A method and system for automatic segmentation of cardiac medical images based on lossless processing

Abstract

The present application provides a method and system for automatic segmentation of cardiac medical images based on lossless processing. Among them, a dynamic sequence of cardiac motion images is received by an imaging device to generate dynamic cardiac imaging data; based on the dynamic cardiac imaging data, a time-frequency analysis algorithm is used to identify the unique vibration modes of each chamber and valve of the heart, and an adaptive harmonic reconstruction technique is used to enhance the clarity and accuracy of important physiological signals, and to generate a temporal and spatial consistent probability map of cardiac motion; based on the temporal and spatial consistent probability map of cardiac motion, a variational level set evolution algorithm is used to intelligently adjust energy changes, and a topological optimization method is used to perform anatomical coherence processing to generate a fine segmentation result of the cardiac structure; based on the fine segmentation result of the cardiac structure, a comprehensive evaluation is performed in combination with clinical diagnostic standards to generate an automatic segmentation report of cardiac medical images. The technical solution provided by the present application significantly improves the accuracy and efficiency of automatic segmentation of cardiac medical images.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of medical image analysis, and in particular to a method and system for automatic segmentation of cardiac medical images based on lossless processing. Background Art

[0002] With the increasing incidence of cardiovascular diseases, the automatic segmentation technology of cardiac medical images plays an increasingly important role in clinical diagnosis and treatment. In modern medicine, receiving dynamic sequence images of cardiac motion through imaging equipment and integrating information from different imaging sources to generate dynamic cardiac imaging data is a key step in evaluating cardiac function and structure. Doctors need accurate cardiac images to diagnose diseases such as cardiomyopathy and valvular disease and develop personalized treatment plans. To achieve this goal, not only high resolution and clarity are required for the images, but also the ability to accurately analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve, and provide a comprehensive and reliable spatial and temporal consistency probability map of cardiac motion.

[0003] Currently, there are many methods for cardiac medical image segmentation, including those based on traditional image processing technology and machine learning. These methods usually rely on static images or limited time series images, use edge detection, region growing and other algorithms for preliminary segmentation, and then combine expert knowledge for manual correction. In recent years, deep learning technology has also been introduced to automatically identify and segment cardiac structures by training convolutional neural network models.

[0004] However, existing methods for automatic segmentation of cardiac medical images have significant deficiencies. Traditional image processing technology has difficulty coping with complex cardiac motion and variable anatomical structures, resulting in insufficient accuracy in automatic segmentation of cardiac medical images. Although machine learning-based methods have been improved, they are still insufficient in dealing with subtle changes in the heart at different phases, especially in capturing the unique vibration patterns of the heart's chambers and valves. In addition, existing methods often ignore the spatiotemporal consistency of cardiac motion and are unable to generate high-quality probability maps, which in turn affects the reliability of the final segmentation results. In particular, the lack of accuracy in automatic segmentation of cardiac medical images still requires a lot of manual intervention in clinical applications, limiting its widespread application and efficiency improvement. Summary of the invention

[0005] The embodiments of the present application provide a method and system for automatic segmentation of cardiac medical images based on lossless processing, so as to solve the problem of insufficient accuracy of automatic segmentation of cardiac medical images in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for automatic segmentation of cardiac medical images based on lossless processing, comprising:

[0007] Receive dynamic sequence images of cardiac motion through imaging equipment, integrate information from different imaging sources, and generate dynamic cardiac imaging data;

[0008] Based on the dynamic cardiac imaging data, a time-frequency analysis algorithm is used to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, and use adaptive harmonic reconstruction technology to enhance the clarity and accuracy of important physiological signals, and generate a temporal and spatial consistent probability map of cardiac motion;

[0009] Based on the spatiotemporal consistent probability map of cardiac motion, the variational level set evolution algorithm is used to intelligently adjust the energy change at the boundary of cardiac tissue to achieve accurate tracking of the myocardial boundary, and the topology optimization method is used to perform anatomical coherence processing to generate a fine segmentation result of the cardiac structure;

[0010] Based on the result of the detailed segmentation of the cardiac structure, morphological measurements are performed, and a comprehensive evaluation is performed in combination with clinical diagnostic standards to generate an automatic segmentation report of cardiac medical images.

[0011] Optionally, based on the dynamic cardiac imaging data, a time-frequency analysis algorithm is used to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, and use adaptive harmonic reconstruction technology to enhance the clarity and accuracy of important physiological signals, and generate a temporal and spatial consistency probability map of cardiac motion, including:

[0012] Based on the dynamic cardiac image data, extracting the spatiotemporal information within the complete cardiac cycle to generate an image sequence of the overall motion characteristics of the heart;

[0013] Based on the image sequence of the overall motion characteristics of the heart, a time-frequency analysis algorithm is used to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, capture the unique frequency components of the heart motion, and generate a structural characteristic spectrum diagram;

[0014] Based on the structural characteristic spectrum diagram, adaptive harmonic reconstruction technology is used to intelligently adjust harmonic parameters, strengthen the detailed performance of the heart structure in different cardiac cycles, enhance the clarity and accuracy of important physiological signals, and generate dynamic and clear images of the heart;

[0015] Based on the dynamic clear images of the heart, the spatial distribution and time-varying laws of the heart structure are comprehensively considered and mapped to a continuous time series to generate a temporal and spatially consistent probability map of the heart movement.

[0016] Optionally, the image sequence based on the overall motion characteristic of the heart uses a time-frequency analysis algorithm to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, capture the unique frequency components of the heart motion, and generate a structural characteristic spectrum diagram, including:

[0017] Based on the image sequence of the overall motion characteristics of the heart, the dynamic changes of each chamber and valve of the heart are extracted to generate temporal and spatial characteristic data of the heart;

[0018] Based on the temporal and spatial characteristic data of the heart, the frequency response of the heart at different phases is analyzed by using a time-frequency analysis algorithm, the unique vibration modes of each chamber and valve of the heart are identified and quantified, the unique frequency components in the heart movement are captured, and the heart vibration spectrum is generated;

[0019] Based on the cardiac vibration spectrum, through frequency domain filtering analysis, we focus on the key vibration characteristics of the cardiac structure and generate effective vibration patterns:

[0020] Based on the effective vibration mode, reference information is displayed through visualization technology to generate a structural feature spectrum diagram.

[0021] Optionally, based on the structural characteristic spectrum diagram, adaptive harmonic reconstruction technology is used to intelligently adjust harmonic parameters, enhance the detail performance of the heart structure in different cardiac cycles, enhance the clarity and accuracy of important physiological signals, and generate dynamic and clear images of the heart, including:

[0022] Based on the structural characteristic spectrum diagram, key frequency components in different cardiac cycles of each chamber and valve of the heart are marked to generate a list of key frequency components;

[0023] Based on the above key frequency component list, adaptive harmonic reconstruction technology is applied to intelligently adjust harmonic parameters, enhance the details of the heart structure in different cardiac cycles, and generate enhanced heart vibration signals;

[0024] Based on the enhanced cardiac vibration signal, further multi-scale detail enhancement processing is performed to enhance the static structure visibility and dynamic change expression, and generate a dynamic enhanced cardiac image;

[0025] Based on the dynamic enhanced cardiac image, the clarity and accuracy of important physiological signals are enhanced to generate a dynamic clear cardiac image.

[0026] Optionally, based on the spatiotemporal consistent probability map of cardiac motion, a variational level set evolution algorithm is used to intelligently adjust the energy change at the boundary of cardiac tissue to achieve accurate tracking of the myocardial boundary, and a topology optimization method is used to perform anatomical coherence processing to generate a fine segmentation result of the cardiac structure, including:

[0027] Based on the cardiac motion spatiotemporal consistent probability map, identifying energy changes at the cardiac tissue boundary and generating a cardiac tissue boundary energy change map;

[0028] Based on the energy change map of the cardiac tissue boundary, the variational level set evolution algorithm is used to intelligently adjust the energy change at the cardiac tissue boundary by optimizing the energy function, so as to achieve accurate tracking of the myocardial boundary, ensure the accuracy of the boundary definition, and generate a preliminary cardiac structure boundary contour;

[0029] Based on the preliminary cardiac structure boundary contour, a topology optimization method is used to perform anatomical coherence processing, eliminate potential discontinuities and unreasonable boundaries, and generate an optimized cardiac structure boundary contour;

[0030] Based on the optimized cardiac structure boundary contour, the cardiac structure boundary smoothness and continuity are further optimized to generate a cardiac structure fine segmentation result.

[0031] Optionally, based on the energy change map of the cardiac tissue boundary, a variational level set evolution algorithm is used to optimize the energy function to intelligently adjust the energy change at the cardiac tissue boundary, achieve accurate tracking of the myocardial boundary, ensure the accuracy of boundary definition, and generate a preliminary cardiac structure boundary contour, including:

[0032] Based on the cardiac tissue boundary energy change map, the physical characteristics of the cardiac tissue boundary and the image intensity information are quantified to generate an energy function:

[0033] Based on the energy function, a variational level set evolution algorithm is used for optimization processing to intelligently adjust the energy change at the boundary of cardiac tissue, achieve accurate tracking of the myocardial boundary, ensure the accuracy of boundary definition, and generate an optimized boundary position;

[0034] Based on the optimized boundary position, the geometric accuracy of boundary definition is improved, key features of the cardiac structure are retained, and a high-precision boundary position is generated;

[0035] Based on the high-precision boundary position, multi-scale spatial correction processing is performed to generate a preliminary cardiac structure boundary contour.

[0036] Optionally, the method of performing morphological measurement based on the result of the fine segmentation of the cardiac structure and conducting a comprehensive evaluation in combination with clinical diagnostic standards to generate an automatic segmentation report of cardiac medical images includes:

[0037] Based on the cardiac structure fine segmentation result, key dimensions and position information of the cardiac structure are quantified to generate a detailed cardiac anatomical parameter list;

[0038] Based on the detailed cardiac anatomical parameter list, a series of morphological measurements are performed to accurately calculate cardiac structural volumetric indices and functional parameters, and generate cardiac morphological measurement data;

[0039] Based on the cardiac morphology measurement data and combined with clinical diagnostic criteria, a comprehensive assessment of cardiac function status is performed to generate a cardiac function assessment report;

[0040] Based on the cardiac function assessment report, a personalized comprehensive clinical interpretation is performed to generate an automatic segmentation report of cardiac medical images.

[0041] In a second aspect, an embodiment of the present application provides a cardiac medical image automatic segmentation system based on lossless processing, comprising:

[0042] A receiving module, used for receiving a dynamic sequence of cardiac motion images through an imaging device, integrating information from different imaging sources, and generating dynamic cardiac imaging data;

[0043] An analysis module is used to analyze the frequency response of the heart at different phases based on the dynamic cardiac imaging data by using a time-frequency analysis algorithm, identify the unique vibration modes of each chamber and valve of the heart, and use adaptive harmonic reconstruction technology to enhance the clarity and accuracy of important physiological signals and generate a temporal and spatial consistency probability map of cardiac motion;

[0044] An adjustment module is used to intelligently adjust the energy change at the boundary of cardiac tissue based on the spatiotemporal consistent probability map of cardiac motion by using a variational level set evolution algorithm to achieve accurate tracking of the myocardial boundary, and to perform anatomical coherence processing by using a topology optimization method to generate a fine segmentation result of the cardiac structure;

[0045] The measurement module is used to perform morphological measurement based on the fine segmentation results of the cardiac structure, conduct a comprehensive evaluation in combination with clinical diagnostic standards, and generate an automatic segmentation report for cardiac medical images.

[0046] In a third aspect, an embodiment of the present application provides a computing device, comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an automatic segmentation method for cardiac medical images based on lossless processing as described in the first aspect.

[0047] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements an automatic segmentation method for cardiac medical images based on lossless processing as described in the first aspect.

[0048] In the embodiment of the present application, a dynamic sequence of cardiac motion images is received by an imaging device, and information from different imaging sources is integrated to generate dynamic cardiac imaging data; based on the dynamic cardiac imaging data, a time-frequency analysis algorithm is used to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, and an adaptive harmonic reconstruction technique is used to enhance the clarity and accuracy of important physiological signals, and generate a temporal and spatial consistent probability map of cardiac motion; based on the temporal and spatial consistent probability map of cardiac motion, a variational level set evolution algorithm is used to intelligently adjust the energy changes at the boundary of cardiac tissue, and accurate tracking of the myocardial boundary is achieved, and a topological optimization method is used to perform anatomical coherence processing to generate a fine segmentation result of the cardiac structure; based on the fine segmentation result of the cardiac structure, morphological measurements are performed, and a comprehensive evaluation is performed in combination with clinical diagnostic standards to generate an automatic segmentation report of cardiac medical images. By providing a comprehensive and automated solution, from image acquisition to the generation of the final segmentation report, all aspects of cardiac image processing are covered. This method not only improves the accuracy and efficiency of cardiac image analysis, but also ensures the reliability and consistency of the segmentation results, greatly facilitating doctors' diagnosis and treatment decisions.

[0049] Furthermore, by describing in detail the specific application steps of time-frequency analysis and adaptive harmonic reconstruction technology, the accurate identification of the vibration modes of each heart chamber and valve and the clarity of important physiological signals are ensured, the quality of cardiac imaging data is enhanced, and high-quality basic data is provided for the subsequent probability map generation, thereby improving the accuracy of the entire segmentation process.

[0050] Furthermore, by introducing the variational level set evolution algorithm and topology optimization method, the accurate tracking of cardiac tissue boundaries and the anatomical continuity processing are ensured. This method not only improves the accuracy of cardiac structure segmentation, but also eliminates potential discontinuities and unreasonable boundaries, making the final segmentation results more consistent with the actual anatomical structure, providing a reliable basis for subsequent clinical evaluation.

[0051] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 A flowchart of a method for automatic segmentation of cardiac medical images based on lossless processing provided in an embodiment of the present application;

[0054] Figure 2 A schematic diagram of the structure of a cardiac medical image automatic segmentation system based on lossless processing provided in an embodiment of the present application;

[0055] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0057] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0059] Figure 1 A flowchart of a method for automatic segmentation of cardiac medical images based on lossless processing is provided for an embodiment of the present application. Figure 1 As shown, the method includes:

[0060] 101. Receive dynamic sequence images of cardiac motion through an imaging device, integrate information from different imaging sources, and generate dynamic cardiac imaging data;

[0061] In this step, the imaging device refers to a medical instrument used to capture dynamic images of the heart, including but not limited to echocardiography, magnetic resonance imaging, computed tomography, etc. These devices can provide high-resolution cardiac structure and function information. The dynamic sequence image of cardiac motion is a data set composed of cardiac images at multiple consecutive time points, which reflects the dynamic changes of the heart in different cardiac cycles and provides rich spatiotemporal information for subsequent analysis. The image data from different imaging devices are aligned through image registration technology to ensure spatial consistency and eliminate errors caused by equipment differences. The dynamic cardiac imaging data is a processed and integrated multi-source imaging data set that contains detailed structural and motion information of the heart at different phases for further time-frequency analysis and segmentation.

[0062] In the embodiments of the present application, multiple imaging devices are first used to collect multiple frames of images of the heart in different cardiac cycles; secondly, the images from different imaging sources are aligned through image registration technology to ensure spatial consistency; thirdly, a noise filtering algorithm is applied to remove artifacts and improve image quality; finally, the processed images are integrated into a complete dynamic cardiac image data set for subsequent analysis.

[0063] Suppose a hospital uses echocardiography equipment to collect a series of dynamic cardiac images. First, it combines magnetic resonance imaging (MRI) images to obtain more detailed myocardial structural information. Second, it uses image registration technology to align the two image sources to eliminate displacement errors. Third, it uses a noise filtering algorithm to process the images and remove artifacts. Finally, it generates a high-quality dynamic cardiac image dataset, laying the foundation for subsequent steps.

[0064] 102. Based on the dynamic cardiac imaging data, a time-frequency analysis algorithm is used to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, and use adaptive harmonic reconstruction technology to enhance the clarity and accuracy of important physiological signals, and generate a temporal and spatial consistent probability map of cardiac motion;

[0065] In this step, the dynamic cardiac imaging data is a high-quality cardiac imaging data set that integrates multi-source imaging information and is used to analyze the characteristics of cardiac motion. The time-frequency analysis algorithm is a signal processing method that can simultaneously analyze signal characteristics in the time and frequency domains, and is particularly suitable for the analysis of non-stationary signals such as cardiac vibration patterns. The frequency response refers to the changes in the frequency components of the heart in different cardiac cycles, reflecting the unique vibration modes of the heart's chambers and valves. The adaptive harmonic reconstruction technology intelligently adjusts the harmonic parameters to enhance the details of the heart structure in different cardiac cycles and enhance the clarity and accuracy of important physiological signals. The temporal and spatial consistency probability map of cardiac motion is a probability map generated by comprehensively considering the spatial distribution and time-varying laws of the cardiac structure and mapping it to a continuous time series, which is used to describe the temporal and spatial consistency of cardiac motion.

[0066] In the embodiments of the present application, first, the spatiotemporal information within the complete cardiac cycle is extracted from the dynamic cardiac imaging data; second, the time-frequency analysis algorithm is applied to analyze the frequency response of the heart in different phases and identify unique vibration patterns; third, the adaptive harmonic reconstruction technology is used to enhance the clarity and accuracy of important physiological signals; finally, a spatiotemporal consistent probability map reflecting the motion characteristics of the heart is generated based on the analysis results.

[0067] For example, continuing with the above example, suppose that the spatiotemporal information within the complete cardiac cycle is extracted from the integrated dynamic cardiac imaging data; first, the short-time Fourier transform and wavelet transform are used to analyze the frequency response of the heart in different phases; second, the harmonic parameters are optimized through adaptive harmonic reconstruction technology to enhance the details of the myocardium and valves; third, a structural feature spectrum is generated; finally, it is mapped to a continuous time series to generate a spatiotemporal consistent probability map of cardiac motion.

[0068] 103. Based on the spatiotemporal consistent probability map of cardiac motion, a variational level set evolution algorithm is used to intelligently adjust the energy change at the boundary of cardiac tissue to achieve accurate tracking of the myocardial boundary, and a topology optimization method is used to perform anatomical coherence processing to generate a fine segmentation result of the cardiac structure;

[0069] In this step, the temporal and spatial consistent probability map of cardiac motion is a probability map describing the characteristics of cardiac motion, providing an accurate representation of cardiac structure in the temporal and spatial dimensions. The variational level set evolution algorithm is an image segmentation algorithm based on energy minimization. It realizes intelligent adjustment of boundaries by optimizing the energy function to ensure the accuracy of boundary definition. The energy change at the boundary of cardiac tissue refers to the change in energy distribution at the boundary of cardiac tissue in different cardiac cycles, reflecting the dynamic characteristics of cardiac structure.

[0070] Intelligent adjustment optimizes the energy function to make the definition of the heart boundary more accurate and ensure that it conforms to the actual anatomical structure. The topological optimization method is a mathematical method used to ensure the consistency and rationality of the segmentation results and eliminate potential discontinuities or unreasonable boundaries. The fine segmentation result of the heart structure is a fine segmentation image of the heart structure generated after anatomical consistency processing, which provides a basis for morphological measurement.

[0071] In an embodiment of the present application, first, the energy changes at the boundaries of cardiac tissue are identified based on the spatiotemporal consistent probability map of cardiac motion; second, the variational level set evolution algorithm is used to optimize the energy function and intelligently adjust the boundary position; third, the topology optimization method is combined to eliminate potential discontinuities or unreasonable boundaries; finally, a fine segmentation result of the cardiac structure processed with anatomical consistency is generated.

[0072] For example, continuing with the above example, assume that an energy change map of the cardiac tissue boundary is generated based on the spatiotemporal consistent probability map of cardiac motion; first, the variational level set evolution algorithm is applied to optimize the energy function and intelligently adjust the position of the cardiac tissue boundary; second, the topological optimization method is combined to ensure the anatomical consistency and rationality of the segmentation results; third, the smoothness and continuity of the cardiac structure boundary are optimized; finally, a fine segmentation result of the cardiac structure is generated to provide a basis for morphological measurement.

[0073] 104. Based on the result of the fine segmentation of the cardiac structure, morphological measurements are performed, and a comprehensive evaluation is performed in combination with clinical diagnostic standards to generate an automatic segmentation report of cardiac medical images.

[0074] In this step, the result of the fine segmentation of the cardiac structure is a fine segmentation image of the cardiac structure generated after anatomical continuity processing, which provides the precise boundaries of each part of the heart. Morphological measurements include calculating the volume, wall thickness, valve opening area and other physical quantities of each chamber of the heart, which are used to evaluate the structure and functional status of the heart. The clinical diagnostic standards are a set of widely accepted medical standards used to assess heart health and disease status and guide doctors in formulating treatment plans. The automatic segmentation report of cardiac medical images is the final detailed report generated, which contains the quantitative analysis results of the structure and function of the heart, providing doctors with an intuitive and quantitative reference.

[0075] In the embodiments of the present application, first, a series of morphological measurements are performed based on the results of fine segmentation of the cardiac structure; second, key physical quantities of the cardiac structure, such as volumetric indicators and functional parameters, are calculated; third, a comprehensive assessment of the cardiac functional status is performed in combination with clinical diagnostic standards; finally, an automatic segmentation report of cardiac medical images is generated, providing detailed quantitative analysis results.

[0076] For example, continuing with the above example, assume that morphological measurements are performed based on the results of fine segmentation of the cardiac structure; first, the volume and wall thickness of the ventricles and atria are calculated; second, the functional status of the heart is comprehensively evaluated by combining functional parameters such as ejection fraction and stroke volume; third, clinical diagnostic standards are compared to identify abnormal indicators and analyze cardiac motor function; finally, an automatic segmentation report of cardiac medical images is generated to provide doctors with a detailed reference for formulating treatment plans.

[0077] In summary, steps 101 to 104 cover the entire process from cardiac image acquisition to final segmentation report generation, aiming to provide a systematic and automated cardiac medical image processing solution to meet the needs of cardiovascular disease diagnosis and treatment. This process not only improves the accuracy and efficiency of cardiac image analysis, but also ensures the reliability and clinical applicability of the segmentation results, greatly promoting the precise diagnosis and treatment of cardiovascular diseases.

[0078] In order to solve the accuracy problem of frequency response and vibration mode recognition in cardiac image analysis, in some embodiments, the time-frequency analysis based on dynamic cardiac image data and the generation of a temporal-spatial consistency probability map of cardiac motion described in step 102 include: based on the dynamic cardiac image data, extracting the temporal-spatial information within the complete cardiac cycle to generate a cardiac overall motion characteristic image sequence; based on the cardiac overall motion characteristic image sequence, using a time-frequency analysis algorithm to analyze the frequency response of the heart in different phases, identify the unique vibration modes of each chamber and valve of the heart, capture the unique frequency components of cardiac motion, and generate a structural characteristic spectrum map; based on the structural characteristic spectrum map, using adaptive harmonic reconstruction technology to intelligently adjust harmonic parameters, strengthen the detailed performance of the cardiac structure in different cardiac cycles, enhance the clarity and accuracy of important physiological signals, and generate a dynamic and clear cardiac image; based on the dynamic and clear cardiac image, comprehensively consider the spatial distribution and time-varying laws of the cardiac structure, map to a continuous time series, and generate a temporal-spatial consistency probability map of cardiac motion.

[0079] In this embodiment, the spatiotemporal information within a complete cardiac cycle refers to the spatial position and morphological change information of the heart at each time point during a complete heartbeat. This information ensures a comprehensive description of the heart's motion characteristics by integrating spatial and temporal features in multiple frames of images.

[0080] The image sequence of the overall motion characteristics of the heart is generated by extracting and integrating the spatiotemporal information within the complete cardiac cycle. This sequence not only contains the changes in the heart structure, but also reflects the movement pattern of the heart throughout the cardiac cycle, providing an important basis for subsequent analysis.

[0081] Frequency response refers to the changes in the frequency components of the heart during different cardiac cycles, reflecting the unique vibration modes of each chamber and valve of the heart.

[0082] The structural characteristic spectrum is a chart generated by analyzing the frequency response of the heart at different phases, which is used to intuitively display the unique vibration modes of each chamber and valve of the heart. This spectrum can capture the unique frequency components in the movement of the heart and provide a basis for subsequent reconstruction.

[0083] The dynamic clear cardiac images are processed by adaptive harmonic reconstruction technology, which significantly enhances the details of the cardiac structure in different cardiac cycles, ensures the clarity and accuracy of important physiological signals, and provides high-quality basic data for subsequent analysis.

[0084] In the embodiments of the present application, firstly, the spatiotemporal information within a complete cardiac cycle is extracted from dynamic cardiac imaging data to generate an image sequence of the overall motion characteristics of the heart; secondly, a time-frequency analysis algorithm is used to parse the frequency response of the heart in different phases, identify the unique vibration modes of each chamber and valve of the heart, and generate a structural characteristic spectrum diagram; thirdly, based on the structural characteristic spectrum diagram, an adaptive harmonic reconstruction technique is used to intelligently adjust the harmonic parameters, enhance the detailed performance of the heart structure in different cardiac cycles, and generate a dynamic and clear image of the heart; finally, the spatial distribution and time-varying laws of the heart structure are comprehensively considered, and the dynamic and clear image of the heart is mapped to a continuous time series to generate a temporal and spatial consistent probability map of the heart motion.

[0085] Suppose a medical institution needs to process the MRI and echocardiography dynamic cardiac imaging data of a patient with dilated cardiomyopathy; first, multimodal image registration technology is used to integrate MRI and echocardiography images, extract the spatiotemporal information within the complete cardiac cycle, and generate an image sequence of the overall motion characteristics of the heart; second, time-frequency analysis methods such as short-time Fourier transform (STFT) and wavelet transform are used to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, and generate a structural feature spectrum; third, based on the generated structural feature spectrum, the harmonic parameters are optimized through adaptive harmonic reconstruction technology, the detailed performance of the myocardium and valves in different cardiac cycles is enhanced, and a dynamic and clear image of the heart is generated; finally, the spatial distribution and time-varying law of the heart structure are combined to map the dynamic and clear image of the heart to a continuous time series, and a temporal and spatial consistent probability map of the heart motion is generated, providing high-precision basic data for subsequent segmentation.

[0086] In order to solve the accuracy problem of automatic segmentation of cardiac medical images, in some embodiments, the time-frequency analysis processing based on the image sequence of the overall motion characteristics of the heart in step 102 includes: based on the image sequence of the overall motion characteristics of the heart, extracting the dynamic changes of the chambers and valves of the heart, and generating temporal and spatial characteristic data of the heart; based on the temporal and spatial characteristic data of the heart, using a time-frequency analysis algorithm, analyzing the frequency response of the heart in different phases, identifying and quantifying the unique vibration modes of the chambers and valves of the heart, capturing the unique frequency components in the heart movement, and generating a heart vibration spectrum; based on the heart vibration spectrum, focusing on the key vibration characteristics of the heart structure through frequency domain filtering analysis, and generating an effective vibration mode; based on the effective vibration mode, displaying reference information through visualization technology, and generating a structural feature spectrum diagram.

[0087] In this embodiment, the cardiac spatiotemporal feature data refers to a data set of spatial positions and morphologies of different parts of the heart (such as the atria, ventricles, and valves) that change over time during the cardiac cycle, extracted from an image sequence of the heart's overall motion characteristics, and is used to quantify the dynamic changes of the heart structure at different time points.

[0088] Time-frequency analysis is a signal processing method that can simultaneously show how the frequency components of a signal change over time. It is particularly important for capturing the vibration characteristics of the heart at different phases.

[0089] The cardiac vibration spectrum refers to a graph or data set obtained through time-frequency analysis, which represents the specific vibration mode of the cardiac structure in different cardiac cycles and its corresponding frequency distribution.

[0090] Effective vibration modes refer to those vibration characteristics that best represent the structural and functional characteristics of the heart that are retained after frequency domain filtering analysis.

[0091] The structural feature spectrum diagram is the final output visualization result, which presents the key vibration information of each part of the heart in an intuitive way, helping doctors better understand the functional status of the heart.

[0092] In the embodiments of the present application, firstly, the original cardiac images are preprocessed by multimodal image registration technology and contrast enhancement algorithm to ensure that all parts of the heart are clearly visible; secondly, advanced time-frequency analysis methods are applied to evaluate the activity of the heart in different cardiac stages and determine the vibration mode of the internal structure of the heart and its frequency composition; thirdly, frequency domain filters are used to separate the main vibration characteristics of the cardiac structure, thereby improving the signal-to-noise ratio and highlighting important information; finally, professional software tools are used to realize the automatic generation of characteristic spectrum diagrams of the cardiac structure, which are provided to clinical physicians as diagnostic aids.

[0093] Here is a specific example:

[0094] In order to further improve the automatic segmentation effect of lossless cardiac medical images, suppose a hospital wants to improve its existing cardiac image analysis process. First, the hospital collected a large number of patients' cardiac MRI and echocardiography data, and used multimodal fusion technology to create a high-resolution cardiac motion characteristic image sequence; second, the researchers developed a set of time-frequency analysis algorithms specifically for these images to accurately measure the motion trajectory of each chamber and valve of the heart in each cardiac cycle, and record the corresponding frequency response; third, they designed a set of customized frequency domain filters based on the obtained cardiac vibration spectrum, successfully filtering out unnecessary noise and highlighting the important vibration characteristics of the heart structure; finally, the team members converted the optimized vibration mode into an easy-to-interpret structural feature spectrum, providing more accurate decision support materials for cardiology experts.

[0095] In order to solve the problem of clarity and accuracy of important physiological signals in cardiac medical images, in some embodiments, the adaptive harmonic reconstruction technology based on the structural feature spectrum diagram described in step 102 includes: based on the structural feature spectrum diagram, marking the key frequency components in different cardiac cycles of each chamber and valve of the heart, and generating a list of key frequency components; based on the above-mentioned list of key frequency components, applying adaptive harmonic reconstruction technology, intelligently adjusting harmonic parameters, enhancing the detail performance of the heart structure in different cardiac cycles, and generating an enhanced heart vibration signal; based on the enhanced heart vibration signal, performing further multi-scale detail enhancement processing, enhancing the static structure visibility and dynamic change expression, and generating a dynamic enhanced image of the heart; based on the dynamic enhanced image of the heart, enhancing the clarity and accuracy of important physiological signals, and generating a dynamic clear image of the heart.

[0096] In this embodiment, the key frequency component list is a data set compiled based on the key frequency information of various parts of the heart marked in the structural characteristic spectrum diagram in different cardiac cycles, and is used to guide the subsequent reconstruction process.

[0097] Enhancing the cardiac vibration signal means that after the adaptive harmonic reconstruction processing, the dynamic change signal of the cardiac structure in different phases is significantly enhanced.

[0098] Multi-scale detail enhancement processing refers to further improving the visibility of static structures and the expression of dynamic changes by targeted enhancement of details at different scales.

[0099] Dynamic clear cardiac images refer to high-quality images generated after the above processing, which ensure the clarity and accuracy of important physiological signals and provide a reliable basis for subsequent analysis.

[0100] In an embodiment of the present application, firstly, based on the structural characteristic spectrum diagram, the key frequency components of each chamber and valve of the heart in different cardiac cycles are marked, and a list of key frequency components is generated; secondly, the adaptive harmonic reconstruction technology is applied to intelligently adjust the harmonic parameters according to the list of key frequency components, so as to enhance the detail performance of the heart structure in different cardiac cycles and generate an enhanced heart vibration signal; thirdly, based on the enhanced heart vibration signal, multi-scale detail enhancement processing is performed to further improve the visibility of the static structure and the expression effect of dynamic changes, and generate a dynamic enhanced image of the heart; finally, all processing results are combined to enhance the clarity and accuracy of important physiological signals and generate a final dynamic clear image of the heart.

[0101] Here is a specific example:

[0102] Suppose a cardiovascular research center needs to process the echocardiographic image data of a patient with arrhythmia; first, based on the structural characteristic spectrum, the key frequency components of each chamber and valve of the heart in different cardiac cycles are accurately marked to generate a list of key frequency components; second, the adaptive harmonic reconstruction technology is applied to intelligently adjust the harmonic parameters according to these key frequency components, enhance the details of the heart structure in different cardiac cycles, and generate an enhanced heart vibration signal; third, based on the enhanced heart vibration signal, multi-scale detail enhancement processing is performed to further improve the visibility of the static structure and the expression of dynamic changes, and generate a dynamic enhanced image of the heart; finally, all processing results are combined to enhance the clarity and accuracy of important physiological signals, and generate a final dynamic and clear heart image, providing doctors with a more detailed basis for diagnosis.

[0103] In order to solve the problem of accuracy in defining the boundaries of cardiac tissues, in some embodiments, the variational level set evolution algorithm processing based on the spatiotemporal consistent probability map of cardiac motion in step 103 includes: based on the spatiotemporal consistent probability map of cardiac motion, identifying energy changes at the boundaries of cardiac tissues, and generating an energy change map of the boundaries of cardiac tissues; based on the energy change map of the boundaries of cardiac tissues, using the variational level set evolution algorithm, by optimizing the energy function, intelligently adjusting the energy changes at the boundaries of cardiac tissues, achieving precise tracking of the boundaries of myocardium, ensuring the accuracy of the boundary definition, and generating a preliminary cardiac structure boundary contour; based on the preliminary cardiac structure boundary contour, using a topological optimization method to perform anatomical continuity processing, eliminating potential discontinuities and unreasonable boundaries, and generating an optimized cardiac structure boundary contour; based on the optimized cardiac structure boundary contour, further optimizing the smoothness and continuity of the cardiac structure boundary, and generating a fine segmentation result of the cardiac structure.

[0104] In this embodiment, the heart tissue boundary energy change map is an image generated by analyzing the spatiotemporal consistent probability map of cardiac motion and identifying the energy change at the heart tissue boundary, which reflects the dynamic characteristics of the heart boundary.

[0105] Boundary definition accuracy refers to the extent to which cardiac tissue boundaries are correctly identified and delineated during image segmentation.

[0106] The preliminary heart structure boundary contour refers to the initial heart boundary shape generated after processing by the variational level set evolution algorithm, which is used for subsequent refinement and optimization.

[0107] Optimizing the cardiac structure boundary contour means further improving the smoothness and continuity of the cardiac structure boundary through a series of processing steps to ensure that it conforms to the geometric characteristics of the actual anatomical structure.

[0108] The cardiac structure fine segmentation result refers to the cardiac structure fine segmentation image generated after anatomical coherence processing, which provides a basis for morphological measurement.

[0109] In an embodiment of the present application, firstly, based on the spatiotemporal consistent probability map of cardiac motion, the energy changes at the boundary of cardiac tissue are identified, and an energy change map of the boundary of cardiac tissue is generated; secondly, the variational level set evolution algorithm is used to intelligently adjust the energy changes at the boundary of cardiac tissue by optimizing the energy function, so as to achieve precise tracking of the myocardial boundary, ensure the accuracy of the boundary definition, and generate a preliminary cardiac structure boundary contour; thirdly, based on the preliminary cardiac structure boundary contour, a topological optimization method is used to perform anatomical continuity processing, eliminate potential discontinuities and unreasonable boundaries, and generate an optimized cardiac structure boundary contour; finally, the smoothness and continuity of the cardiac structure boundary are further optimized to generate the final cardiac structure fine segmentation result.

[0110] Here is a specific example:

[0111] Assume that a large medical research center is analyzing the CT image data of a patient with aortic stenosis; first, based on the consistent spatiotemporal probability map of cardiac motion, the energy changes at the boundary of cardiac tissue are identified, and a map of the energy changes at the boundary of cardiac tissue is generated; secondly, the variational level set evolution algorithm is applied to intelligently adjust the energy changes at the boundary of cardiac tissue by optimizing the energy function, so as to achieve precise tracking of the myocardial boundary, ensure the accuracy of the boundary definition, and generate a preliminary cardiac structure boundary contour; thirdly, based on the preliminary cardiac structure boundary contour, the topological optimization method is used to perform anatomical coherence processing, eliminate potential discontinuities and unreasonable boundaries, and generate an optimized cardiac structure boundary contour; finally, the smoothness and continuity of the cardiac structure boundary are further optimized to generate the final cardiac structure fine segmentation result, providing doctors with detailed anatomical information to assist in diagnosis and treatment planning.

[0112] In order to solve the problem of accuracy in defining the cardiac tissue boundary, in some embodiments, the variational level set evolution algorithm processing based on the cardiac tissue boundary energy change map in step 103 includes: based on the cardiac tissue boundary energy change map, quantifying the physical characteristics of the cardiac tissue boundary and the image intensity information, and generating an energy function; based on the energy function, using the variational level set evolution algorithm for optimization processing to intelligently adjust the energy changes at the cardiac tissue boundary, achieve accurate tracking of the myocardial boundary, ensure the accuracy of the boundary definition, and generate an optimized boundary position; based on the optimized boundary position, improve the geometric accuracy of the boundary definition, retain the key features of the cardiac structure, and generate a high-precision boundary position; based on the high-precision boundary position, perform multi-scale spatial correction processing to generate a preliminary cardiac structure boundary contour.

[0113] In an embodiment of the present application, first, based on the energy change map of the cardiac tissue boundary, the physical characteristics of the cardiac tissue boundary and the image intensity information are quantified to generate an energy function; secondly, based on the generated energy function, the variational level set evolution algorithm is used to perform optimization processing, and the energy change at the cardiac tissue boundary is intelligently adjusted to achieve precise tracking of the myocardial boundary, ensure the accuracy of the boundary definition, and generate an optimized boundary position; thirdly, based on the optimized boundary position, the geometric accuracy of the boundary definition is improved, the key features of the cardiac structure are retained, and a high-precision boundary position is generated; finally, the high-precision boundary position is subjected to multi-scale spatial correction processing to generate a preliminary cardiac structure boundary contour.

[0114] Suppose a cardiovascular imaging analysis laboratory needs to process echocardiographic image data of a patient with hypertrophic cardiomyopathy; first, based on the energy change map of the cardiac tissue boundary, the physical characteristics of the cardiac tissue boundary and the image intensity information are quantified to generate an energy function; second, based on the generated energy function, the variational level set evolution algorithm is applied for optimization processing to intelligently adjust the energy change at the cardiac tissue boundary, achieve precise tracking of the myocardial boundary, ensure the accuracy of the boundary definition, and generate an optimized boundary position; third, based on the optimized boundary position, the geometric accuracy of the boundary definition is improved, the key features of the cardiac structure are retained, and a high-precision boundary position is generated; finally, the high-precision boundary position is subjected to multi-scale spatial correction processing to generate a preliminary cardiac structure boundary contour, providing doctors with detailed anatomical information to assist in diagnosis and treatment planning.

[0115] In order to solve the problem of comprehensiveness and accuracy of the automatic segmentation report of cardiac medical images, in some embodiments, the morphological measurement based on the fine segmentation result of the cardiac structure and the generation of the report described in step 104 include: based on the fine segmentation result of the cardiac structure, quantifying the key dimensions and position information of the cardiac structure, and generating a detailed cardiac anatomical parameter list; based on the detailed cardiac anatomical parameter list, performing a series of morphological measurements, accurately calculating the volume index and functional parameters of the cardiac structure, and generating cardiac morphological measurement data; based on the cardiac morphological measurement data, combined with clinical diagnostic standards, a comprehensive assessment of the cardiac function status is performed to generate a cardiac function assessment report; based on the cardiac function assessment report, a personalized comprehensive clinical interpretation is performed to generate an automatic segmentation report of cardiac medical images.

[0116] In this embodiment, the fine segmentation result of the cardiac structure refers to the precise boundary and position information of different parts of the heart (such as ventricles, atria, valves, blood vessels, etc.) obtained after automatic or semi-automatic analysis of medical imaging data such as MRI or CT by advanced image processing algorithms (such as deep learning, machine learning, etc.). These segmentation results are used for subsequent quantification of key dimensions and position information of the cardiac structure, that is, determining the size, shape and spatial positioning of each component of the heart.

[0117] The detailed cardiac anatomical parameter list is a dataset generated by quantifying the key dimensions and positional information of cardiac structures for subsequent morphological measurements.

[0118] Morphological measurements are a series of geometric and functional quantitative analyses of cardiac structure, aimed at evaluating the structural and functional status of the heart. These measurements include but are not limited to calculating the volume indicators of each cardiac chamber (such as ventricular end-diastolic volume, end-systolic volume), wall thickness, valve opening area and other physical quantities, as well as functional parameters (such as ejection fraction, stroke volume) to comprehensively reflect the functional characteristics of the heart.

[0119] Volumetric indicators refer to the volume changes of each heart chamber in different cardiac cycles, including end-diastolic volume, end-systolic volume, etc.

[0120] Functional parameters refer to quantitative indicators used to evaluate the heart's pumping efficiency and functional status, such as ejection fraction, stroke volume, cardiac output, etc.

[0121] The cardiac function assessment report is a detailed report generated based on cardiac morphology measurement data, which comprehensively evaluates the cardiac function status in combination with clinical diagnostic criteria.

[0122] Personalized comprehensive clinical interpretation refers to a detailed interpretation by a professional doctor based on the patient's specific situation and combined with the cardiac function assessment report.

[0123] In the embodiment of the present application, a detailed list of cardiac anatomical parameters is first generated based on the results of the above-mentioned fine segmentation. This step involves extracting and calculating the geometric features of various cardiac structures, such as the inner diameter, wall thickness, and myocardial mass of the left and right ventricles. Then, using these anatomical parameter lists, the system will perform a series of morphological measurements to accurately calculate the volume indicators (such as end-diastolic volume, end-systolic volume) and functional parameters (such as ejection fraction and cardiac output) of the cardiac structure, thereby obtaining cardiac morphological measurement data. Then, combined with clinical diagnostic criteria, the system will conduct a comprehensive assessment of the functional state of the heart and generate a cardiac function assessment report, which contains professional insights into the health of the heart. Finally, based on this assessment report, medical experts can provide personalized comprehensive clinical interpretations, and ultimately form an automatic segmentation report for cardiac medical images, which not only summarizes the results of the technical analysis, but also provides treatment recommendations for specific patient conditions.

[0124] Here is a specific example:

[0125] Suppose there is a patient with dilated cardiomyopathy whose cardiac imaging shows an enlarged heart and reduced pumping efficiency. In the process of applying step 104, the patient's MRI scan image is first automatically segmented using computer-assisted software to identify the specific locations of the heart's chambers and major blood vessels. The software then automatically generates a detailed list of anatomical parameters that lists all important dimensions and location information. Next, the software calculates the volume changes of the ventricles and the proportion of blood ejected with each heartbeat (ejection fraction). If the ejection fraction is found to be significantly lower than normal, it indicates that the heart's ability to pump blood is impaired. Finally, the doctor will receive a comprehensive cardiac function assessment report that indicates the severity of the patient's condition and provides possible treatment options, such as medication, lifestyle adjustments, or surgical intervention. In addition, a personalized cardiac medical image automatic segmentation report will be attached to help the doctor better understand the progression of the disease and develop the most appropriate treatment plan.

[0126] This application considers that in order to solve the problem of inaccurate recognition of cardiac vibration patterns caused by noise interference and multi-source data synchronization errors in the prior art, the embodiment of the invention proposes this optional solution. This solution aims to accurately analyze the frequency response of the heart at different phases, capture unique frequency components, and generate a high-precision cardiac vibration spectrum by removing noise, correcting time synchronization errors, performing spatial normalization processing, and applying time-frequency analysis algorithms, so as to solve the accuracy problem in the automatic segmentation of cardiac medical images. Therefore, a new optional solution is proposed, which includes:

[0127] Based on the temporal and spatial characteristic data of the heart, the frequency response of the heart at different phases is analyzed using a time-frequency analysis algorithm, the unique vibration modes of each chamber and valve of the heart are identified and quantified, the unique frequency components in the heart movement are captured, and the heart vibration spectrum is generated, including:

[0128] Based on the cardiac spatiotemporal characteristic data, noise is removed to improve signal quality, and time synchronization errors are corrected to ensure consistency of data from different imaging sources;

[0129] The cardiac structures are aligned to a standard coordinate system through spatial normalization, and the local frequency content is analyzed by applying a short-time Fourier transform to generate an energy response:

[0130] The energy response is calculated using the following formula:

[0131] ;

[0132] in, For time Energy response at For the The amplitude of the vibration mode of each chamber or valve; For the Attenuation coefficient of each chamber or valve vibration mode; For the The time shift of the vibration mode of each chamber or valve; For the The frequency content of the vibration modes of each chamber or valve; For the The phase angle of each chamber or valve vibration mode; is the index of the unique vibration mode identified, from 1 to ; is the total number of unique vibration patterns identified; is the time variable;

[0133] Based on the energy response, filtering is performed using a nonlinear time-varying filter function to highlight the unique frequency components of the heart, and a window function is introduced to limit the influence within the time range to ensure that the frequency components in each cardiac cycle are accurately captured to generate the heart vibration spectrum intensity;

[0134] The intensity of the heart vibration spectrum is calculated using the following formula:

[0135] ;

[0136] in, For frequency The intensity of the heart vibration spectrum at For time Energy response at It is a nonlinear time-varying filter function used to capture the unique frequency components in cardiac motion; It is a window function, which is used to limit the impact within the time range; is an imaginary unit; is a frequency variable; is the time variable; is the adjustment factor used to balance the contribution of the two integral terms; is another nonlinear time-varying filter function used to capture more complex frequency components; is a weight function used to adjust the importance of different time periods;

[0137] Based on the intensity of the heart vibration spectrum, a smoothing filter is applied to reduce artifacts to ensure continuity and smoothness, and key frequency components are identified and marked through peak detection technology to generate a heart vibration spectrum.

[0138] This method aims to effectively analyze the frequency response of the heart in different cardiac cycles and capture its unique frequency components. To this end, this solution adopts a series of steps, from denoising to energy response calculation, to spectrum intensity generation, and finally smoothing filtering and peak detection to ensure that the final generated heart vibration spectrum can truly reflect the unique vibration modes of each chamber and valve of the heart. The overall design is based on improving the reliability and clinical applicability of the heart vibration spectrum through mathematical model and algorithm optimization.

[0139] In the energy response, the amplitude decay term : This sub-item is used to describe the amplitude of the vibration mode of each chamber or valve of the heart and its attenuation characteristics over time. The dynamic changes of the heart structure are extracted through image processing technology, and its movement amplitude is quantified and its rate of change over time is considered; frequency component item :This sub-item is used to describe the frequency components and phase angles of the vibration modes of each chamber or valve of the heart. It is obtained by analyzing the local frequency components through short-time Fourier transform (STFT) and is used to describe the vibration characteristics of the heart structure at different phases.

[0140] in, It is extracted through image processing technology; and Determined by cardiac motion model; and Obtained through STFT analysis; Obtained through time-frequency analysis algorithm statistics; The time axis is taken from the spatial and temporal characteristics data of the heart;

[0141] In the spectrum intensity of cardiac vibration, the spectrum intensity term : This item is used to capture the unique frequency components in cardiac motion and limit the impact within the time range through the window function to ensure that the frequency components in each cardiac cycle are accurately captured; nonlinear adjustment item : This sub-item is used to capture more complex frequency components and introduce a weight function to adjust the importance of different time periods to ensure the accuracy and reliability of spectrum intensity;

[0142] in, Calculated by the energy response formula; and This is achieved through nonlinear time-varying filter design; and Choose appropriate window function and weight function; Through experimental or empirical setting; Obtained through imaginary units; Generated by frequency variables; The time axis is taken from the spatial and temporal characteristics data of the heart;

[0143] Assume that a cardiovascular hospital is processing echocardiographic imaging data of a patient with valvular heart disease;

[0144] Assumptions ;

[0145] ;

[0146] Assumptions and Designed, Select Hanning Window, Set to uniform distribution, ; The value range is ;

[0147] ;

[0148] Assuming that the threshold is set to 0.85, since the calculated result 0.88 is greater than the set threshold, it shows that the patient's heart vibration spectrum has high accuracy and reliability, and can effectively capture the unique vibration modes of each chamber and valve of the heart. This is because the higher spectrum intensity index reflects the stability of the heart vibration mode in the current cardiac cycle, while ensuring the clarity of different frequency components. Through the above steps, the accurate analysis of the heart vibration mode is ensured, the scientificity and accuracy of heart disease diagnosis are improved, the practicality and reliability of the entire cardiac medical image automatic segmentation system are enhanced, and the diagnosis and treatment level of cardiovascular diseases is significantly improved.

[0149] This application considers that in order to solve the problem of cardiac tissue boundary tracking caused by noise interference, displacement error and inaccurate boundary definition in the prior art, the embodiment of the invention proposes this optional solution. This solution aims to achieve accurate tracking of myocardial boundaries, ensure the accuracy of boundary definition, generate optimized boundary positions, and solve the accuracy problem in the automatic segmentation of cardiac medical images by removing noise, eliminating displacement errors, grayscale normalization processing, and applying variational level set evolution algorithm, intelligently adjusting the energy changes at the cardiac tissue boundary, and ensuring the accuracy of boundary definition. Therefore, a new optional solution is proposed, which includes:

[0150] Based on the energy function, the variational level set evolution algorithm is used for optimization processing to intelligently adjust the energy change at the boundary of the cardiac tissue, achieve accurate tracking of the myocardial boundary, ensure the accuracy of the boundary definition, and generate the optimized boundary position, including:

[0151] Based on the energy function, Gaussian filtering is applied to remove noise, image registration technology is used to eliminate displacement errors, and grayscale normalization is performed to generate energy distribution of cardiac tissue boundaries;

[0152] The energy distribution of cardiac tissue boundary is calculated by the following formula:

[0153] ;

[0154] in, is the energy distribution at the cardiac tissue boundary; is the image intensity; and are the average gray values ​​of the inner and outer areas of the heart, respectively; and is the weight coefficient, which adjusts the energy contribution of the inner and outer regions of the heart respectively; is the regularization parameter used to control the boundary smoothness; is the domain of the image; is the gradient operator;

[0155] Based on the energy distribution of the cardiac tissue boundary, the level set function is initialized, the evolution rule is set, and the curvature term and the regularization parameter are combined to control the boundary smoothness to ensure the stability of the evolution process, so as to generate the evolution of the level set function over time;

[0156] The evolution of the level set function over time is calculated using the following formula:

[0157] ;

[0158] in, is the evolution of the level set function over time; is the Diracdelta function; and is the weight coefficient; is the regularization parameter; is the nonlinear adjustment factor; is the curvature term; and is a nonlinear parameter; is the level set function; is the image intensity; and are the average gray values ​​of the inner and outer areas of the heart, respectively; is the gradient operator; is the level set function The gradient vector divided by its modulus is the unit gradient vector; is the domain of the image;

[0159] Based on the evolution of the level set function over time, phase contrast technology is applied to enhance the visibility of the cardiac tissue boundary, adjust the small deformation of the cardiac boundary in different cardiac cycles, and generate an optimized boundary position.

[0160] This method aims to further improve the accuracy of the definition of cardiac tissue boundaries, effectively analyze the changes in the boundaries of the heart during different cardiac cycles, and capture its unique energy distribution. To this end, this scheme adopts a series of steps, from denoising to energy distribution calculation, to level set function evolution, and finally phase contrast enhancement, to ensure that the final generated cardiac tissue boundary can truly reflect the changes in cardiac structure. The overall design is based on improving the reliability and clinical applicability of cardiac tissue boundaries through mathematical model and algorithm optimization.

[0161] In the energy distribution of cardiac tissue boundary, the internal region energy term : This item is used to describe the energy contribution of the heart region through the weight coefficient Adjust the energy influence of the inner area of ​​the heart; the outer area energy term :This item is used to describe the energy contribution of the extracardiac region, through the weight coefficient Adjust the energy influence of the heart's outer area; boundary smoothing term :This item is used to control the boundary smoothness through the regularization parameter Control the degree of boundary smoothing;

[0162] in, Obtained through image preprocessing; and Obtained through image segmentation technology; and Through experimental or empirical setting; Through experimental or empirical setting; is the gradient operator; Obtained through the definition domain of the image;

[0163] In the time evolution of the level set function, the energy difference term : This item is used to describe the energy difference between the inner and outer regions of the heart, through the Diracdelta function and weight coefficient Adjust the contribution of energy differences to the boundary; boundary smoothing term :This item is used to control the boundary smoothness through the regularization parameter and the unit gradient vector Implement; curvature adjustment item :This sub-item is used to combine the curvature term with the nonlinear parameter to control the stability of the boundary evolution. , curvature term and nonlinear parameters accomplish;

[0164] in, Obtained through the Dirac delta function; and Through experimental or empirical setting; Through experimental or empirical setting; Through experimental or empirical setting; Obtained through curvature calculation; and Through experimental or empirical setting; is the level set function; Obtained through image preprocessing; and Obtained through image segmentation technology; is the gradient operator; is the calculated level set function The unit gradient vector of ; Obtained through the definition domain of the image;

[0165] Assume that the cardiology department of a large general hospital is processing cardiac MRI image data of a healthy volunteer to study the normal anatomical structure of the heart; ;

[0166] ;

[0167] Assumptions Already designed, Obtained through curvature calculation, ;

[0168] ;

[0169] Assuming that the threshold is set to 0.9, since the calculated result 0.98 is greater than the set threshold, it shows that the cardiac tissue boundary of the healthy volunteer has a high degree of accuracy and reliability, and can accurately depict the normal anatomical structure of each chamber and valve of the heart. This is because the higher evolution rate reflects the stability and clarity of the cardiac tissue boundary in the current cardiac cycle, while ensuring the accuracy of different frequency components. Through the above steps, the accurate analysis of the cardiac tissue boundary is achieved, the scientificity and accuracy of the study of cardiac anatomical structure are improved, the performance of the entire medical image automatic segmentation system is optimized, and the quality and reliability of cardiovascular image analysis are significantly enhanced.

[0170] Figure 2 A schematic diagram of a structure of a cardiac medical image automatic segmentation system based on lossless processing is provided for an embodiment of the present application. Figure 2 As shown, the device comprises:

[0171] The receiving module 21 is used to receive a dynamic sequence of cardiac motion images through an imaging device, integrate information from different imaging sources, and generate dynamic cardiac imaging data;

[0172] The analysis module 22 is used to analyze the frequency response of the heart at different phases based on the dynamic cardiac imaging data by using a time-frequency analysis algorithm, identify the unique vibration modes of each chamber and valve of the heart, and use adaptive harmonic reconstruction technology to enhance the clarity and accuracy of important physiological signals, and generate a temporal and spatial consistency probability map of cardiac motion;

[0173] An adjustment module 23 is used to intelligently adjust the energy change at the boundary of cardiac tissue based on the spatiotemporal consistent probability map of cardiac motion by using a variational level set evolution algorithm to achieve accurate tracking of the myocardial boundary, and to perform anatomical coherence processing by using a topology optimization method to generate a fine segmentation result of the cardiac structure;

[0174] The measurement module 24 is used to perform morphological measurement based on the result of the fine segmentation of the cardiac structure, conduct a comprehensive evaluation in combination with clinical diagnostic standards, and generate an automatic segmentation report of cardiac medical images.

[0175] Figure 2 The automatic segmentation system of cardiac medical images based on lossless processing can be performed Figure 1 The implementation principle and technical effect of the method for automatic segmentation of cardiac medical images based on lossless processing described in the embodiment shown are not repeated here. The specific way in which each module and unit performs operations in the automatic segmentation system for cardiac medical images based on lossless processing in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0176] In one possible design, Figure 2The embodiment shown in the figure can be implemented as a computing device, such as a cardiac medical image automatic segmentation system based on lossless processing. Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0177] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0178] The processing component 32 is used to: receive dynamic sequence images of cardiac motion through an imaging device, integrate information from different imaging sources, and generate dynamic cardiac imaging data; based on the dynamic cardiac imaging data, use a time-frequency analysis algorithm to analyze the frequency response of the heart in different phases, identify the unique vibration modes of each chamber and valve of the heart, use adaptive harmonic reconstruction technology to enhance the clarity and accuracy of important physiological signals, and generate a temporal and spatial consistent probability map of cardiac motion; based on the temporal and spatial consistent probability map of cardiac motion, use a variational level set evolution algorithm to intelligently adjust the energy changes at the boundaries of cardiac tissues to achieve accurate tracking of myocardial boundaries, use a topology optimization method to perform anatomical continuity processing, and generate a fine segmentation result of the cardiac structure; based on the fine segmentation result of the cardiac structure, perform morphological measurements, combine clinical diagnostic standards for comprehensive evaluation, and generate an automatic segmentation report of cardiac medical images.

[0179] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0180] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0181] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0182] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0183] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0184] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0185] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment is a method for automatic segmentation of cardiac medical images based on lossless processing.

[0186] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0187] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0188] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0189] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application 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 make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for automatic segmentation of cardiac medical images based on lossless processing, characterized in that: include: Receive dynamic sequence images of cardiac motion through imaging equipment, integrate information from different imaging sources, and generate dynamic cardiac imaging data; Based on the dynamic cardiac imaging data, a time-frequency analysis algorithm is used to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, and use adaptive harmonic reconstruction technology to enhance the clarity and accuracy of important physiological signals, and generate a temporal and spatial consistent probability map of cardiac motion; Based on the spatiotemporal consistent probability map of cardiac motion, the variational level set evolution algorithm is used to intelligently adjust the energy change at the boundary of cardiac tissue to achieve accurate tracking of the myocardial boundary, and the topology optimization method is used to perform anatomical coherence processing to generate a fine segmentation result of the cardiac structure; Based on the result of the detailed segmentation of the cardiac structure, morphological measurements are performed, and a comprehensive evaluation is performed in combination with clinical diagnostic standards to generate an automatic segmentation report of cardiac medical images.

2. The method according to claim 1, characterized in that Based on the dynamic cardiac imaging data, the time-frequency analysis algorithm is used to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, and use adaptive harmonic reconstruction technology to enhance the clarity and accuracy of important physiological signals, and generate a temporal and spatial consistency probability map of cardiac motion, including: Based on the dynamic cardiac image data, extracting the spatiotemporal information within the complete cardiac cycle to generate an image sequence of the overall motion characteristics of the heart; Based on the image sequence of the overall motion characteristics of the heart, a time-frequency analysis algorithm is used to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, capture the unique frequency components of the heart motion, and generate a structural characteristic spectrum diagram; Based on the structural characteristic spectrum diagram, adaptive harmonic reconstruction technology is used to intelligently adjust harmonic parameters, strengthen the detailed performance of the heart structure in different cardiac cycles, enhance the clarity and accuracy of important physiological signals, and generate dynamic and clear images of the heart; Based on the dynamic clear images of the heart, the spatial distribution and time-varying laws of the heart structure are comprehensively considered and mapped to a continuous time series to generate a temporal and spatially consistent probability map of the heart movement.

3. The method according to claim 2, characterized in that The method based on the image sequence of the overall motion characteristics of the heart uses a time-frequency analysis algorithm to analyze the frequency response of the heart at different phases, identify the unique vibration modes of each chamber and valve of the heart, capture the unique frequency components of the heart motion, and generate a structural characteristic spectrum diagram, including: Based on the image sequence of the overall motion characteristics of the heart, the dynamic changes of each chamber and valve of the heart are extracted to generate temporal and spatial characteristic data of the heart; Based on the temporal and spatial characteristic data of the heart, the frequency response of the heart at different phases is analyzed by using a time-frequency analysis algorithm, the unique vibration modes of each chamber and valve of the heart are identified and quantified, the unique frequency components in the heart movement are captured, and the heart vibration spectrum is generated; Based on the cardiac vibration spectrum, through frequency domain filtering analysis, we focus on the key vibration characteristics of the cardiac structure and generate effective vibration patterns: Based on the effective vibration mode, reference information is displayed through visualization technology to generate a structural feature spectrum diagram.

4. The method according to claim 2, characterized in that: Based on the structural characteristic spectrum diagram, the adaptive harmonic reconstruction technology is used to intelligently adjust the harmonic parameters, strengthen the detailed performance of the heart structure in different cardiac cycles, enhance the clarity and accuracy of important physiological signals, and generate a dynamic and clear image of the heart, including: Based on the structural characteristic spectrum diagram, key frequency components in different cardiac cycles of each chamber and valve of the heart are marked to generate a list of key frequency components; Based on the above key frequency component list, adaptive harmonic reconstruction technology is applied to intelligently adjust harmonic parameters, enhance the details of the heart structure in different cardiac cycles, and generate enhanced heart vibration signals; Based on the enhanced cardiac vibration signal, further multi-scale detail enhancement processing is performed to enhance the static structure visibility and dynamic change expression, and generate a dynamic enhanced cardiac image; Based on the dynamic enhanced cardiac image, the clarity and accuracy of important physiological signals are enhanced to generate a dynamic clear cardiac image.

5. The method according to claim 1, characterized in that Based on the spatiotemporal consistent probability map of cardiac motion, the variational level set evolution algorithm is used to intelligently adjust the energy change at the boundary of cardiac tissue to achieve accurate tracking of the myocardial boundary, and the topology optimization method is used to perform anatomical coherence processing to generate a fine segmentation result of the cardiac structure, including: Based on the cardiac motion spatiotemporal consistent probability map, identifying energy changes at the cardiac tissue boundary and generating a cardiac tissue boundary energy change map; Based on the energy change map of the cardiac tissue boundary, the variational level set evolution algorithm is used to intelligently adjust the energy change at the cardiac tissue boundary by optimizing the energy function, so as to achieve accurate tracking of the myocardial boundary, ensure the accuracy of the boundary definition, and generate a preliminary cardiac structure boundary contour; Based on the preliminary cardiac structure boundary contour, a topology optimization method is used to perform anatomical coherence processing, eliminate potential discontinuities and unreasonable boundaries, and generate an optimized cardiac structure boundary contour; Based on the optimized cardiac structure boundary contour, the cardiac structure boundary smoothness and continuity are further optimized to generate a cardiac structure fine segmentation result.

6. The method according to claim 5, characterized in that Based on the energy change map of the cardiac tissue boundary, the variational level set evolution algorithm is used to optimize the energy function to intelligently adjust the energy change at the cardiac tissue boundary, achieve accurate tracking of the myocardial boundary, ensure the accuracy of boundary definition, and generate a preliminary cardiac structure boundary contour, including: Based on the cardiac tissue boundary energy change map, the physical characteristics of the cardiac tissue boundary and the image intensity information are quantified to generate an energy function: Based on the energy function, a variational level set evolution algorithm is used for optimization processing to intelligently adjust the energy change at the boundary of cardiac tissue, achieve accurate tracking of the myocardial boundary, ensure the accuracy of boundary definition, and generate an optimized boundary position; Based on the optimized boundary position, the geometric accuracy of boundary definition is improved, key features of the cardiac structure are retained, and a high-precision boundary position is generated; Based on the high-precision boundary position, multi-scale spatial correction processing is performed to generate a preliminary cardiac structure boundary contour.

7. The method according to claim 1, characterized in that Based on the result of the fine segmentation of the cardiac structure, morphological measurement is performed, and a comprehensive evaluation is performed in combination with clinical diagnostic standards to generate an automatic segmentation report of cardiac medical images, including: Based on the cardiac structure fine segmentation result, key dimensions and position information of the cardiac structure are quantified to generate a detailed cardiac anatomical parameter list; Based on the detailed cardiac anatomical parameter list, a series of morphological measurements are performed to accurately calculate cardiac structural volumetric indices and functional parameters, and generate cardiac morphological measurement data; Based on the cardiac morphology measurement data and combined with clinical diagnostic criteria, a comprehensive assessment of cardiac function status is performed to generate a cardiac function assessment report; Based on the cardiac function assessment report, a personalized comprehensive clinical interpretation is performed to generate an automatic segmentation report of cardiac medical images.

8. An automatic segmentation system for cardiac medical images based on lossless processing, characterized in that: include: A receiving module, used for receiving a dynamic sequence of cardiac motion images through an imaging device, integrating information from different imaging sources, and generating dynamic cardiac imaging data; An analysis module is used to analyze the frequency response of the heart at different phases based on the dynamic cardiac imaging data by using a time-frequency analysis algorithm, identify the unique vibration modes of each chamber and valve of the heart, and use adaptive harmonic reconstruction technology to enhance the clarity and accuracy of important physiological signals and generate a temporal and spatial consistency probability map of cardiac motion; An adjustment module is used to intelligently adjust the energy change at the boundary of cardiac tissue based on the spatiotemporal consistent probability map of cardiac motion by using a variational level set evolution algorithm to achieve accurate tracking of the myocardial boundary, and to perform anatomical coherence processing by using a topology optimization method to generate a fine segmentation result of the cardiac structure; The measurement module is used to perform morphological measurement based on the fine segmentation results of the cardiac structure, conduct a comprehensive evaluation in combination with clinical diagnostic standards, and generate an automatic segmentation report for cardiac medical images.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for automatic segmentation of cardiac medical images based on lossless processing as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, the method for automatic segmentation of cardiac medical images based on lossless processing as described in any one of claims 1 to 7 is implemented.

Images