 5 As shown, step S103 includes:

[0139] S501. Determine the initial weighted mask image;

[0140] S502. Perform Gaussian pyramid decomposition on the initial weighted mask image to obtain multiple weighted mask frequency band images;

[0141] or,

[0142] S503. Filter the initial weighted mask based on a high-pass filter to obtain multiple weighted mask frequency band images;

[0143] or,

[0144] S504. Decompose the initial weighted mask based on principal component analysis to obtain multiple weighted mask frequency band images;

[0145] or,

[0146] S505. The initial weighted mask is decomposed based on the wavelet transform method to obtain multiple weighted mask frequency band images.

[0147] The size of the initial weighted mask can be set according to the fusion region of the image fusion. In a specific embodiment of the present invention, the image fusion is to fuse overlapping regions, so the size of the initial weighted mask should be consistent with the size of the overlapping regions.

[0148] In a specific embodiment of the present invention, the pixel value of the left half of the initial weighted mask image is 1, and the pixel value of the right half is 0.

[0149] It should be understood that the pixel values ​​in the initial weighted mask image can also be set to 0 for the left half and 1 for the right half, or 1 for the left 1 / 3 and 0 for the right 1 / 3, etc. These settings can be adjusted and set according to the actual application scenario and fusion requirements, and will not be elaborated here.

[0150] It should be noted that the method for obtaining multiple first-band images should be the same as the method for obtaining multiple weighted mask band images to ensure the reliability and accuracy of the obtained fused band images.

[0151] It should be noted that the specific methods of performing Gaussian pyramid decomposition, principal component analysis-based decomposition of the initial weighted mask image, and wavelet transform-based decomposition of the initial weighted mask image in steps S502-S505 are the same as the specific methods of processing the first local projection image described above, and will not be repeated here.

[0152] Since the embodiments of the present invention require the fusion of multiple first frequency band images and multiple second frequency band images based on multiple fusion frequency band images, the number of multiple fusion frequency band images should be the same as the number of multiple first frequency band images and multiple second frequency band images, and the resolution of the fusion frequency band images located in the same frequency band should also be the same as the resolution of the first frequency band images.

[0153] In some embodiments of the present invention, step S104 specifically involves obtaining a fused frequency band image based on each weighted mask frequency band image, a first frequency band image, a second frequency band image, and a preset fusion model.

[0154] In a specific embodiment of the present invention, the fusion model is as follows:

[0155]

[0156] In the formula, T k For fused frequency band images; L k This is the k-th image in the first frequency band; For the k-th second frequency band image; M k This is the k-th weighted mask frequency band image.

[0157] In a specific embodiment of the present invention, taking the constructed first Laplacian pyramid and second Laplacian pyramid as examples, the fusion process is described. Specifically, the multiple weighted mask frequency band images include a first weighted mask frequency band image, a second weighted mask frequency band image, and a third weighted mask frequency band image; the second Laplacian pyramid includes a first adjacent layer Laplacian image, a second adjacent layer Laplacian image, and a third adjacent layer Laplacian image, respectively corresponding to the first layer Laplacian image, the second layer Laplacian image, and the third layer Laplacian image. Each layer Laplacian image is a first frequency band image, and each adjacent layer Laplacian image is a second frequency band image; then, as... Figure 6 As shown, step S104 includes:

[0158] S601. Obtain the first layer fused Laplacian image based on the first weighted mask frequency band image, the first layer Laplacian image, the first adjacent layer Laplacian image, and the preset fusion model.

[0159] S602. Obtain the second-layer fused Laplacian image based on the second weighted mask frequency band image, the second-layer Laplacian image, the second adjacent-layer Laplacian image, and the preset fusion model;

[0160] S603. Obtain the third-layer fused Laplacian image based on the third weighted mask frequency band image, the third-layer Laplacian image, the third adjacent layer Laplacian image, and the preset fusion model.

[0161] S604. Obtain the fused Laplacian pyramid based on the first layer fused Laplacian image, the second layer fused Laplacian image, and the third layer fused Laplacian image.

[0162] In some embodiments of the present invention, when multiple first frequency band images and multiple second frequency band images are obtained by means of a high-pass filter, step S105 specifically involves superimposing the low-frequency fused frequency band image and the high-frequency fused frequency band image to obtain a positioning patch image.

[0163] In some embodiments of the present invention, when multiple first frequency band images and multiple second frequency band images are obtained by principal component analysis, step S105 specifically involves stitching together the images of different frequency bands and then performing inverse PCA transformation (i.e., multiplying the first local projection image, the second local projection image, and the transpose of the feature vector) to obtain the positioning patch image.

[0164] In some embodiments of the present invention, when multiple first frequency band images and multiple second frequency band images are obtained by wavelet transform, step S105 specifically involves stitching together the images of different frequency bands and then performing inverse wavelet transform to obtain a positioning patch image.

[0165] In some embodiments of the present invention, when multiple first frequency band images and multiple second frequency band images are obtained by constructing a first Laplacian pyramid and a second Laplacian pyramid, step S105 specifically involves: determining a positioning patch image based on the first-layer fused Laplacian image, the second-layer fused Laplacian image, and the third-layer fused Laplacian image.

[0166] like Figure 7 As shown, determining the localization patch image based on the first-layer fused Laplacian image, the second-layer fused Laplacian image, and the third-layer fused Laplacian image specifically includes:

[0167] S701. Upsample the third-layer fused Laplacian image to obtain a first upsampled image, and superimpose the first upsampled image and the second-layer fused Laplacian image to obtain a first superimposed image;

[0168] S702. Upsample the first overlay image to obtain a second upsampled image, and overlay the second upsampled image and the first layer fused Laplacian image to obtain a second overlay image;

[0169] S703. Upsample the second superimposed image to obtain the positioning patch image.

[0170] In some application scenarios, due to the influence of scanning parameters of the scanning equipment (e.g., low X-ray tube dose), the detailed features on the first and second local projection images are not obvious, resulting in low clarity of the determined positioning slice image. To solve this technical problem, in some embodiments of the present invention, such as... Figure 8 As shown, step S105 includes:

[0171] Determine the weights of each fused frequency band image, and determine the location patch image based on the weights and the fused frequency band image.

[0172] This invention, by assigning weights to images in different frequency bands, can enhance the detail features of high frequencies by adjusting the weights, thereby improving the clarity of the determined positioning image.

[0173] Taking the determination of a localization patch image based on the first-layer fused Laplacian image, the second-layer fused Laplacian image, and the third-layer fused Laplacian image as an example, in a specific embodiment of the present invention, such as... Figure 8 As shown, step S105 includes:

[0174] S801, Determine the first weight of the first layer fused Laplacian image, the second weight of the second layer fused Laplacian image, and the third weight of the third layer fused Laplacian image;

[0175] S802. Determine the localization patch image based on the first weight, the second weight, the third weight, the first-layer fused Laplacian image, the second-layer fused Laplacian image, and the third-layer fused Laplacian image.

[0176] It should be noted that: since the larger the weight corresponding to each layer of the fused Laplacian image, the more obvious the feature scale represented by the corresponding layer, and the high-frequency detail features are mainly determined by the high-level fused Laplacian image, in some embodiments of the present invention, the third weight is greater than the second weight, and the second weight is greater than the first weight, thereby enhancing the high-frequency detail features of the image and improving the clarity of the positioning patch image.

[0177] It should be understood that the first, second, and third weights can be set based on empirical values ​​or a preset weight model. To improve the determination efficiency of the positioning image, in some embodiments of the present invention, the first, second, and third weights are set according to a preset weight model. Specifically, the preset weight model is as follows:

[0178]

[0179] In the formula, α is the weight of the k-th layer of the fused Laplacian image; N is the total number of layers of the fused Laplacian pyramid.

[0180] In specific embodiments of the present invention, such as Figure 9 As shown, step S802 includes:

[0181] S901. Upsample the third-layer fused Laplacian image to obtain a first upsampled image, and weight the first upsampled image based on the third weight to obtain a first weighted image. Superimpose the first weighted image and the second-layer fused Laplacian image to obtain a first superimposed image.

[0182] S902. Upsample the first overlay image to obtain a second upsampled image, and weight the second upsampled image based on the second weight to obtain a second weighted image. Overlay the second weighted image with the first layer fused Laplacian image to obtain a second overlay image.

[0183] S903. Upsample the second overlay image to obtain a third upsampled image, and weight the third upsampled image based on the first weight to obtain a positioning patch image.

[0184] The embodiments of the present invention determine the positioning patch image by using a first weight, a second weight, and a third weight, which can enhance the high-frequency detail features in the determined positioning patch image, thereby improving the clarity of the positioning patch image.

[0185] To better implement the positioning patch image reconstruction method in this embodiment of the invention, based on the positioning patch image reconstruction method, correspondingly, as follows: Figure 10 As shown, this embodiment of the invention also provides a positioning patch image reconstruction device, the positioning patch image reconstruction device 1000 comprising:

[0186] The local projection image acquisition unit 1001 is used to acquire a first local projection image and a second local projection image of the positioning patch. The first local projection image includes an overlapping area that overlaps with the second local projection image.

[0187] The local projection image frequency division unit 1002 is used to determine multiple first frequency band images of the first local projection image and multiple second frequency band images of the second local projection image, wherein the number of multiple first frequency band images is the same as the number of multiple second frequency band images.

[0188] The overlapping region frequency division unit 1003 is used to determine multiple weighted mask frequency band images of the overlapping region;

[0189] The frequency band fusion unit 1004 is used to fuse multiple first frequency band images and multiple second frequency band images based on multiple weighted mask frequency band images to obtain multiple fused frequency band images;

[0190] The positioning image determination unit 1005 is used to determine the positioning image based on multiple fused frequency band images.

[0191] The positioning patch image reconstruction device 1000 provided in the above embodiments can realize the technical solutions described in the above positioning patch image reconstruction method embodiments. The specific implementation principles of each module or unit can be found in the corresponding content in the above positioning patch image reconstruction method embodiments, and will not be repeated here.

[0192] like Figure 11 As shown, the present invention also provides a computed tomography (CT) scanner 1100. The CT scanner 1100 includes a processor 1101, a memory 1102, and a display 1103. Figure 11 Only some components of the computed tomography (CT) scanner 1100 are shown; however, it should be understood that implementation of all shown components is not required, and more or fewer components may be implemented instead.

[0193] In some embodiments, memory 1102 may be an internal storage unit of the computed tomography (CT) scanner 1100, such as a hard disk or memory of the CT scanner 1100. In other embodiments, memory 1102 may also be an external storage device of the CT scanner 1100, such as a pluggable hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on the CT scanner 1100.

[0194] Furthermore, the memory 1102 may include both internal storage units of the computed tomography (CT) scanner 1100 and external storage devices. The memory 1102 is used to store application software and various types of data for which the CT scanner 1100 is installed.

[0195] In some embodiments, processor 1101 may be a central processing unit (CPU), microprocessor or other data processing chip, used to run program code stored in memory 1102 or process data, such as the positioning patch image reconstruction method of the present invention.

[0196] In some embodiments, display 1103 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 1103 is used to display information from the computed tomography (CT) scanner 1100 and to display a visual user interface. Components 1101-1103 of the CT scanner 1100 communicate with each other via a system bus.

[0197] In one embodiment of the present invention, when the processor 1101 executes the positioning patch image reconstruction program in the memory 1102, the following steps can be implemented:

[0198] Acquire a first local projection image and a second local projection image of the positioning patch, wherein the first local projection image includes an overlapping region that overlaps with the second local projection image;

[0199] A plurality of first frequency band images of a first local projection image and a plurality of second frequency band images of a second local projection image are determined, wherein the number of the plurality of first frequency band images is the same as the number of the plurality of second frequency band images;

[0200] Determine multiple weighted mask frequency band images of overlapping regions;

[0201] Multiple first-band images and multiple second-band images are fused based on multiple weighted mask frequency band images to obtain multiple fused frequency band images;

[0202] The location patch image is determined based on images from multiple fused frequency bands.

[0203] It should be understood that when the processor 1101 executes the positioning patch image reconstruction program in the memory 1102, in addition to the functions mentioned above, it can also perform other functions, as can be found in the description of the corresponding method embodiments above.

[0204] Furthermore, this embodiment of the invention does not specifically limit the type of the computed tomography (CT) scanner 1100 mentioned. The CT scanner 1100 can be a portable CT scanner such as a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, or laptop computer. Exemplary embodiments of portable CT scanners include, but are not limited to, portable CT scanners running iOS, Android, Microsoft, or other operating systems. The aforementioned portable CT scanner can also be other portable CT scanners, such as laptop computers with touch-sensitive surfaces (e.g., touch panels). It should also be understood that in some other embodiments of the invention, the CT scanner 1100 may not be a portable CT scanner, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0205] Accordingly, this application also provides a computer-readable storage medium for storing computer-readable programs or instructions. When the programs or instructions are executed by a processor, they can implement the positioning patch image reconstruction method steps or functions provided in the above-described method embodiments.

[0206] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0207] The above provides a detailed description of the positioning slice image reconstruction method, apparatus, and computed tomography (CT) scanning device provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for reconstructing a localization patch image, characterized in that, include: A first local projection image and a second local projection image of the positioning patch are obtained, wherein the first local projection image includes an overlapping region that overlaps with the second local projection image; A plurality of first frequency band images of the first local projection image and a plurality of second frequency band images of the second local projection image are determined, wherein the number of the plurality of first frequency band images is the same as the number of the plurality of second frequency band images; Multiple weighted mask frequency band images are determined for the overlapping region; the number of the multiple weighted mask frequency band images is the same as the number of the multiple first frequency band images; Based on the multiple weighted mask frequency band images, the multiple first frequency band images and the multiple second frequency band images are fused to obtain multiple fused frequency band images; The location patch image is determined based on the multiple fused frequency band images; the fused frequency band images include a first-layer fused Laplacian image, a second-layer fused Laplacian image, and a third-layer fused Laplacian image; The step of determining the location patch image based on the multiple fused frequency band images includes: Determine the first weight of the first layer fused Laplacian image, the second weight of the second layer fused Laplacian image, and the third weight of the third layer fused Laplacian image; The localization patch image is determined based on the first weight, the second weight, the third weight, the first layer fused Laplacian image, the second layer fused Laplacian image, and the third layer fused Laplacian image; The first, second, and third weights are set according to a preset weight model, which is as follows: In the formula, represents the weights of the k-th layer of the merged Laplacian image; N is the total number of layers in the merged Laplacian pyramid. Determining multiple weighted mask frequency band images of the overlapping region includes: determining an initial weighted mask image; performing Gaussian pyramid decomposition on the initial weighted mask image to obtain the multiple weighted mask frequency band images; the pixel values ​​in the initial weighted mask image are 0 for the left half and 1 for the right half, or 1 for the left 1 / 3 and 0 for the right 1 / 3.

2. The method for reconstructing a localization patch image according to claim 1, characterized in that, Determining the plurality of first frequency band images of the first local projection image and the plurality of second frequency band images of the second local projection image includes: Construct a first Laplacian pyramid of the first local projection image and a second Laplacian pyramid of the second local projection image to obtain the plurality of first frequency band images and the plurality of second frequency band images accordingly; or, The first local projection image and the second local projection image are filtered by a high-pass filter respectively to obtain the plurality of first frequency band images and the plurality of second frequency band images. or, Principal component analysis is used to decompose the first local projection image and the second local projection image to obtain the plurality of first frequency band images and the plurality of second frequency band images. or, The first local projection image and the second local projection image are decomposed based on the wavelet transform method to obtain the plurality of first frequency band images and the plurality of second frequency band images.

3. The method for reconstructing a positioning patch image according to claim 2, characterized in that, The construction of the first Laplacian pyramid of the first local projection image includes: The first local projection image is decomposed into a Gaussian pyramid to obtain the first Gaussian pyramid. The first Laplace pyramid is determined based on the first Gaussian pyramid.

4. The positioning patch image reconstruction method according to claim 3, characterized in that, The first Gaussian pyramid includes a first layer of Gaussian images, a second layer of Gaussian images, and a third layer of Gaussian images; The step of performing Gaussian pyramid decomposition on the first local projection image to obtain the first Gaussian pyramid includes: The first local projection image is used as the first layer Gaussian image; The first local projection image is subjected to Gaussian filtering based on preset Gaussian filtering parameters to obtain the first preprocessed projection image. The first preprocessed projection image is downsampled based on preset downsampling parameters to obtain the second layer Gaussian image; The second layer Gaussian image is subjected to Gaussian filtering based on preset Gaussian filtering parameters to obtain the second preprocessed projection image. The second preprocessed projection image is downsampled based on preset downsampling parameters to obtain the third layer Gaussian image.

5. The method for reconstructing a positioning patch image according to claim 4, characterized in that, The first Laplacian pyramid includes a first layer of Laplacian images, a second layer of Laplacian images, and a third layer of Laplacian images; Determining the first Laplace pyramid based on the first Gaussian pyramid includes: Upsample the second-layer Gaussian image and the third-layer Gaussian image to obtain the second-layer upsampled image and the third-layer upsampled image respectively; Gaussian convolution is performed on the second-layer upsampled image and the third-layer upsampled image to obtain the second-layer Gaussian convolution image and the third-layer Gaussian convolution image respectively; The difference between the first Gaussian image and the second Gaussian convolution image is used as the first Laplacian image. The difference between the second Gaussian image and the third Gaussian convolution image is used as the second Laplacian image. The third Gaussian image is used as the third Laplacian image.

6. The method for reconstructing a positioning patch image according to claim 5, characterized in that, The determination of the multiple weighted mask frequency band images of the overlapping region includes: Determine the initial weighted mask image; Gaussian pyramid decomposition is performed on the initial weighted mask image to obtain the multiple weighted mask frequency band images; or, The initial weighted mask is filtered using a high-pass filter to obtain the frequency band images of the multiple weighted masks; or, The initial weighted mask is decomposed based on principal component analysis to obtain the frequency band images of the multiple weighted masks; or, The initial weighted mask is decomposed using wavelet transform to obtain the multiple weighted mask frequency band images.

7. The method for reconstructing a positioning patch image according to claim 6, characterized in that, The plurality of weighted mask frequency band images include a first weighted mask frequency band image, a second weighted mask frequency band image, and a third weighted mask frequency band image; the second Laplacian pyramid includes a first adjacent layer Laplacian image, a second adjacent layer Laplacian image, and a third adjacent layer Laplacian image, which are respectively corresponding to the first layer Laplacian image, the second layer Laplacian image, and the third layer Laplacian image; The process of fusing the multiple first frequency band images and the multiple second frequency band images based on the multiple weighted mask frequency band images to obtain multiple fused frequency band images includes: The first layer fused Laplacian image is obtained based on the first weighted mask frequency band image, the first layer Laplacian image, the first adjacent layer Laplacian image, and the preset fusion model. The second-layer fused Laplacian image is obtained based on the second weighted mask frequency band image, the second-layer Laplacian image, the second adjacent-layer Laplacian image, and the preset fusion model; The third-layer fused Laplacian image is obtained based on the third weighted mask frequency band image, the third-layer Laplacian image, the third adjacent layer Laplacian image, and the preset fusion model. The fused Laplacian pyramid is obtained based on the first layer fused Laplacian image, the second layer fused Laplacian image, and the third layer fused Laplacian image.

8. A positioning patch image reconstruction device, characterized in that, The apparatus applicable to the localization patch image reconstruction method according to any one of claims 1-7 comprises: A local projection image acquisition unit is used to acquire a first local projection image and a second local projection image of the positioning patch, wherein the first local projection image includes an overlapping region that overlaps with the second local projection image; A local projection image frequency division unit is used to determine a plurality of first frequency band images of the first local projection image and a plurality of second frequency band images of the second local projection image, wherein the number of the plurality of first frequency band images is the same as the number of the plurality of second frequency band images; An overlapping region frequency division unit is used to determine multiple weighted mask frequency band images of the overlapping region; the number of the multiple weighted mask frequency band images is the same as the number of the multiple first frequency band images; determining the multiple weighted mask frequency band images of the overlapping region includes: determining an initial weighted mask image; performing Gaussian pyramid decomposition on the initial weighted mask image to obtain the multiple weighted mask frequency band images; the pixel values ​​in the initial weighted mask image are either 0 for the left half and 1 for the right half, or 1 for the left 1 / 3 and 0 for the right 1 / 3. A frequency band fusion unit is used to fuse the multiple first frequency band images and the multiple second frequency band images based on the multiple weighted mask frequency band images to obtain multiple fused frequency band images; The positioning image determination unit is used to determine the positioning image based on the multiple fused frequency band images.

9. A computed tomography (CT) scanner, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the localization patch image reconstruction method according to any one of claims 1 to 7.

Images