CT image reconstruction method and system

Abstract

The invention relates to the technical field of medical image processing, in particular to a CT image reconstruction method and system. The method comprises the following steps: acquiring a tomography image through medical equipment, and carrying out image segmentation on the tomography image to obtain a target area; extracting feature points in the target area, and identifying a focus part according to the feature points; a CT image interlayer registration model is constructed, and interlayer registration is carried out on the tomography image through the CT image interlayer registration model based on the focus part; the registration results are stacked into three-dimensional data, the three-dimensional data are resampled, and then a three-dimensional image of the focus part is reconstructed. According to the method, the target area is extracted after the tomography image is segmented, and then the focus part is recognized to reconstruct the three-dimensional image, so that the problem of clearly displaying the three-dimensional form of human tissues or organs is solved, and doctors can analyze and research focus information from different angles so as to improve the operation success rate.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to a CT image reconstruction method and system. Background Art

[0002] Computed tomography technology provides doctors with rich information in clinical diagnosis, but this requires experienced doctors to reconstruct the corresponding three-dimensional structure of these sequence images in the brain, which makes the diagnosis results very subjective and may lead to misjudgment or failure to determine the lesion information. In order to display the lesion information more objectively and accurately, reduce the workload of doctors and thus improve the treatment effect, it is necessary to clearly display the three-dimensional morphology of human tissues or organs. Summary of the invention

[0003] In view of the shortcomings of existing methods and the needs of practical applications, in order to assist doctors in formulating personalized treatment plans, help doctors accurately locate important functional areas during surgery, and solve the problem of clearly displaying the three-dimensional morphology of human tissues or organs. On the one hand, the present invention provides a CT image reconstruction method, comprising the following steps: acquiring a tomographic image through a medical device, performing image segmentation on the tomographic image to obtain a target area; extracting feature points in the target area, and identifying the lesion site according to the feature points; constructing a CT image inter-layer registration model, and based on the lesion site, performing inter-layer registration on the tomographic image through the CT image inter-layer registration model; stacking the registration results into three-dimensional data, resampling the three-dimensional data, and then reconstructing a three-dimensional image of the lesion site. The present invention extracts the target area after segmenting the tomographic image and then identifies the lesion site, and then registers the tomographic images of different angles through the CT image inter-layer registration model and then reconstructs the three-dimensional image, which solves the problem of clearly displaying the three-dimensional morphology of human tissues or organs, and is conducive to doctors to analyze and study lesion information from different angles to improve the success rate of surgery.

[0004] Optionally, performing image segmentation on the tomographic image to obtain a target area comprises the following steps: Decompose the tomographic image to obtain a texture layer image; construct a first metric model, and obtain a first metric value based on the texture layer image and the first metric model; obtain the number of cluster centers based on the tomographic image; perform cluster segmentation on the tomographic image based on the first metric value and the number of cluster centers to obtain the target area. The present invention performs clustering iteration based on the first metric model in the texture layer image, which is conducive to accurately segmenting the target area, and further facilitates accurate identification of the lesion site.

[0005] Optionally, the first measurement model satisfies the following formula: in, represents the first metric value, Indicates the number of pixels in the texture layer image. represents the number of cluster centers, Indicates The pixel belongs to The membership degree of a cluster, represents the fuzzy factor, Indicates Pixels to The measure of the center of a cluster.

[0006] Optionally, the Pixels to The measure of the cluster center satisfies the following formula: in, Indicates Pixels to The measure of the cluster centers, Indicates The spatial frequency of pixels, Indicates Pixels to The horizontal spacing between cluster centers, Indicates Pixels to The vertical spacing of cluster centers, Indicates Pixel to The Euclidean distance between cluster centers, Indicates The horizontal position of the pixel, Indicates The vertical position of the pixel, Indicates The gray value of the pixel, Indicates The present invention uses the model formula to accurately obtain the first metric value, which is conducive to the objective and accurate construction of the three-dimensional image of the lesion site.

[0007] Optionally, obtaining the number of cluster centers according to the tomographic image comprises the following steps: According to the grayscale value of the tomographic image, a grayscale value histogram is constructed; data processing is performed according to the grayscale value histogram, and then the number of cluster centers is obtained according to the data processing result. The present invention uses the grayscale value of the tomographic image to determine the number of cluster centers, which is further conducive to the objective and accurate construction of a three-dimensional image of the lesion site. Optionally, extracting feature points in the target area and identifying the lesion site according to the feature points comprises the following steps: Kernels of different sizes are set, and black hat operations are performed on the target area based on the kernels; feature points are extracted according to the black hat operation results corresponding to the kernels of different sizes; edge detection is performed with the feature points as the center to identify the lesion site. The present invention uses black hat operations to extract feature points, and then identifies the lesion site based on the feature points. The calculation is simple, which is conducive to improving the reconstruction efficiency of the present invention.

[0008] Optionally, performing edge detection with the feature point as the center to identify the lesion site includes the following steps: Construct a filtering window, and filter the target area according to the filtering window; calculate the pixel gradient, and filter the candidate edge points according to the pixel gradient with the feature point as the center; set a screening threshold, and obtain the contour of the lesion site by combining the screening threshold and the candidate edge points. The present invention filters edge points twice by using pixel gradient and screening threshold, which is conducive to accurately identifying the lesion site and improving the accuracy of the present invention.

[0009] Optionally, the CT image inter-layer registration model satisfies the following formula: in, Representing images With image The registration cost function is Representing images With image The amount of interactive information, Representing images With image The normalized mutual information of Representing images With image The interaction variance of Representing images With image The Euclidean closeness of Representing images With image The maximum and minimum closeness of Representing images With image The registration cost function constructed by the present invention can accurately evaluate the image registration result, which is further conducive to fast and accurate image reconstruction.

[0010] Optionally, stacking the registration results into three-dimensional data comprises the following steps: According to the gray value difference of the registered image, an interpolation method is selected to interpolate the registered image; and the three-dimensional data is obtained by stacking the interpolation results. The present invention interpolates the registration results in different ways according to the interpolation threshold, which is conducive to improving the image interpolation effect and further improving the image reconstruction accuracy of the present invention.

[0011] In the second aspect, in order to efficiently execute the CT image reconstruction method provided by the present invention, the present invention also provides a CT image reconstruction system, including a processor, an input device, an output device and a memory, wherein the processor, the input device, the output device and the memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the CT image reconstruction method as described in the first aspect of the present invention. The CT image reconstruction system of the present invention has a compact structure and stable performance, and can stably execute the CT image reconstruction method provided by the present invention, further improving the overall applicability and practical application ability of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A flow chart of a CT image reconstruction method provided by an embodiment of the present invention; Figure 2 Comparison diagram of CT images before and after reconstruction provided by an embodiment of the present invention; Figure 3 A framework diagram of a CT image reconstruction system provided by an embodiment of the present invention; Figure 4 A schematic diagram of the structure of a CT image reconstruction device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0013] The specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described herein are only for illustration and are not intended to limit the present invention. In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention. However, it is obvious to those of ordinary skill in the art that these specific details do not need to be adopted to implement the present invention. In other examples, in order to avoid confusing the present invention, known circuits, software or methods are not specifically described.

[0014] Throughout the specification, references to "one embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in conjunction with the embodiment or example is included in at least one embodiment of the present invention. Therefore, the phrases "in one embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily all refer to the same embodiment or example. In addition, particular features, structures, or characteristics may be combined in one or more embodiments or examples in any suitable combination and / or subcombination. In addition, it should be understood by those of ordinary skill in the art that the figures provided herein are for illustrative purposes and that the figures are not necessarily drawn to scale.

[0015] See also Figure 1 , in order to assist doctors in developing personalized treatment plans, help doctors accurately locate important functional areas during surgery, and solve the problem of clearly displaying the three-dimensional morphology of human tissues or organs. Figure 1 As shown, in one embodiment, the method includes the following steps: S1. Acquire a tomographic image through a medical device, and perform image segmentation on the tomographic image to obtain a target area.

[0016] CT images, or tomographic images, refer to images obtained through computed tomography (CT) technology. CT scans generate cross-sectional images of the body through X-ray beams and computers. When X-rays penetrate the human body, they collide with atoms in the human body and are attenuated. Tissues of different densities and thicknesses attenuate X-rays to different degrees. These attenuation information is recorded by the detector and processed by a computer to generate tomographic images. CT tomographic images play an important role in medical diagnosis, especially in the diagnosis of tumors, vascular lesions, trauma and other diseases. They have an irreplaceable position. Before surgery, doctors can use CT tomographic images to understand the condition of the lesion and develop a more accurate surgical plan. During treatment, CT tomographic images can be used to evaluate the treatment effect and observe changes in lesions.

[0017] CT scan images have high resolution and can clearly display the fine structures inside the human body. In addition, CT scans provide cross-sectional images of the body, that is, cross-sectional images perpendicular to the long axis of the body, which helps doctors to more accurately determine the location and range of lesions. Furthermore, the image data obtained through CT scans can be reconstructed in three dimensions to generate realistic three-dimensional models, which helps doctors to understand the lesions more intuitively.

[0018] Specifically, obtaining tomographic images through medical equipment refers to the use of CT scanning equipment and other key equipment for obtaining tomographic images, including X-ray sources, detectors, computer processing systems and other parts. The X-ray source emits an X-ray beam, which penetrates the human body and is received by the detector. The detector converts the received X-ray signal into an electrical signal and transmits it to the computer processing system. The computer processing system amplifies, converts analog-to-digital signals, reconstructs the image and other processes on the received electrical signal to ultimately generate a tomographic image.

[0019] Furthermore, the step S1 of segmenting the tomographic image to obtain the target area includes the following steps: S11, decomposing the tomographic image to obtain a texture layer image.

[0020] In the embodiment, the texture layer image is obtained by performing Fourier transform on the tomographic image.

[0021] It can be understood that the texture layer refers to the attribute of the image that reflects the arrangement of the surface structure of the object, which is usually manifested as recurring local patterns in the image and their arrangement rules. Therefore, in some other embodiments, it can also be obtained based on other statistical methods, structural methods, transformation-based methods, model-based methods or learning-based methods.

[0022] Statistical methods are based on the grayscale properties of pixels and their neighborhoods to study the statistical characteristics of texture areas. This type of method extracts texture features by analyzing the distribution and relationship of grayscale values ​​in an image, including grayscale co-occurrence matrix statistical methods, local binary pattern methods, and grayscale run length matrix methods.

[0023] The structural method is based on the theory of texture primitives (basic texture elements), which believes that texture is composed of a number of simple texture primitives arranged repeatedly in a certain regular form, including syntactic texture description algorithms and mathematical morphology methods.

[0024] Transformation-based methods describe texture by converting the image into different spaces or domains and then extracting relatively stable feature values, including Fourier transform, Gabor transform and wavelet transform.

[0025] The model-based method constructs an image model and uses the model parameters as its texture features, including Markov random field (MRF) model, MRF model and fractal model.

[0026] Learning-based methods automatically extract texture features through machine learning algorithms, including deep learning and visual dictionary methods.

[0027] S12: construct a first measurement model, and obtain a first measurement value according to the texture layer image and the first measurement model.

[0028] In an embodiment, the first measurement model satisfies the following formula: in, represents the first metric value, Indicates the number of pixels in the texture layer image. represents the number of cluster centers, Indicates The pixel belongs to The membership degree of a cluster, represents the fuzzy factor, Indicates Pixels to The measure of the center of a cluster.

[0029] Furthermore, the Pixels to The measure of the cluster center satisfies the following formula: in, Indicates Pixels to The measure of the cluster centers, Indicates The spatial frequency of pixels, Indicates Pixels to The horizontal spacing between cluster centers, Indicates Pixels to The vertical spacing of cluster centers, Indicates Pixel to The Euclidean distance between cluster centers, Indicates The horizontal position of the pixel, Indicates The vertical position of the pixel, Indicates The gray value of the pixel, Indicates The gray value of the cluster center point.

[0030] Specifically, taking the texture layer image as an object, the first measurement value of all pixel points of the texture layer image and the cluster center point is obtained through the first measurement model.

[0031] S13. Obtain the number of cluster centers according to the tomographic image.

[0032] In an embodiment, obtaining the number of cluster centers according to the tomographic image comprises the following steps: S131. Construct a grayscale value histogram according to the grayscale value of the tomographic image.

[0033] First, you need to read the image to be processed, which usually involves using an image processing library (such as OpenCV, PIL / Pillow, scikit-image, etc.) to load the image file; second, convert the grayscale image. If the image is in color (such as an RGB image), you need to convert it to a grayscale image. A grayscale image only contains brightness information, not color information, which makes processing simpler; then, initialize the histogram: create an array or list to store the number of pixels at each grayscale level. For an 8-bit grayscale image, the size of the array should be 256 (because the grayscale levels range from 0 to 255). Next, iterate over each pixel of the image and count the number of pixels at each grayscale level. This usually involves nested loops to iterate over the width and height of the image, and then update the histogram. For each pixel, update the count at the corresponding position in the histogram array according to its grayscale value. Finally, use a drawing library (such as matplotlib) to draw a grayscale value histogram, which usually involves setting the x-axis to grayscale level and the y-axis to the number of pixels, and drawing a bar chart or line chart.

[0034] In order to better compare the grayscale distribution of different images, the histogram can be normalized so that the number of pixels at each grayscale level is divided by the total number of pixels in the image, thus obtaining the probability distribution of the grayscale levels.

[0035] S132, performing data processing according to the gray value histogram, and then obtaining the number of cluster centers according to the data processing result.

[0036] Specifically, the gray value histogram is smoothed (such as by using kernel density estimation or other methods) to obtain a continuous probability density function, and then the function is differentiated to obtain the peak points in the histogram according to the derivative result, and the number of these peak points is used as the number of cluster centers.

[0037] S14. Based on the first metric value and the number of cluster centers, cluster segment the tomographic image to obtain the target area.

[0038] Specifically, first set the fuzzy factor, which controls the fuzziness of clustering. It is generally a real number greater than 1. In most applications, there is no strict limit on the value range of the fuzzy factor, but a value that is too large may increase the computational complexity of the algorithm and over-fuzzify the results, while a value that is too small may not fully reflect the fuzziness of the data. In this embodiment, the fuzzy factor is set to 2, which can provide a reasonable clustering result in most cases, but it can also be adjusted according to the specific data set and requirements.

[0039] Secondly, the cluster centers are randomly initialized according to the number of cluster centers, and the Lagrange multiplier method is used to solve the first The pixels belong to The membership of the class is then updated, and the membership and cluster center are updated. The first metric values ​​of the cluster center before and after the update are compared and iterated until convergence. According to the final clustering result, a cluster segmentation result is generated to obtain the target area.

[0040] S2. Extracting characteristic points in the target area, and identifying the lesion site according to the characteristic points.

[0041] In the embodiment, step S2, extracting feature points in the target area and identifying the lesion site according to the feature points, comprises the following steps: S21. Setting kernels of different sizes, and performing black hat operations on the target area based on the kernels.

[0042] Black hat operation is the operation of subtracting the original image from the closed image. The closing operation is the process of dilating the image and then corroding it. It can fill the small holes or dark areas in the image. The black hat operation obtains the details in the image that are darker than the surrounding area by calculating the difference between the closing operation result and the original image.

[0043] Furthermore, if the grayscale image is corroded by the structural element and then processed by the dilation operation, this process is called an opening operation. The output result of the opening operation is that it can retain bright spots larger than the structural unit, and can also keep all grayscale levels unchanged and the properties of the bright area relatively stable. Therefore, in some other embodiments, the target area can also be opened before the black hat operation to increase the sensitivity of the lesion area to the black hat operation kernel size selection. In the embodiment, the kernel is set to 3×3 and 7×7.

[0044] S22. Extract feature points according to the black hat operation results corresponding to kernels of different sizes.

[0045] The lesion area is sensitive to the selection of the black hat operation kernel size. The pixels at the lesion show obvious aggregation as the black hat operation kernel size increases. Therefore, we can compare the black hat operation results of two different kernel sizes and set a suitable threshold to extract a feature point at the lesion location.

[0046] S23, performing edge detection with the feature point as the center to identify the lesion site.

[0047] Specifically, the edge detection is performed with the feature point as the center to identify the lesion site, including the following steps: S231, constructing a filtering window, and performing filtering processing on the target area according to the filtering window.

[0048] The filter window is a matrix or frame used to perform filtering operations on an image. The size and shape of the window depend on the type and target of the filter. The window size determines the image area that the filtering operation will cover. A larger window can smooth more image details, but may also cause image blur. A smaller window can retain more image details, but may not completely remove noise. The shape of the window can be rectangular, circular, elliptical, etc. The choice of shape depends on the image features to be processed and the filtering target.

[0049] For some types of filters (such as weighted average filters), weight coefficients need to be set for each pixel position in the window. These coefficients determine the degree of influence of each pixel on the filtering result.

[0050] Select the area to be filtered on the target area image, which can be the entire image or a subset of the image. Move the filter window to each pixel position in the target area. At each position, use the pixel value in the window and the corresponding filter coefficient (if applicable) to calculate the filtering result; calculate the filtering result of the current pixel position based on the pixel value in the filtering window and the filter coefficient (if set). For a weighted average filter, this usually involves dividing the weighted sum of the pixel values ​​in the window by the total number of pixels in the window (or the sum of the weighted sums). Assign the calculated filtering result to the current pixel position, thereby updating the image.

[0051] S232, setting a screening threshold and calculating pixel gradients, screening candidate edge points based on the pixel gradients and prime screening thresholds with the feature point as the center, and obtaining the lesion site contour.

[0052] In the embodiment, the gradient amplitude is calculated in a 3×3 neighborhood, and the partial derivatives in the horizontal, vertical, 45° and 135° directions as well as the horizontal and vertical differences are calculated for the pixel point to obtain the amplitude and direction angle at the point.

[0053] Furthermore, according to the criterion of non-maximum suppression, possible edge points are screened out based on the amplitude and direction angle at the point, that is, by comparing the gradient amplitude of a point with the two adjacent points in the gradient direction, if the point is not greater than the amplitude of the adjacent pixel point, the point is excluded, otherwise it is retained in the set of candidate edge points.

[0054] S233, setting a screening threshold, and combining the screening threshold and the candidate edge points to obtain the lesion site contour.

[0055] Specifically, according to the grayscale histogram distribution, the average of the maximum grayscale value and the minimum grayscale value of the target area is used as the initial threshold, and the target area image is divided into a part greater than the initial threshold and a part not greater than the initial threshold, and then the average grayscale values ​​of the two parts are calculated respectively as the maximum screening threshold and the minimum screening threshold of the two parts.

[0056] Furthermore, the candidate edge points are screened according to the maximum screening threshold and the minimum screening threshold, and a graph composed of the screened edge points is obtained, that is, the contour of the lesion site is obtained.

[0057] Specifically, if the grayscale value of a candidate edge point is greater than the maximum screening threshold, it is an edge point. If it is between the maximum and minimum screening thresholds, it is necessary to determine whether there are points greater than the maximum screening threshold in the surrounding eight neighborhoods. If there are points that meet this condition, it is an edge point, otherwise it is not. If it is less than the minimum screening threshold, it is not an edge point.

[0058] S3. Construct a CT image inter-layer registration model, and based on the lesion location, perform inter-layer registration on the tomographic image using the CT image inter-layer registration model.

[0059] In an embodiment, the CT image inter-layer registration model constructed in step S3 satisfies the following formula: in, Representing images With image The registration cost function is Representing images With image The amount of interactive information, Representing images With image The normalized mutual information of Representing images With image The interaction variance of Representing images With image The Euclidean closeness of Representing images With image The maximum and minimum closeness of Representing images With image Hamming closeness.

[0060] Specifically, based on the lesion location, the tomographic image is inter-layeredly registered by the CT image inter-layer registration model. In CT image registration, the cost function is used to guide the optimization process of image transformation parameters. The essence of the registration problem is to apply transformation to move the coordinate points of the floating image to an optimal position so that the pixel intensity value at the new position has the greatest approximation to the pixel intensity value at the corresponding position of the reference image, that is, the cost function value is minimized.

[0061] S4. Stacking the registration results into three-dimensional data, resampling the three-dimensional data, and reconstructing a three-dimensional image of the lesion site.

[0062] In an embodiment, stacking the registration results into three-dimensional data includes the following steps: S41 . According to the gray value difference of the registered image, an interpolation method is selected to interpolate the registered image.

[0063] Specifically, the difference in pixel grayscale values ​​of corresponding points of the upper and lower images of the registered image is calculated. If it is less than the preset interpolation threshold, the grayscale value of the intermediate slice is directly calculated by linear interpolation. Otherwise, the pixel points of the upper and lower images are selected as the center, and the ratio of the layer spacing to the image pixel spacing is used as the radius to construct two corresponding regions of the same size. Then the normalized correlation of the two regions is calculated. If it is greater than the correlation threshold, the Lagrange interpolation method is selected to find the grayscale value of the corresponding point of the interpolation layer; if it is not greater than the correlation threshold, all corresponding pixels in the two regions are compared in terms of grayscale value, gradient size and direction, distance between two points in horizontal plane projection, and mean and variance in the neighborhood. The maximum similarity function value is selected as the corresponding pixel pair, and the grayscale value of the pixel point of the intermediate slice is calculated by linear interpolation.

[0064] S42, stacking according to the interpolation results to obtain the three-dimensional data.

[0065] Specifically, stacking is performed according to the interpolation results based on a volume rendering algorithm to obtain the three-dimensional data.

[0066] The volume rendering algorithm constructs three-dimensional space volume data through a sequence of two-dimensional images, directly generates a three-dimensional structure model and projects it onto the screen. It can be divided into image space method and object space method, such as ray casting method, snowball method and shear-curve method.

[0067] See also Figure 2, the left side is a three-dimensional CT image obtained by a conventional method, and the right side is a three-dimensional CT image obtained by the method provided by the present invention. It can be seen from the effect in the figure that the three-dimensional model of the medical three-dimensional image reconstructed by the method of the present invention has a clearer contour, a better three-dimensional visual effect, and eliminates the step-like phenomenon on the surface.

[0068] See also Figure 3 In an embodiment, in order to efficiently execute a CT image reconstruction method provided by the present invention, the present invention further provides a CT image reconstruction system, including: an input device, an output device, a processor, and a memory, wherein the input device, the output device, the processor, and the memory are interconnected, and the memory contains program instructions, and the program instructions are used for the steps of the CT image reconstruction method. The CT image reconstruction system of the present invention has a compact structure and stable performance, and can stably execute a CT image reconstruction method of the present invention, further improving the overall applicability and practical application ability of the present invention.

[0069] In an embodiment, the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. An input device may be used to obtain data information. An output device may be used to output the results obtained by storing program instructions contained in a computer program in a memory provided by the present invention. The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory.

[0070] In yet another alternative embodiment, see Figure 4 In order to efficiently execute the CT image reconstruction method provided by the present invention, this embodiment also provides a CT image reconstruction device, such as Figure 4 As shown, including: The memory 10 is used to store computer programs; the processor 20 is used to execute the computer programs to implement the above CT image reconstruction method. The memory 10, the processor 20, the communication interface 31 and the communication bus 32. The memory 10, the processor 20 and the communication interface 31 all communicate with each other through the communication bus 32.

[0071] In an embodiment, the memory 10 is used to store one or more program instructions. The memory 10 may store program instructions for implementing the following functions: A tomographic image is acquired through medical equipment, and the tomographic image is segmented to obtain a target area; feature points are extracted from the target area, and the lesion site is identified based on the feature points; a CT image inter-layer registration model is constructed, and based on the lesion site, the tomographic image is inter-layer registered through the CT image inter-layer registration model; the registration results are stacked into three-dimensional data, and the three-dimensional data is resampled to reconstruct a three-dimensional image of the lesion site.

[0072] In one possible implementation, the memory 10 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, and at least one application required for a function, etc.; the data storage area may store data created during use. In addition, the memory 10 may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include NVRAM. The memory stores an operating system and operating instructions, executable modules or data structures, or a subset thereof, or an extended set thereof, wherein the operating instructions may include various operating instructions for implementing various operations. The operating system may include various system programs for implementing various basic tasks and processing hardware-based tasks.

[0073] The processor 20 may be a central processing unit (CPU), an application specific integrated circuit, a digital signal processor, a field programmable gate array or other programmable logic device, and the processor 20 may be a microprocessor or any conventional processor, etc. The processor 20 may call a program stored in the memory 10. The communication interface 31 may be an interface of a communication module, and is used to connect with other devices or systems.

[0074] Of course, it should be noted that Figure 4 The structure shown does not constitute a limitation on the CT image reconstruction device in this embodiment. In actual applications, the CT image reconstruction device may include Figure 4 More or fewer components than shown, or combinations of certain components.

[0075] The embodiment further provides a storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above-mentioned CT image reconstruction method are implemented.

[0076] The storage medium may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0077] To summarize, the present invention extracts the target area after segmenting the tomographic image and then identifies the lesion site, and then reconstructs the three-dimensional image after aligning the tomographic images at different angles through the CT image inter-layer alignment model, thereby solving the problem of clearly displaying the three-dimensional morphology of human tissues or organs, which is beneficial for doctors to analyze and study lesion information from different angles to improve the success rate of surgery.

[0078] Therefore, the present invention effectively overcomes various shortcomings in the prior art and has high industrial utilization value.

[0079] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope recorded in the present invention.

Claims

1. A CT image reconstruction method, characterized in that: The CT image reconstruction method comprises the following steps: Acquiring a tomographic image through a medical device, and performing image segmentation on the tomographic image to obtain a target area; Extracting characteristic points in the target area, and identifying the lesion site according to the characteristic points; Constructing a CT image inter-layer registration model, and based on the lesion location, performing inter-layer registration on the tomographic image through the CT image inter-layer registration model; The registration results are stacked into three-dimensional data, and the three-dimensional data are resampled to reconstruct a three-dimensional image of the lesion site.

2. The CT image reconstruction method according to claim 1, characterized in that: The step of segmenting the tomographic image to obtain the target area comprises the following steps: Decomposing the tomographic image to obtain a texture layer image; Constructing a first measurement model, and obtaining a first measurement value according to the texture layer image and the first measurement model; According to the tomographic image, obtaining the number of cluster centers; Based on the first metric value and the number of cluster centers, cluster segmentation is performed on the tomographic image to obtain the target area.

3. The CT image reconstruction method according to claim 2, characterized in that: The first measurement model satisfies the following formula: in, represents the first metric value, Indicates the number of pixels in the texture layer image. represents the number of cluster centers, Indicates The pixel belongs to The membership degree of a cluster, represents the fuzzy factor, Indicates Pixels to The measure of the center of a cluster.

4. The CT image reconstruction method according to claim 3, characterized in that: The said Pixels to The measure of the cluster center satisfies the following formula: in, Indicates Pixels to The measure of the cluster centers, Indicates The spatial frequency of pixels, Indicates Pixels to The horizontal spacing between cluster centers, Indicates Pixels to The vertical spacing of cluster centers, Indicates Pixel to The Euclidean distance between cluster centers, Indicates The horizontal position of the pixel, Indicates The vertical position of the pixel, Indicates The gray value of the pixel, Indicates The gray value of the cluster center point.

5. The CT image reconstruction method according to claim 2, characterized in that: The step of obtaining the number of cluster centers according to the tomographic image comprises the following steps: Constructing a grayscale value histogram according to the grayscale value of the tomographic image; Data processing is performed according to the gray value histogram, and then the number of cluster centers is obtained according to the data processing result.

6. The CT image reconstruction method according to claim 1, characterized in that: The step of extracting characteristic points from the target area and identifying the lesion site according to the characteristic points comprises the following steps: Setting kernels of different sizes, and performing black hat operations on the target area based on the kernels; Extract feature points based on the black hat operation results corresponding to kernels of different sizes; Edge detection is performed with the feature point as the center to identify the lesion site.

7. The CT image reconstruction method according to claim 6, characterized in that: The step of performing edge detection with the feature point as the center to identify the lesion site comprises the following steps: Constructing a filtering window, and performing filtering processing on the target area according to the filtering window; Calculate pixel gradients, and select candidate edge points based on the pixel gradients with the feature point as the center; A screening threshold is set, and the lesion site contour is obtained by combining the screening threshold and the candidate edge points.

8. The CT image reconstruction method according to claim 1, characterized in that: The CT image inter-layer registration model satisfies the following formula: in, Representing images With image The registration cost function is Representing images With image The amount of interactive information, Representing images With image The normalized mutual information of Representing images With image The interaction variance of Representing images With image The Euclidean closeness of Representing images With image The maximum and minimum closeness of Representing images With image Hamming closeness.

9. The CT image reconstruction method according to claim 1, characterized in that: The step of stacking the registration results into three-dimensional data comprises the following steps: According to the gray value difference of the registered image, an interpolation method is selected to interpolate the registered image; The three-dimensional data is obtained by stacking the interpolation results.

10. A CT image reconstruction system, characterized in that: The CT image reconstruction system comprises: an input device, an output device, a processor, and a memory, wherein the input device, the output device, the processor, and the memory are interconnected, and the memory comprises program instructions, and the program instructions are used to execute the CT image reconstruction method according to any one of claims 1 to 9.

Images