Cardiovascular three-dimensional optical coherence image system

Abstract

The invention discloses a cardiovascular three-dimensional optical coherence imaging system, and belongs to the technical field of medical instruments. A cardiovascular three-dimensional optical coherence imaging system comprises a host integration unit and an auxiliary diagnosis unit. The system solves the problem that accurate diagnosis and treatment requirements are difficult to meet in the prior art, interference signals are generated through the host integration unit, high-speed data processing is carried out in combination with the GPU, high-resolution cardiovascular tissue two-dimensional and three-dimensional images can be generated, powerful support is provided for early diagnosis and treatment of diseases, and the system is suitable for popularization and application. The auxiliary diagnosis unit establishes a cardiovascular disease diagnosis knowledge base based on a large amount of clinical case data, automatically identifies plaque types in a three-dimensional image by using a machine learning algorithm, can quickly and accurately provide diagnosis reference for doctors, and calculates a calcification radian score based on the distribution and geometric morphology of calcified plaques, so that the accuracy of diagnosis is improved. And the severity of lesion can be evaluated more accurately by quantitatively analyzing the poor attachment degree of the stent and the tissue prolapse volume.

Description

Technical Field

[0001] The present invention relates to the technical field of medical devices, and in particular to a cardiovascular three-dimensional optical coherence imaging system. Background Art

[0002] In the diagnosis and treatment of cardiovascular diseases, traditional imaging technologies such as digital subtraction angiography (DSA) and intravascular ultrasound (IVUS) can provide information on the internal structure of blood vessels to a certain extent. However, they have problems such as limited resolution, insufficient three-dimensional imaging capabilities, and the inability to clearly observe the microstructure of the blood vessel wall, making it difficult to meet the needs of precise diagnosis and treatment. Therefore, we propose a cardiovascular three-dimensional optical coherence imaging system to address this problem. Summary of the Invention

[0003] The purpose of the present invention is to provide a cardiovascular three-dimensional optical coherence imaging system that generates two-dimensional and three-dimensional images of cardiovascular tissue to comprehensively display the internal structure and pathological conditions of the cardiovascular system. It can also automatically identify plaque types, quantitatively analyze the severity of lesions, and provide doctors with diagnostic references and treatment recommendations, thereby solving the problems raised in the above-mentioned background technology.

[0004] To achieve the above-mentioned object, the present invention provides the following technical solutions: a cardiovascular three-dimensional optical coherence imaging system, the system comprising a host integrated unit and an auxiliary diagnosis unit, the host integrated unit comprising a laser light source module and an acquisition and processing module;

[0005] The laser light source module is configured to split the light beam generated by the laser light source into a reference beam and a signal beam via a beam splitter, receive the returned reference beam and the signal beam reflected or scattered by the cardiovascular tissue to generate an interference signal;

[0006] The acquisition and processing module is configured to acquire the generated interference signals, combine with the GPU, and integrate optical coherence imaging technology and laser scanning confocal technology to generate two-dimensional and three-dimensional images of cardiovascular tissue;

[0007] The auxiliary diagnosis unit is configured to establish a cardiovascular disease diagnosis knowledge base based on a large amount of clinical case data, and automatically identify plaque types in three-dimensional images, including ruptured plaques, calcified plaques, and vulnerable plaques, using a machine learning algorithm based on the morphological and texture features of optical coherence tomography images;

[0008] At the same time, based on the distribution and geometric morphology of calcified plaques, the calcification arc score is calculated, the degree of stent poor adhesion and tissue prolapse volume are quantitatively analyzed, and a visual report is generated.

[0009] Furthermore, the acquisition and processing module includes:

[0010] A signal acquisition module is configured to capture the interference signal generated by the laser light source module, and perform preliminary processing such as amplification and filtering on the collected interference signal to remove noise interference;

[0011] The data processing module is configured as a GPU-based parallel computing architecture. It uses the parallel computing power of the GPU to calculate the large number of collected interference signals. It uses the Fourier transform algorithm to extract cardiovascular tissue information and optical characteristic parameters from the interference signals and construct a two-dimensional cross-sectional image of the tissue.

[0012] an image fusion module configured to acquire a structural image of cardiovascular tissue using optical coherence imaging technology, acquire image data at the cellular and tissue levels using laser scanning confocal technology, fuse the structural image of cardiovascular tissue with the image data at the cellular and tissue levels using an image registration algorithm, determine the positional relationship and corresponding points of the two images in a spatial coordinate system, and align the two images at the pixel level;

[0013] a three-dimensional reconstruction module configured to construct a three-dimensional model of the cardiovascular tissue based on the fused data using a three-dimensional reconstruction algorithm, wherein the three-dimensional reconstruction algorithm includes a voxel reconstruction technology and a surface rendering algorithm;

[0014] In the voxel reconstruction process, each voxel is assigned a corresponding optical characteristic value and position coordinate to construct a three-dimensional data set with spatial volume information;

[0015] The surface rendering algorithm displays the surface morphology and internal structure contours of cardiovascular tissue based on the reflective characteristics of the tissue surface.

[0016] The image optimization module is configured to perform post-processing optimization on the generated two-dimensional and three-dimensional images using an image enhancement algorithm. At the same time, it performs image noise reduction to further remove artifacts and noise that may be generated during the reconstruction process. The image enhancement algorithm includes contrast adjustment, sharpening, and edge detection.

[0017] Furthermore, the step of sharpening the generated two-dimensional and three-dimensional images includes:

[0018] Extract 2D and 3D images;

[0019] Performing grayscale processing on the two-dimensional and three-dimensional images to obtain the two-dimensional and three-dimensional images after the grayscale processing as images to be processed;

[0020] Extracting the grayscale value of each pixel contained in the image to be processed;

[0021] Comparing the grayscale value of each pixel in the image to be processed with a preset grayscale reference value;

[0022] Filter pixels whose grayscale values are lower than a preset grayscale reference value as candidate pixels;

[0023] Performing ratio processing on the grayscale value of the candidate pixel and the grayscale reference value to obtain a first grayscale ratio;

[0024] Obtaining the grayscale standard deviation of the candidate pixel points using the grayscale value of the candidate pixel points;

[0025] Obtain the grayscale standard deviation of the grayscale values of all pixels contained in the image to be processed;

[0026] Performing ratio processing on the grayscale standard deviation of the candidate pixel point and the grayscale standard deviation corresponding to the image to be processed to obtain a second grayscale ratio;

[0027] comparing the first grayscale ratio with the second grayscale ratio;

[0028] Retrieving a candidate pixel point corresponding to a first grayscale ratio not lower than a second grayscale ratio as a target pixel point;

[0029] Using the grayscale value corresponding to the target pixel point to set an image division coefficient, and using the image division coefficient to divide the image to be processed to obtain multiple image blocks;

[0030] Each image block contained in the image to be processed is traversed in sequence, and each image block is sharpened until the image to be processed is sharpened.

[0031] Furthermore, the grayscale value corresponding to the target pixel point is used to set an image division coefficient, and the image to be processed is divided using the image division coefficient to obtain multiple image blocks, including:

[0032] Retrieve the grayscale value corresponding to each target pixel;

[0033] Retrieving a preset initial image block size from a database, wherein the initial image block size is n×n; wherein n represents the number of pixels contained in the side length of the image block;

[0034] Extract the standard deviation of the grayscale values of candidate pixels;

[0035] Obtaining an image division coefficient by using the grayscale value of each target pixel and the standard deviation of the grayscale values of the candidate pixels;

[0036] Adjusting the preset initial image block size using the image division coefficient;

[0037] The image to be processed is divided according to the adjusted image block size.

[0038] Furthermore, the auxiliary diagnosis unit includes:

[0039] a knowledge base building module configured to collect a large amount of clinical case data from multiple cardiovascular disease hospitals and research institutions to build a cardiovascular disease diagnosis knowledge base, including patients' medical history, symptoms, diagnosis results, treatment processes, and corresponding optical coherence tomography images;

[0040] An image feature analysis module is configured to extract morphological features of cardiovascular tissue and plaques based on optical coherence tomography images, including the boundary shape of the plaque, the size and morphological changes of the vascular lumen, and the length and width of the plaque. The module also analyzes the texture features of the optical coherence tomography images, reflecting the characteristics of the internal microstructure of cardiovascular tissue by calculating indicators such as the distribution of image grayscale levels, texture uniformity, and contrast. Different types of plaques have different texture manifestations.

[0041] a plaque recognition module configured to build a model using a machine learning algorithm to automatically identify plaque types in a three-dimensional image, including ruptured plaques, calcified plaques, and vulnerable plaques;

[0042] a data analysis module configured to calculate a calcification arc score based on the distribution and geometry of calcified plaques, quantitatively analyze the degree of stent malapposition and tissue prolapse volume, perform risk assessment based on the calcification arc score and quantitative analysis results, and provide treatment recommendations based on the risk assessment results, including but not limited to adjusting the drug treatment plan or conducting further interventional treatment;

[0043] The visualization module is configured to automatically generate text content of a report based on the analysis results of the data analysis module, including patient information, plaque type identification results, calcification arc score, and quantitative analysis data of stent malapposition degree and tissue prolapse volume.

[0044] Furthermore, the image optimization module includes:

[0045] an image evaluation module configured to establish an image quality evaluation index system to evaluate the quality of the generated two-dimensional and three-dimensional images, wherein the evaluation indexes include image resolution, signal-to-noise ratio, contrast, and edge definition;

[0046] The evaluation feedback module is configured to perform comprehensive calculations based on the results of various evaluation indicators according to pre-set weight coefficients to generate image quality scores, and organize the evaluation data of various indicators and the image quality scores into a feedback report for feedback;

[0047] The optimization and adjustment module is configured to take corresponding optimization measures for different problems based on the feedback report, including but not limited to adjusting the algorithm parameters in the image generation and processing process, and adjusting the parameters of the laser light source module.

[0048] Furthermore, the knowledge base establishment module includes:

[0049] A data annotation module is configured to annotate the collected optical coherence tomography images. The annotation content includes plaque type, location, size, morphological characteristics, and the presence of pathological conditions such as stent malapposition and tissue prolapse. The annotation results are used for model training and optimization;

[0050] The update and maintenance module is configured to establish a dynamic knowledge base update mechanism, regularly collect new clinical case data, update and maintain the cardiovascular disease diagnosis knowledge base, and at the same time, monitor and clean the data in the cardiovascular disease diagnosis knowledge base to remove erroneous or outdated data.

[0051] Furthermore, the plaque identification module includes:

[0052] a model training module configured to build a model using a machine learning algorithm and train the model using annotated data in a cardiovascular disease diagnosis knowledge base;

[0053] During the training process, the extracted patch morphology and texture features are used as input feature vectors, and the corresponding patch type is used as the output label. The model parameters are adjusted through the optimization algorithm to enable the model to learn the characteristic patterns of different types of patches.

[0054] The model validation module is configured to use the cross-validation method to validate and evaluate the trained model, and optimize the model based on the validation results, including adjusting the model structure, parameters and feature selection;

[0055] The automatic recognition module is configured to input the three-dimensional image into the trained machine learning model. The model automatically recognizes the plaques in the image and outputs the type of plaques and the location information of the plaques in the three-dimensional image.

[0056] Furthermore, the data analysis module includes:

[0057] The analysis and quantification module is configured to calculate the calcification arc score based on the distribution and geometry of calcified plaques, and to quantitatively analyze the degree of stent malapposition and tissue prolapse volume. Specifically:

[0058] Calcified plaque analysis stage: Set an appropriate grayscale threshold and use the threshold segmentation algorithm to preliminarily screen out possible calcified plaque areas in the 3D image;

[0059] Extract the geometric features of calcified plaques, including area, volume, perimeter, major diameter, minor diameter, thickness, and curvature. The extracted features are used to describe the size, shape, and distribution of calcified plaques on the vessel wall.

[0060] Based on the extracted geometric features, the calcification arc score is calculated according to the pre-set scoring criteria;

[0061] Stent malapposition analysis stage: Use image edge detection algorithms to extract the edge contour of the stent and determine the position and shape of the stent within the blood vessel;

[0062] Image segmentation technology is used to determine the gap area between the stent and the vessel wall, calculate the area and width parameters of the gap, and quantify the degree of stent malapposition;

[0063] Tissue prolapse volume analysis stage: Identify tissue prolapse at the edge and inside of the stent and calculate the volume of the prolapsed tissue using a volume measurement algorithm;

[0064] A risk assessment module is configured to establish a risk assessment model based on a large amount of collected clinical case data. The risk assessment model evaluates the pathological risk of the calcification arc score, the degree of stent malapposition, and the quantitative analysis results of tissue prolapse volume, and classifies the risk into different levels, including low risk, medium risk, and high risk;

[0065] The recommendation generation module is configured to formulate corresponding treatment recommendations for different risk levels in advance, and generate corresponding treatment recommendations for patients based on the evaluation results.

[0066] Furthermore, the laser light source module is specifically:

[0067] The laser light source module realizes imaging of the sample under test based on the principle of optical interference, and is composed of a laser light source, a beam splitter, and a beam guiding and receiving component;

[0068] The beam generated by the laser light source is divided into a reference beam and a signal beam after passing through a beam splitter;

[0069] The reference beam is reflected by a mirror in the laser light source module and returns to the beam splitter. The signal beam is guided by the laser light source module and transmitted to the sample area under test. When the signal beam hits the surface of the cardiovascular tissue under test, it is reflected or scattered inside the tissue. The light modulated by the tissue returns to the beam splitter and reunites with the reference beam.

[0070] At the beam splitter, the two beams of light accurately converge again at the meeting area to produce interference, forming interference fringes and then generating an interference signal.

[0071] Compared with the prior art, the present invention has the following beneficial effects:

[0072] The present invention generates interference signals through the host integrated unit and combines it with a GPU for high-speed data processing. It can generate high-resolution two-dimensional and three-dimensional images of cardiovascular tissue, accurately displaying the subtle structure and lesions inside the cardiovascular system, facilitating the early detection of cardiovascular diseases and providing strong support for the early diagnosis and treatment of the disease. The auxiliary diagnosis unit establishes a cardiovascular disease diagnosis knowledge base based on a large amount of clinical case data, and uses machine learning algorithms to automatically identify plaque types in three-dimensional images. It can quickly and accurately provide doctors with diagnostic references, reduce the occurrence of human misjudgment, and improve diagnostic efficiency and accuracy. The calcification arc score is calculated based on the distribution and geometric morphology of calcified plaques, and the degree of stent poor adhesion and tissue prolapse volume are quantitatively analyzed, enabling doctors to more accurately assess the severity of the lesion and improve treatment efficacy and prognosis quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 This is a diagram of the overall structure of the cardiovascular three-dimensional optical coherence imaging system of the present invention. DETAILED DESCRIPTION

[0074] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0075] To address the technical issues that existing imaging technologies can provide information on the internal structure of blood vessels to a certain extent, but have limited resolution, insufficient 3D imaging capabilities, and inability to clearly observe the microstructure of the blood vessel wall, making it difficult to meet the needs of accurate diagnosis and treatment, please refer to Figure 1 , this embodiment provides the following technical solutions:

[0076] A cardiovascular three-dimensional optical coherence imaging system, comprising a host integrated unit and an auxiliary diagnosis unit, wherein the host integrated unit comprises a laser light source module and an acquisition and processing module;

[0077] The laser light source module is configured to split the light beam generated by the laser light source into a reference beam and a signal beam via a beam splitter, receive the returned reference beam and the signal beam reflected or scattered by the cardiovascular tissue under test, and generate an interference signal, specifically:

[0078] The laser light source module realizes imaging of the sample under test based on the principle of optical interference, and is composed of a laser light source, a beam splitter, and a beam guiding and receiving component;

[0079] The beam generated by the laser light source is divided into a reference beam and a signal beam after passing through a beam splitter;

[0080] The reference beam is reflected by a mirror in the laser light source module and returns to the beam splitter. The signal beam is guided by the laser light source module and transmitted to the sample area under test. When the signal beam hits the surface of the cardiovascular tissue under test, it is reflected or scattered inside the tissue. The light modulated by the tissue returns to the beam splitter and reunites with the reference beam.

[0081] At the beam splitter, the two beams of light accurately converge again at the meeting area to produce interference, forming interference fringes and then generating an interference signal;

[0082] The acquisition and processing module is configured to acquire the generated interference signals, combine with the GPU, and integrate optical coherence imaging technology and laser scanning confocal technology to generate two-dimensional and three-dimensional images of cardiovascular tissue;

[0083] The auxiliary diagnosis unit is configured to establish a cardiovascular disease diagnosis knowledge base based on a large amount of clinical case data, and automatically identify plaque types in three-dimensional images, including ruptured plaques, calcified plaques, and vulnerable plaques, using a machine learning algorithm based on the morphological and texture features of optical coherence tomography images;

[0084] At the same time, based on the distribution and geometric morphology of calcified plaques, the calcification arc score is calculated, the degree of stent poor adhesion and tissue prolapse volume are quantitatively analyzed, and a visual report is generated.

[0085] In this embodiment, the laser light source module is mainly composed of a laser light source, a beam splitter, and a beam guiding and receiving component. First, the laser light source generates a beam of laser light with a specific wavelength and high coherence. After passing through the beam splitter, this laser beam is divided into two parts: a reference beam and a signal beam. This splitting process is the key foundation for achieving subsequent interferometric imaging.

[0086] The signal beam is precisely guided by the laser light source module and transmitted to the cardiovascular tissue area being measured. When the signal beam hits the surface of the cardiovascular tissue, complex reflection and scattering phenomena occur within the tissue. Some light is specularly reflected from the tissue surface, while more light enters the tissue. After multiple scattering, some of the scattered light returns to the tissue surface and propagates outward. The light modulated by the tissue, carrying rich structural and physiological information of the cardiovascular tissue, returns to the spectrometer.

[0087] At the same time, the reference beam is reflected and controlled by a dedicated mirror inside the laser light source module. Parameters such as the mirror's position, angle, and number of reflections are precisely calculated and optimized to ensure that the reference beam accurately reconverges with the returning signal beam at the beam splitter at the appropriate time and phase.

[0088] At the beam splitter, the reference beam and the signal beam, which has been reflected or scattered by cardiovascular tissue, meet. Due to the difference in optical path length between the two beams, an optical path difference is generated, which in turn causes interference in the area where they meet. The generation of interference is based on the wave principle of light. When two coherent light beams meet, the light waves of the two beams will superimpose, forming interference fringes of alternating light and dark. The interference fringes contain optical information of the cardiovascular tissue being measured. By detecting and analyzing the interference signal, imaging of the cardiovascular tissue structure and pathological conditions can be achieved.

[0089] In summary, the laser light source module, a series of laser reflection-based design and implementation methods, can provide a high-quality interference signal source for the entire cardiovascular three-dimensional optical coherence imaging system, thereby ensuring the accuracy and reliability of the generated two-dimensional and three-dimensional images of cardiovascular tissue, laying a solid foundation for the precise diagnosis and treatment of cardiovascular diseases.

[0090] The technical effects of the above technical solution are as follows: the laser light source module of the host integrated unit generates a reference beam and a signal beam through a spectrometer to generate an interference signal, which provides a basic signal source for subsequent image generation and realizes the effective acquisition of optical signals of cardiovascular tissue. The acquisition and processing module, combined with the GPU and multiple technology integration, can efficiently convert the interference signal into two-dimensional and three-dimensional images, improving the speed and quality of image generation, so that the structure of cardiovascular tissue can be presented intuitively and clearly, helping doctors to observe the cardiovascular morphology more accurately. The knowledge base established by the auxiliary diagnosis unit provides rich data support for diagnosis, automatically identifies plaque types based on image features and machine learning algorithms, improves the accuracy and efficiency of diagnosis, avoids the subjectivity and errors of manual identification, calculates the calcification arc score and quantitatively analyzes the degree of stent poor adhesion and tissue prolapse volume, realizes the accurate quantitative evaluation of cardiovascular disease-related indicators, and provides doctors with a more targeted and scientific basis for formulating treatment plans.

[0091] Acquisition and processing module, including:

[0092] A signal acquisition module is configured to capture the interference signal generated by the laser light source module, and perform preliminary processing such as amplification and filtering on the collected interference signal to remove noise interference;

[0093] The data processing module is configured as a GPU-based parallel computing architecture. It uses the parallel computing power of the GPU to calculate the large number of collected interference signals. It uses the Fourier transform algorithm to extract cardiovascular tissue information and optical characteristic parameters from the interference signals and construct a two-dimensional cross-sectional image of the tissue.

[0094] an image fusion module configured to acquire a structural image of cardiovascular tissue using optical coherence imaging technology, acquire image data at the cellular and tissue levels using laser scanning confocal technology, fuse the structural image of cardiovascular tissue with the image data at the cellular and tissue levels using an image registration algorithm, determine the positional relationship and corresponding points of the two images in a spatial coordinate system, and align the two images at the pixel level;

[0095] a three-dimensional reconstruction module configured to construct a three-dimensional model of the cardiovascular tissue based on the fused data using a three-dimensional reconstruction algorithm, wherein the three-dimensional reconstruction algorithm includes a voxel reconstruction technology and a surface rendering algorithm;

[0096] During the voxel reconstruction process, each voxel is assigned a corresponding optical property value and position coordinates to construct a three-dimensional data set with spatial volume information, which can reflect detailed information such as the volume and density of cardiovascular tissue in space.

[0097] The surface rendering algorithm displays the surface morphology and internal structural contours of cardiovascular tissue based on the reflective characteristics of the tissue surface, providing a more intuitive understanding of the three-dimensional structure of cardiovascular tissue.

[0098] The image optimization module is configured to perform post-processing optimization on the generated two-dimensional and three-dimensional images using an image enhancement algorithm. At the same time, it performs image noise reduction to further remove artifacts and noise that may be generated during the reconstruction process. The image enhancement algorithm includes contrast adjustment, sharpening, and edge detection.

[0099] The technical effects of the above technical solution are as follows: the signal acquisition module can capture the interference signal generated by the laser light source module, and perform preliminary processing of amplification and filtering on it, effectively removing noise interference, thereby ensuring the quality of the collected signal and providing an accurate and reliable signal source for subsequent image generation and other processing. The data processing module is based on the parallel computing architecture of GPU, which can quickly calculate a large number of collected interference signals. Through the Fourier transform algorithm, it can accurately extract the information and optical characteristic parameters of cardiovascular tissue from the interference signal, and then construct a two-dimensional cross-sectional image of cardiovascular tissue, realizing efficient conversion from signal to two-dimensional image, and improving the efficiency and quality of image generation. The image fusion module combines the cardiovascular tissue structure image obtained by optical coherence imaging technology with the laser scanning The combination of cellular and tissue-level image data acquired by confocal technology can more comprehensively and meticulously display the structure of cardiovascular tissue and detailed information at the cellular and tissue levels, providing a richer image data basis for subsequent diagnosis and analysis. The 3D reconstruction module then uses a 3D reconstruction algorithm to construct a 3D model of cardiovascular tissue based on the fused data and generate a 3D image. The image optimization module uses an image enhancement algorithm to perform post-processing on the generated 2D and 3D images, further highlighting the key details and features of the image, making the image clearer and easier to identify. At the same time, image noise reduction is performed to effectively remove artifacts and noise that may be generated during the 3D reconstruction process, further improving the quality and reliability of the image, and helping doctors to more accurately observe and analyze the pathological conditions of cardiovascular tissue.

[0100] Specifically, the step of sharpening the generated two-dimensional and three-dimensional images includes:

[0101] Extract 2D and 3D images;

[0102] Performing grayscale processing on the two-dimensional and three-dimensional images to obtain the two-dimensional and three-dimensional images after the grayscale processing as images to be processed;

[0103] Extracting the grayscale value of each pixel contained in the image to be processed;

[0104] Comparing the grayscale value of each pixel in the image to be processed with a preset grayscale reference value;

[0105] Filter pixels whose grayscale values are lower than a preset grayscale reference value as candidate pixels;

[0106] Performing ratio processing on the grayscale value of the candidate pixel and the grayscale reference value to obtain a first grayscale ratio;

[0107] Obtaining the grayscale standard deviation of the candidate pixel points using the grayscale value of the candidate pixel points;

[0108] Obtain the grayscale standard deviation of the grayscale values of all pixels contained in the image to be processed;

[0109] Performing ratio processing on the grayscale standard deviation of the candidate pixel point and the grayscale standard deviation corresponding to the image to be processed to obtain a second grayscale ratio;

[0110] comparing the first grayscale ratio with the second grayscale ratio;

[0111] Retrieving a candidate pixel point corresponding to a first grayscale ratio not lower than a second grayscale ratio as a target pixel point;

[0112] Using the grayscale value corresponding to the target pixel point to set an image division coefficient, and using the image division coefficient to divide the image to be processed to obtain multiple image blocks;

[0113] Each image block contained in the image to be processed is traversed in sequence, and each image block is sharpened until the image to be processed is sharpened.

[0114] The technical effect of the above-mentioned technical solution is as follows: by screening candidate pixels with grayscale values below a preset reference value, obtaining and comparing the relevant grayscale ratios, determining the target pixel, and then setting the image partition coefficient to divide the image into blocks. Sharpening these image blocks effectively enhances detail in darker areas of the image (determined by the candidate pixels), making image edges clearer and textures more pronounced, improving image clarity and readability, and facilitating subsequent identification and analysis of target structures (such as blood vessels). This solution comprehensively considers pixel grayscale values and their grayscale standard deviations to target specific areas of the image (represented by the image block containing the target pixel). Compared to uniformly sharpening the entire image, this approach avoids problems such as increased image noise and distortion caused by oversharpening. While enhancing image detail, it maintains overall image quality, resulting in a more natural visual effect after sharpening and better suited to applications such as medical diagnosis. Based on grayscale processing, the processing targets (target pixels and corresponding image blocks) are determined through a rigorous process of pixel grayscale value screening and grayscale ratio comparison, making the sharpening operation more targeted and accurate. It can precisely target areas requiring detail enhancement, reducing unnecessary interference with normal image regions, improving the effectiveness of sharpening processing, and providing more reliable image data for subsequent image-based analysis and diagnosis. This technical solution processes images based on characteristics such as the grayscale value and grayscale standard deviation of the image's pixels, independent of specific image types or imaging conditions. Whether for two-dimensional or three-dimensional vascular ultrasound images, the appropriate sharpening area and method can be determined based on their inherent grayscale characteristics. This highly adaptable approach can be widely applied to sharpening images from diverse sources and characteristics.

[0115] Specifically, the grayscale value corresponding to the target pixel is used to set an image division coefficient, and the image to be processed is divided using the image division coefficient to obtain multiple image blocks, including:

[0116] Retrieve the grayscale value corresponding to each target pixel;

[0117] Retrieving a preset initial image block size from a database, wherein the initial image block size is n×n; wherein n represents the number of pixels contained in the side length of the image block;

[0118] Extract the standard deviation of the grayscale values of candidate pixels;

[0119] Obtaining an image division coefficient by using the grayscale value of each target pixel and the standard deviation of the grayscale values of the candidate pixels;

[0120] The image division coefficient is obtained by the following formula:

[0121]

[0122] Among them, J represents the image division coefficient; m represents the number of target pixels; Z i Represents the grayscale value corresponding to the i-th target pixel; Z b Indicates the standard deviation of the grayscale value of the candidate pixel; specifically, The grayscale discreteness of the candidate pixel is associated with the grayscale value of the target pixel to build a relative relationship. The denominator is added with 1 to avoid Z i = 0, the denominator is 0, which also makes the range of the formula more reasonable. This formula measures the relative proportional relationship of the grayscale value of the target pixel taking into account the grayscale discreteness of the candidate pixels. The range of the sine function sin is [-1,1]. Multiply the result by The sine function then serves as the independent variable. On the one hand, the nonlinear characteristics of the sine function can perform a nonlinear transformation on the relative proportional relationship obtained previously, highlighting the differences between different combinations of target pixel grayscale values and candidate pixel grayscale discreteness. On the other hand, limiting the result to the range of [-1, 1] facilitates subsequent summation and averaging operations, and also allows this part of the calculation result to participate in the calculation of the image partition coefficient in an appropriate proportional form. The sine transform results of the m target pixels are summed, comprehensively considering all target pixels and reflecting the combined effect of the grayscale values of the multiple target pixels and the grayscale dispersion of the candidate pixels. The average is then divided by m to normalize the combined effect of all target pixels, resulting in a single value that represents the overall characteristics of all target pixels, namely the image partition coefficient J. This value reflects the comprehensive grayscale characteristics of the target pixels in the image and is used to determine the appropriate image partitioning method. This formula comprehensively considers the target pixel grayscale value and the grayscale standard deviation of the candidate pixels, and calculates the image partition coefficient using a specific calculation method. It accurately reflects the grayscale distribution and discrete characteristics of the target area in the image, making the image partition coefficient more accurate to the actual image. The image blocks partitioned based on this method can better match the image content structure, improving the relevance and accuracy of subsequent image processing (such as sharpening). Because the formula calculates the image partition coefficient based on the image's own grayscale characteristic parameters (the target pixel grayscale value and the candidate pixel grayscale standard deviation), it can dynamically generate appropriate coefficients for images with different grayscale characteristics. Whether the image has uniform grayscale or complex variations, the method can be adapted to its characteristics and perform reasonable image block division, enhancing its adaptability and versatility for diverse images. The use of a sine function and the summation and averaging operation in the formula prevent the excessive influence of individual extreme pixel grayscale values on the final image division coefficient. By comprehensively processing multiple target pixels, the calculation results are more stable and reliable, preventing unreasonable fluctuations in the image division coefficient caused by outliers from a few pixels, thus ensuring the stability and reliability of the image division process.

[0123] Adjusting the preset initial image block size using the image division coefficient;

[0124] The adjusted image block size is obtained by the following formula:

[0125] n t =(1+J)*n

[0126] Among them, n t Indicates the number of pixels contained in the side length corresponding to the adjusted image block size, and n t is rounded up; J represents the image division coefficient; n represents the number of pixels contained in the side length of the image block;

[0127] The image to be processed is divided according to the adjusted image block size.

[0128] The technical effect of the above technical solution is: determining an image partition coefficient based on the grayscale value of the target pixel and the standard deviation of the grayscale values of the candidate pixels, and then adjusting the initial image block size accordingly to divide the image. This method can determine the appropriate image block size based on the image's inherent grayscale feature information. Compared to fixed-size partitioning, the partitioned image blocks can better match the image's actual content and structure, more accurately capture image details, and improve the relevance and accuracy of subsequent processing. Because the image partition coefficient is related to the image's inherent grayscale characteristics, the image block size can be dynamically adjusted for images with different grayscale distributions and features. Whether the grayscale changes are smooth or complex, the appropriate partitioning method can be found, making this technical solution highly adaptable to images of different types and imaging conditions, broadening its application range. If the image block size is too large, it will contain too many different feature areas, resulting in poor subsequent processing effects such as sharpening; if the size is too small, the processing load will increase and efficiency will decrease. By adjusting the image block size based on the image's grayscale features, it is possible to ensure processing results while avoiding unnecessary waste of computing resources, achieving a good balance between processing efficiency and results, and improving overall image processing performance. The image optimization module includes:

[0129] an image evaluation module configured to establish an image quality evaluation index system to evaluate the quality of the generated two-dimensional and three-dimensional images, wherein the evaluation indexes include image resolution, signal-to-noise ratio, contrast, and edge definition;

[0130] The evaluation feedback module is configured to perform comprehensive calculations based on the results of various evaluation indicators according to pre-set weight coefficients to generate image quality scores, and organize the evaluation data of various indicators and the image quality scores into a feedback report for feedback;

[0131] The optimization and adjustment module is configured to take corresponding optimization measures for different problems based on the feedback report, including but not limited to adjusting the algorithm parameters in the image generation and processing process, and adjusting the parameters of the laser light source module.

[0132] The technical effect of the above technical solution is: the image quality evaluation index system constructed by the image evaluation module can accurately quantify the quality of two-dimensional and three-dimensional images, providing a scientific and objective quantitative basis for image optimization. The evaluation feedback module generates an image quality score based on the evaluation index results and comprehensive calculation according to the weights, and compiles it into a feedback report, which can clearly understand the current image quality status and the specific performance of each indicator, and provide a quantitative reference for the subsequent optimization and adjustment module to avoid blind adjustment and improve optimization efficiency. For example: changing the parameters of the Fourier transform algorithm to optimize image information extraction, or adjusting the image registration algorithm parameters to improve image fusion accuracy. At the same time, adjustment of the laser light source module parameters, such as adjusting the wavelength and power of the laser light source, can improve the interference signal quality from the source.

[0133] Auxiliary diagnostic unit, including:

[0134] a knowledge base building module configured to collect a large amount of clinical case data from multiple cardiovascular disease hospitals and research institutions to build a cardiovascular disease diagnosis knowledge base, including patients' medical history, symptoms, diagnosis results, treatment processes, and corresponding optical coherence tomography images;

[0135] An image feature analysis module is configured to extract morphological features of cardiovascular tissue and plaques based on optical coherence tomography images, including the boundary shape of the plaque, the size and morphological changes of the vascular lumen, and the length and width of the plaque. The module also analyzes the texture features of the optical coherence tomography images, reflecting the characteristics of the internal microstructure of cardiovascular tissue by calculating indicators such as the distribution of image grayscale levels, texture uniformity, and contrast. Different types of plaques have different texture manifestations.

[0136] a plaque recognition module configured to build a model using a machine learning algorithm to automatically identify plaque types in a three-dimensional image, including ruptured plaques, calcified plaques, and vulnerable plaques;

[0137] a data analysis module configured to calculate a calcification arc score based on the distribution and geometry of calcified plaques, quantitatively analyze the degree of stent malapposition and tissue prolapse volume, perform risk assessment based on the calcification arc score and quantitative analysis results, and provide treatment recommendations based on the risk assessment results, including but not limited to adjusting the drug treatment plan or conducting further interventional treatment;

[0138] The visualization module is configured to automatically generate text content of a report based on the analysis results of the data analysis module, including patient information, plaque type identification results, calcification arc score, and quantitative analysis data of stent malapposition degree and tissue prolapse volume.

[0139] The technical effects of the above technical solution are as follows: the knowledge base establishment module establishes a cardiovascular disease diagnosis knowledge base based on a large amount of clinical case data, providing rich data support and knowledge reserves for diagnosis, which helps to improve the comprehensiveness and accuracy of diagnosis; the image feature analysis module can accurately extract the morphological characteristics of cardiovascular tissue and plaques based on optical coherence tomography images, thereby providing a key basis for subsequent plaque identification; the plaque identification module uses a machine learning algorithm to build a model to realize the automatic identification of ruptured plaques, calcified plaques and vulnerable plaques in three-dimensional images, changing the subjectivity and errors of traditional manual identification, and improving diagnostic efficiency and Accuracy helps to timely discover potential cardiovascular disease risks. The data analysis module calculates the calcification curvature score based on the distribution and geometric morphology of calcified plaques, quantitatively analyzes the degree of stent poor apposition and tissue prolapse volume, and conducts risk assessment based on this, providing doctors with more accurate disease assessment methods and helping to formulate personalized treatment plans. Finally, the visualization module automatically generates report text content based on the analysis results, covering patient information, plaque type identification results, calcification curvature score, and quantitative analysis data of stent poor apposition and tissue prolapse, so that doctors can quickly obtain key information and improve the efficiency of diagnosis and treatment.

[0140] Knowledge base building module, including:

[0141] A data annotation module is configured to annotate the collected optical coherence tomography images. The annotation content includes plaque type, location, size, morphological characteristics, and the presence of pathological conditions such as stent malapposition and tissue prolapse. The annotation results are used for model training and optimization;

[0142] The update and maintenance module is configured to establish a dynamic knowledge base update mechanism, regularly collect new clinical case data, update and maintain the cardiovascular disease diagnosis knowledge base, and at the same time, monitor and clean the data in the cardiovascular disease diagnosis knowledge base to remove erroneous or outdated data.

[0143] The technical effects of the above technical solution are: the data annotation module can perform detailed annotations on the collected optical coherence tomography images, providing accurate annotation data for subsequent model training and optimization, thereby improving the accuracy of model recognition and diagnosis, and the update and maintenance module establishes a dynamic knowledge base update mechanism, regularly collects new case data and updates the knowledge base, while monitoring data quality, cleaning up erroneous and outdated data, and ensuring the timeliness, accuracy and reliability of the knowledge base data, thereby improving the quality of diagnostic recommendations and better assisting doctors in diagnosing cardiovascular diseases.

[0144] Plaque recognition module, including:

[0145] a model training module configured to build a model using a machine learning algorithm and train the model using labeled data in a cardiovascular disease diagnosis knowledge base;

[0146] During the training process, the extracted patch morphology and texture features are used as input feature vectors, and the corresponding patch type is used as the output label. The model parameters are adjusted through the optimization algorithm to enable the model to learn the characteristic patterns of different types of patches.

[0147] The model validation module is configured to use the cross-validation method to validate and evaluate the trained model, and optimize the model based on the validation results, including adjusting the model structure, parameters and feature selection;

[0148] The automatic recognition module is configured to input the three-dimensional image into the trained machine learning model. The model automatically recognizes the plaques in the image and outputs the type of plaques and the location information of the plaques in the three-dimensional image.

[0149] The technical effects of the above technical solution are as follows: the model training module is based on the annotated data in the cardiovascular disease diagnosis knowledge base, takes the extracted plaque morphology and texture features as input feature vectors, and the corresponding plaque type as output label. The model parameters are adjusted through the optimization algorithm, so that the model can accurately learn the characteristic patterns of different types of plaques, which not only improves the recognition accuracy of the model, but also enhances its ability to distinguish different types of plaques, laying a solid foundation for subsequent automatic recognition. The model verification module uses the cross-validation method to verify and evaluate the trained model, which can objectively reflect the performance of the model on different data sets and timely discover the potential of the model. Problems and overfitting can be detected, so that the model can be optimized and adjusted according to the verification results, which can further improve the stability and generalization ability of the model and ensure its accuracy and reliability for different patients and different image data in practical applications. The automatic recognition module can input three-dimensional images into the trained machine learning model to realize fast, automatic and accurate recognition of plaques in the image, and output the plaque type and location information in the three-dimensional image, thereby improving the diagnostic efficiency and reducing the errors caused by human factors. It provides a more accurate and objective basis for the diagnosis of cardiovascular diseases and helps doctors make more timely and accurate clinical diagnosis and treatment decisions.

[0150] Data analysis module, including:

[0151] The analysis and quantification module is configured to calculate the calcification arc score based on the distribution and geometry of calcified plaques, and to quantitatively analyze the degree of stent malapposition and tissue prolapse volume. Specifically:

[0152] Calcified plaque analysis stage: Set an appropriate grayscale threshold and use the threshold segmentation algorithm to preliminarily screen out possible calcified plaque areas in the 3D image;

[0153] Extract the geometric features of calcified plaques, including area, volume, perimeter, major diameter, minor diameter, thickness, and curvature. The extracted features are used to describe the size, shape, and distribution of calcified plaques on the vessel wall, providing accurate data support for subsequent calcification curvature scoring;

[0154] Based on the extracted geometric features, the calcification arc score is calculated according to the pre-set scoring criteria;

[0155] Stent malapposition analysis stage: Use image edge detection algorithms to extract the edge contour of the stent and determine the position and shape of the stent within the blood vessel;

[0156] Image segmentation technology is used to determine the gap area between the stent and the vessel wall, calculate the area and width parameters of the gap, quantify the degree of stent malapposition, and achieve accurate quantitative assessment of the degree of stent malapposition;

[0157] Tissue prolapse volume analysis phase: Identify tissue prolapse at the edge and inside the stent, and use volume measurement algorithms to calculate the volume of prolapsed tissue, providing key quantitative evidence for subsequent risk assessment;

[0158] A risk assessment module is configured to establish a risk assessment model based on a large amount of collected clinical case data. The risk assessment model evaluates the pathological risk of the calcification arc score, the degree of stent malapposition, and the quantitative analysis results of tissue prolapse volume, and classifies the risk into different levels, including low risk, medium risk, and high risk;

[0159] The recommendation generation module is configured to formulate corresponding treatment recommendations for different risk levels in advance, and generate corresponding treatment recommendations for patients based on the evaluation results.

[0160] The technical effects of the above technical solution are as follows: the analysis and quantification module can calculate the calcification curvature score based on the distribution and geometric morphology of calcified plaques, and quantitatively analyze the degree of stent malapposition and tissue prolapse volume, thereby providing a basis for subsequent risk assessment. The risk assessment module is based on a risk assessment model established based on a large amount of clinical case data. It can combine the calcification curvature score with the quantitative analysis results of the degree of stent malapposition and tissue prolapse volume to accurately assess the pathological risk and divide the risk into low risk, medium risk and high risk levels, providing doctors with clear risk grading information to help doctors better understand the severity of the patient's condition and potential risks. The recommendation generation module pre-forms corresponding treatment recommendations for different risk levels. After assessing the patient's risk level, it can quickly and accurately generate corresponding personalized treatment recommendations, providing strong support for doctors to formulate treatment plans, improve the targetedness and effectiveness of treatment, and improve the patient's treatment effect and prognosis.

[0161] Working Principle: Leveraging the powerful computing power of GPUs and integrating optical coherence imaging technology with laser scanning confocal technology, it can quickly and efficiently generate high-resolution two-dimensional and three-dimensional images of cardiovascular tissue, providing doctors with clear and intuitive visualization, helping to more accurately observe the subtle structure and pathological conditions of cardiovascular tissue. A cardiovascular disease diagnostic knowledge base built based on a large amount of clinical case data and the use of machine learning algorithms to automatically identify different types of plaques in three-dimensional images improves diagnostic accuracy and efficiency, reduces subjective errors in manual identification, and facilitates early detection of potential cardiovascular disease risks. The system also calculates the calcification arc score based on the distribution and geometric morphology of calcified plaques, and quantitatively analyzes the degree of stent poor apposition and tissue prolapse volume, achieving accurate quantitative assessment of cardiovascular disease-related pathological characteristics and providing doctors with more comprehensive and objective data support for formulating personalized treatment plans.

[0162] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0163] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A cardiovascular three-dimensional optical coherence imaging system, characterized in that: The system includes a host integrated unit and an auxiliary diagnosis unit, wherein the host integrated unit includes a laser light source module and an acquisition and processing module; The laser light source module is configured to split the light beam generated by the laser light source into a reference beam and a signal beam via a beam splitter, receive the returned reference beam and the signal beam reflected or scattered by the cardiovascular tissue to generate an interference signal; The acquisition and processing module is configured to acquire the generated interference signals, combine with the GPU, and integrate optical coherence imaging technology and laser scanning confocal technology to generate two-dimensional and three-dimensional images of cardiovascular tissue; The generated two-dimensional and three-dimensional images are sharpened, including image extraction, grayscale processing, pixel screening, ratio processing, setting image division coefficients, dividing the image blocks, and sharpening each image block; The image division coefficient is set using the grayscale value of the target pixel, and the size of the initial image block is adjusted accordingly; The auxiliary diagnosis unit is configured to establish a cardiovascular disease diagnosis knowledge base based on a large amount of clinical case data, and automatically identify plaque types in three-dimensional images, including ruptured plaques, calcified plaques, and vulnerable plaques, using a machine learning algorithm based on the morphological and texture features of optical coherence tomography images; At the same time, based on the distribution and geometric morphology of calcified plaques, the calcification arc score is calculated, the degree of stent poor adhesion and tissue prolapse volume are quantitatively analyzed, and a visual report is generated.

2. The cardiovascular three-dimensional optical coherence imaging system according to claim 1, characterized in that: The acquisition and processing module includes: A signal acquisition module is configured to capture the interference signal generated by the laser light source module, and perform preliminary processing such as amplification and filtering on the collected interference signal to remove noise interference; The data processing module is configured as a GPU-based parallel computing architecture. It uses the parallel computing power of the GPU to calculate the large number of collected interference signals. It uses the Fourier transform algorithm to extract cardiovascular tissue information and optical characteristic parameters from the interference signals and construct a two-dimensional cross-sectional image of the tissue. an image fusion module configured to acquire a structural image of cardiovascular tissue using optical coherence imaging technology, acquire image data at the cellular and tissue levels using laser scanning confocal technology, fuse the structural image of cardiovascular tissue with the image data at the cellular and tissue levels using an image registration algorithm, determine the positional relationship and corresponding points of the two images in a spatial coordinate system, and align the two images at the pixel level; a three-dimensional reconstruction module configured to construct a three-dimensional model of the cardiovascular tissue based on the fused data using a three-dimensional reconstruction algorithm, wherein the three-dimensional reconstruction algorithm includes a voxel reconstruction technology and a surface rendering algorithm; In the voxel reconstruction process, each voxel is assigned a corresponding optical characteristic value and position coordinate to construct a three-dimensional data set with spatial volume information; The surface rendering algorithm displays the surface morphology and internal structure contours of cardiovascular tissue based on the reflective characteristics of the tissue surface. The image optimization module is configured to perform post-processing optimization on the generated two-dimensional and three-dimensional images using an image enhancement algorithm. At the same time, it performs image noise reduction to further remove artifacts and noise that may be generated during the reconstruction process. The image enhancement algorithm includes contrast adjustment, sharpening, and edge detection.

3. The cardiovascular three-dimensional optical coherence imaging system according to claim 2, characterized in that: The step of sharpening the generated two-dimensional and three-dimensional images comprises: Extract 2D and 3D images; Performing grayscale processing on the two-dimensional and three-dimensional images to obtain the two-dimensional and three-dimensional images after the grayscale processing as images to be processed; Extracting the grayscale value of each pixel contained in the image to be processed; Comparing the grayscale value of each pixel in the image to be processed with a preset grayscale reference value; Filter pixels whose grayscale values are lower than a preset grayscale reference value as candidate pixels; Performing ratio processing on the grayscale value of the candidate pixel and the grayscale reference value to obtain a first grayscale ratio; Obtaining the grayscale standard deviation of the candidate pixel points using the grayscale value of the candidate pixel points; Obtain the grayscale standard deviation of the grayscale values of all pixels contained in the image to be processed; Performing ratio processing on the grayscale standard deviation of the candidate pixel point and the grayscale standard deviation corresponding to the image to be processed to obtain a second grayscale ratio; comparing the first grayscale ratio with the second grayscale ratio; Retrieving a candidate pixel point corresponding to a first grayscale ratio not lower than a second grayscale ratio as a target pixel point; Using the grayscale value corresponding to the target pixel point to set an image division coefficient, and using the image division coefficient to divide the image to be processed to obtain multiple image blocks; Each image block contained in the image to be processed is traversed in sequence, and each image block is sharpened until the image to be processed is sharpened.

4. The cardiovascular three-dimensional optical coherence imaging system according to claim 3, characterized in that: The image division coefficient is set using the grayscale value corresponding to the target pixel point, and the image to be processed is divided using the image division coefficient to obtain multiple image blocks, including: Retrieve the grayscale value corresponding to each target pixel; Retrieving a preset initial image block size from a database, wherein the initial image block size is n×n; wherein n represents the number of pixels contained in the side length of the image block; Extract the standard deviation of the grayscale values of candidate pixels; Obtaining an image division coefficient by using the grayscale value of each target pixel and the standard deviation of the grayscale values of the candidate pixels; Adjusting the preset initial image block size using the image division coefficient; The image to be processed is divided according to the adjusted image block size.

5. The cardiovascular three-dimensional optical coherence imaging system according to claim 1, characterized in that: The auxiliary diagnosis unit comprises: a knowledge base building module configured to collect a large amount of clinical case data from multiple cardiovascular disease hospitals and research institutions to build a cardiovascular disease diagnosis knowledge base, including patients' medical history, symptoms, diagnosis results, treatment processes, and corresponding optical coherence tomography images; An image feature analysis module is configured to extract morphological features of cardiovascular tissue and plaques based on optical coherence tomography images, including the boundary shape of the plaque, the size and morphological changes of the vascular lumen, and the length and width of the plaque. The module also analyzes the texture features of the optical coherence tomography images, reflecting the characteristics of the internal microstructure of cardiovascular tissue by calculating indicators such as the distribution of image grayscale levels, texture uniformity, and contrast. Different types of plaques have different texture manifestations. a plaque recognition module configured to build a model using a machine learning algorithm to automatically identify plaque types in a three-dimensional image, including ruptured plaques, calcified plaques, and vulnerable plaques; a data analysis module configured to calculate a calcification arc score based on the distribution and geometry of calcified plaques, quantitatively analyze the degree of stent malapposition and tissue prolapse volume, perform risk assessment based on the calcification arc score and quantitative analysis results, and provide treatment recommendations based on the risk assessment results, including but not limited to adjusting the drug treatment plan or conducting further interventional treatment; The visualization module is configured to automatically generate text content of a report based on the analysis results of the data analysis module, including patient information, plaque type identification results, calcification arc score, and quantitative analysis data of stent malapposition degree and tissue prolapse volume.

6. The cardiovascular three-dimensional optical coherence imaging system according to claim 2, characterized in that: The image optimization module includes: an image evaluation module configured to establish an image quality evaluation index system to evaluate the quality of the generated two-dimensional and three-dimensional images, wherein the evaluation indexes include image resolution, signal-to-noise ratio, contrast, and edge definition; The evaluation feedback module is configured to perform comprehensive calculations based on the results of various evaluation indicators according to pre-set weight coefficients to generate image quality scores, and organize the evaluation data of various indicators and the image quality scores into a feedback report for feedback; The optimization and adjustment module is configured to take corresponding optimization measures for different problems based on the feedback report, including but not limited to adjusting the algorithm parameters in the image generation and processing process, and adjusting the parameters of the laser light source module.

7. The cardiovascular three-dimensional optical coherence imaging system according to claim 5, characterized in that: The knowledge base establishment module includes: A data annotation module is configured to annotate the collected optical coherence tomography images. The annotation content includes plaque type, location, size, morphological characteristics, and the presence of pathological conditions such as stent malapposition and tissue prolapse. The annotation results are used for model training and optimization; The update and maintenance module is configured to establish a dynamic knowledge base update mechanism, regularly collect new clinical case data, update and maintain the cardiovascular disease diagnosis knowledge base, and at the same time, monitor and clean the data in the cardiovascular disease diagnosis knowledge base to remove erroneous or outdated data.

8. The cardiovascular three-dimensional optical coherence imaging system according to claim 5, characterized in that: The plaque identification module includes: a model training module configured to build a model using a machine learning algorithm and train the model using annotated data in a cardiovascular disease diagnosis knowledge base; During the training process, the extracted patch morphology and texture features are used as input feature vectors, and the corresponding patch type is used as the output label. The model parameters are adjusted through the optimization algorithm to enable the model to learn the characteristic patterns of different types of patches. The model validation module is configured to use the cross-validation method to validate and evaluate the trained model, and optimize the model based on the validation results, including adjusting the model structure, parameters and feature selection; The automatic recognition module is configured to input the three-dimensional image into the trained machine learning model. The model automatically recognizes the plaques in the image and outputs the type of plaques and the location information of the plaques in the three-dimensional image.

9. The cardiovascular three-dimensional optical coherence imaging system according to claim 5, characterized in that: The data analysis module includes: The analysis and quantification module is configured to calculate the calcification arc score based on the distribution and geometry of calcified plaques, and to quantitatively analyze the degree of stent malapposition and tissue prolapse volume. Specifically: Calcified plaque analysis stage: Set an appropriate grayscale threshold and use the threshold segmentation algorithm to preliminarily screen out possible calcified plaque areas in the 3D image; Extract the geometric features of calcified plaques, including area, volume, perimeter, major diameter, minor diameter, thickness, and curvature. The extracted features are used to describe the size, shape, and distribution of calcified plaques on the vessel wall. Based on the extracted geometric features, the calcification arc score is calculated according to the pre-set scoring criteria; Stent malapposition analysis stage: Use image edge detection algorithms to extract the edge contour of the stent and determine the position and shape of the stent within the blood vessel; Image segmentation technology is used to determine the gap area between the stent and the vessel wall, calculate the area and width parameters of the gap, and quantify the degree of stent malapposition; Tissue prolapse volume analysis stage: Identify tissue prolapse at the edge and inside of the stent and calculate the volume of the prolapsed tissue using a volume measurement algorithm; A risk assessment module is configured to establish a risk assessment model based on a large amount of collected clinical case data. The risk assessment model evaluates the pathological risk of the calcification arc score, the degree of stent malapposition, and the quantitative analysis results of tissue prolapse volume, and classifies the risk into different levels, including low risk, medium risk, and high risk; The recommendation generation module is configured to formulate corresponding treatment recommendations for different risk levels in advance, and generate corresponding treatment recommendations for patients based on the evaluation results.

10. The cardiovascular three-dimensional optical coherence imaging system according to claim 1, characterized in that: The laser light source module is specifically: The laser light source module realizes imaging of the sample under test based on the principle of optical interference, and is composed of a laser light source, a beam splitter, and a beam guiding and receiving component; The beam generated by the laser light source is divided into a reference beam and a signal beam after passing through a beam splitter; The reference beam is reflected by a mirror in the laser light source module and returns to the beam splitter. The signal beam is guided by the laser light source module and transmitted to the sample area under test. When the signal beam hits the surface of the cardiovascular tissue under test, it is reflected or scattered inside the tissue. The light modulated by the tissue returns to the beam splitter and reunites with the reference beam. At the beam splitter, the two beams of light accurately converge again at the meeting area to produce interference, forming interference fringes and then generating an interference signal.

Images