Auxiliary diagnosis method, device, medium and product based on historical image data

Abstract

This invention discloses an auxiliary diagnostic method, device, medium, and product based on historical image data. The method relates to the field of medical image processing. The method includes: acquiring an image data set of a target object; determining a feature matrix based on the image data set, the feature matrix including at least two feature vectors, each feature vector including multiple quantitative feature change rates corresponding to a target follow-up time, the quantitative feature change rate being the rate of change of a quantitative feature between the target follow-up time and a reference follow-up time; determining a correlation coefficient matrix corresponding to the feature matrix based on a pre-trained correlation model; determining at least one predetermined correlation coefficient progression curve corresponding to the disease identifier of the target object; and determining the stage identifier of the current disease stage of the target object based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve. This invention addresses the problem of low accuracy in the auxiliary diagnostic results determined by existing auxiliary diagnostic methods.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of medical image processing technology, and in particular to an auxiliary diagnostic method, device, medium and product based on historical image data. Background Technology

[0002] In the imaging diagnosis of follow-up patients, longitudinal comparison of historical imaging data is the core basis for assessing recovery status. However, the auxiliary diagnostic results determined by existing technologies based on longitudinal comparison of historical imaging data have low accuracy. Summary of the Invention

[0003] This invention provides an auxiliary diagnostic method, device, medium, and product based on historical image data to address the problem of low accuracy in the auxiliary diagnostic results determined by existing auxiliary diagnostic methods.

[0004] According to one aspect of the present invention, an auxiliary diagnosis based on historical image data is provided, the method comprising:

[0005] Obtain an image data set of the target object, the image data set including multiple image data combinations, the image data combination including all image data of the target object at one follow-up time;

[0006] A feature matrix is ​​determined based on the image data set. The feature matrix includes at least two feature vectors. Each feature vector includes multiple quantitative feature change rates at a corresponding target follow-up time. The quantitative feature change rate is the rate of change of a quantitative feature between the corresponding target follow-up time and a reference follow-up time. The reference follow-up time is the previous follow-up time or the first follow-up time. The target follow-up time is a time other than the first follow-up time.

[0007] The correlation coefficient matrix corresponding to the feature matrix is ​​determined based on the pre-trained correlation model. The pre-trained correlation model is used to determine the nonlinear dependency weights of each feature rate of change pair in the feature matrix, and the normalized result of all the nonlinear dependency weights is output as the correlation coefficient matrix.

[0008] Determine at least one predetermined correlation coefficient progression curve corresponding to the disease identifier of the target object, and determine the stage identifier of the current disease stage of the target object based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve.

[0009] According to another aspect of the present invention, an auxiliary diagnostic device based on historical image data is provided, the device comprising:

[0010] The acquisition module is used to acquire a set of image data of a target object. The set of image data includes multiple combinations of image data, and the combination of image data includes all image data of the target object during a follow-up time.

[0011] The feature matrix determination module is used to determine a feature matrix based on the image data set. The feature matrix includes at least two feature vectors. Each feature vector includes multiple quantitative feature change rates at a corresponding target follow-up time. The quantitative feature change rate is the rate of change of a quantitative feature between the corresponding target follow-up time and a reference follow-up time. The reference follow-up time is the previous follow-up time or the first follow-up time. The target follow-up time is a time other than the first follow-up time.

[0012] The coefficient matrix determination module is used to determine the correlation coefficient matrix corresponding to the feature matrix based on the pre-trained correlation model. The pre-trained correlation model is used to determine the nonlinear dependency weights of each feature change rate pair in the feature matrix, and outputs the normalized result of all the nonlinear dependency weights as the correlation coefficient matrix.

[0013] The stage identifier determination module is used to determine at least one predetermined correlation coefficient progress curve corresponding to the disease identifier of the target object, and determine the stage identifier of the current disease stage of the target object based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progress curve.

[0014] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0015] One or more processors;

[0016] Storage device for storing one or more programs.

[0017] When one or more programs are executed by one or more processors, the one or more processors implement the assisted diagnostic method based on historical image data as described in any embodiment of the present invention.

[0018] According to another aspect of the present invention, a computer-readable storage medium is provided, which stores computer instructions for causing a processor to execute and implement the auxiliary diagnostic method based on historical image data as described in any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer program product is provided, which, when executed by a processor, implements the auxiliary diagnostic method based on historical image data as described in any embodiment of the present invention.

[0020] The technical solution of this invention includes an image dataset comprising multiple image data combinations, each including all image data of the target object at a specific follow-up time. The feature matrix includes at least two feature vectors, each including multiple quantitative feature change rates corresponding to the target follow-up time. These quantitative feature change rates represent the rate of change of a quantitative feature between the target follow-up time and a reference follow-up time, where the reference follow-up time is the previous or first follow-up time, and the target follow-up time is a time other than the first follow-up time. The correlation coefficient matrix includes the correlation coefficients of each feature change rate pair in the feature matrix. Therefore, it at least includes the synergistic changes of different biological processes at the same follow-up time point and the autocorrelation of the same quantitative feature at different time points. This encodes a high-dimensional spatiotemporal dependency network of disease evolution, enabling a paradigm shift from single-indicator threshold judgment to multi-indicator relationship modeling. The target model includes at least one predetermined correlation coefficient progression curve corresponding to the target object's disease identifier. Therefore, based on the correlation coefficient matrix and the target model, the current disease stage of the target object can be accurately determined, providing accurate data support for doctors to make an overall judgment on the disease progression of the target object.

[0021] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A flowchart illustrating the auxiliary diagnostic method based on historical image data provided in an embodiment of the present invention;

[0024] Figure 2 This is another flowchart illustrating the auxiliary diagnostic method based on historical image data provided in an embodiment of the present invention.

[0025] Figure 3A This is a schematic diagram of the structure of the auxiliary diagnostic device based on historical image data provided in an embodiment of the present invention;

[0026] Figure 3B This is another structural schematic diagram of the auxiliary diagnostic device based on historical image data provided in an embodiment of the present invention;

[0027] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0028] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0029] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0030] Figure 1 This is a flowchart illustrating an auxiliary diagnostic method based on historical image data provided in an embodiment of the present invention. This embodiment is applicable to situations where auxiliary diagnostic results are automatically determined based on image data from multiple follow-up periods. The method can be executed by an auxiliary diagnostic device based on historical image data, which can be implemented in hardware and / or software and can be configured in electronic devices such as computers or servers. Figure 1 As shown, the method in this embodiment includes:

[0031] S110. Obtain the image data set of the target object. The image data set includes multiple image data combinations, and the image data combinations include all image data of the target object during a follow-up period.

[0032] The target group is also known as the patient.

[0033] Considering that the imaging data required for each follow-up visit of the target subject may be collected within one day or may take several days, the follow-up time in this embodiment can optionally include the follow-up visit date and a predetermined time period before that date. The follow-up visit date is the date on which the method described in this embodiment of the invention is used to assist in diagnosis. Of course, the method described in this embodiment of the invention is usually used by the target subject's attending physician to assist in diagnosis; therefore, the follow-up visit date is usually also the date on which the target subject sees a clinician. The predetermined time period is a configurable item and can be set according to actual needs, such as one week.

[0034] All imaging data of the target subject at each follow-up time is combined into one imaging data set. For example, if the target subject has an ultrasound image taken at the three-month follow-up, an ultrasound and CT image taken at the six-month follow-up, and an ultrasound and CT image taken at the one-year follow-up, the ultrasound image taken at three months post-surgery is used as the first imaging data set; the ultrasound and CT image taken at six months post-surgery is used as the second imaging data set; and the ultrasound and CT image taken at one year post-surgery is used as the third imaging data set.

[0035] S120. Determine a feature matrix based on the image data set. The feature matrix includes at least two feature vectors. Each feature vector includes multiple quantitative feature change rates at the corresponding target follow-up time. The quantitative feature change rate is the rate of change of the quantitative feature between the corresponding target follow-up time and the reference follow-up time. The reference follow-up time is the previous follow-up time or the first follow-up time, and the target follow-up time is a time other than the first follow-up time.

[0036] The feature matrix comprises multiple feature vectors, each containing the rate of change of multiple quantitative features at the corresponding follow-up time. Quantitative features can be understood as indicators that objectively reflect the pathological and physiological properties of tissues in a continuous numerical form, such as volume and diameter, extracted through mathematical algorithms that calculate pixels / voxels in medical images. The rate of change of quantitative features is the rate of change of the quantitative features between the target follow-up time and the reference follow-up time.

[0037] For example, the reference follow-up time is the first follow-up time. The image dataset includes a combination of projection data at four follow-up times, arranged from earliest to latest: the first follow-up time, the second follow-up time, the third follow-up time, and the fourth follow-up time. The first follow-up time is set as the reference follow-up time. The fourth follow-up time corresponds to the A4 feature vector, which includes the rate of change of at least one quantitative feature between the fourth follow-up time and the first follow-up time; the third follow-up time corresponds to the A3 feature vector, which includes the rate of change of at least one quantitative feature between the third follow-up time and the first follow-up time; the second follow-up time corresponds to the A2 feature vector, which includes the rate of change of at least one quantitative feature between the second follow-up time and the first follow-up time.

[0038] For example, the reference follow-up time is the previous follow-up time. The image dataset includes a combination of image data at four follow-up times, arranged from earliest to latest: the first follow-up time, the second follow-up time, the third follow-up time, and the fourth follow-up time. The fourth follow-up time corresponds to a B4 feature vector, which includes the rate of change of at least one quantitative feature between the fourth and third follow-up times; the third follow-up time corresponds to a B3 feature vector, which includes the rate of change of at least one quantitative feature between the third and second follow-up times; and the second follow-up time corresponds to a B2 feature vector, which includes the rate of change of at least one quantitative feature between the second and first follow-up times.

[0039] In one embodiment, the feature vector corresponding to the fourth follow-up time includes the aforementioned A4 feature vector and the aforementioned B4 feature vector; similarly, the feature vector corresponding to the third follow-up time includes the aforementioned A3 feature vector and the aforementioned A4 feature vector, and the feature vector corresponding to the second follow-up time includes the aforementioned A2 feature vector and the aforementioned B2 feature vector.

[0040] In one embodiment, the feature vector includes the rate of change of multiple quantitative features at the target follow-up time, multiple quantitative feature data at the target follow-up time, and the multiple quantitative feature data at a reference follow-up time. In this embodiment, the feature vector carries more comprehensive data information, so the feature matrix containing the feature vector can provide richer feature information, resulting in higher accuracy of subsequent data analysis results.

[0041] The feature vector may include at least one of the following: structural feature change rate, metabolic feature change rate, and functional feature change rate; simultaneously, the feature matrix may include at least two of the following: structural feature change rate, metabolic feature change rate, and functional feature change rate. Taking brain follow-up as an example, structural features include, but are not limited to, lesion diameter, morphology, density, and signal changes; metabolic features, taking magnetic resonance spectroscopy as an example, may include the ratio of N-acetylaspartate (NAA) to choline (Cho); functional features, taking diffusion tensor imaging as an example, may include microstructural integrity parameters of white matter fiber tracts obtained from diffusion tensor imaging, and activated areas of functional cortical areas displayed by functional magnetic resonance imaging.

[0042] In summary, the rate of change of a feature can eliminate the influence of time length. Therefore, in the embodiments of the present invention, the rate of change of the same quantitative feature in different feature vectors is comparable; each quantitative feature in the same feature vector can reflect the rate of change of biological features in different dimensions.

[0043] S130. Based on the pre-trained correlation model, determine the correlation coefficient matrix corresponding to the feature matrix. The pre-trained correlation model is used to determine the nonlinear dependency weights of each feature change rate pair in the feature matrix, and the normalized result of all nonlinear dependency weights is output as the correlation coefficient matrix.

[0044] The feature rate of change pairs include the rate of change of two quantitative features in the quantitative feature matrix. A pre-trained association model is used to determine the nonlinear dependency weights between pairwise quantitative feature rates of change in the feature matrix.

[0045] After the feature matrix is ​​determined, it is input into a pre-trained correlation model, which then performs data analysis on the feature matrix to obtain a correlation coefficient matrix. Since the correlation coefficient matrix includes the correlation coefficients of all feature rate of change pairs in the feature matrix, it achieves full coverage encoding of each feature rate of change pair in the feature matrix. This fully preserves the spatiotemporal dependencies between multidimensional biological feature changes during disease evolution, such as directional and structural information. It provides comprehensive data information for understanding disease mechanisms at the system level, evaluating treatment effectiveness, and predicting disease progression, solving the problems of low utilization efficiency and incomplete correlation analysis of historical image data.

[0046] In one embodiment, the first element of the feature vector is a time element, that is, the specific time of the target follow-up visit. A feature vector is a row vector of the feature matrix. In this way, the feature matrix effectively associates the follow-up visit time with the feature vector.

[0047] A pre-trained correlation model is used to determine the nonlinear dependency weights between the rates of change of two quantitative features at different follow-up times in the feature matrix. These nonlinear dependency weights reflect the correlation between the two rates of change. Then, the nonlinear dependency weights of all pairs of quantitative feature rate changes are normalized to obtain a correlation coefficient matrix. Thus, the larger the nonlinear dependency weight, the larger its corresponding correlation coefficient, and the stronger the correlation between the two rates of change; conversely, the smaller the corresponding correlation coefficient, the weaker the correlation between the two rates of change.

[0048] The correlation matrix includes the correlation coefficients of each pair of feature rate of change in the feature matrix. The correlation coefficient matrix can be an asymmetric matrix, reflecting the directionality of the nonlinear dependence of the feature rate of change; if a symmetric constraint is applied, a symmetric correlation coefficient matrix can be obtained. If the correlation model is a standard multi-head attention network, the correlation matrix is ​​an asymmetric matrix.

[0049] S140. Determine at least one predetermined correlation coefficient progression curve corresponding to the disease identifier of the target object, and determine the stage identifier of the current disease stage of the target object based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve.

[0050] To facilitate the description of disease progression, clinical practice uniformly discretizes the continuous disease process into several clinically significant stages, such as early, middle, late, remission, and progression stages. Stage identifiers are used to characterize the current stage of the target disease.

[0051] The predetermined association coefficient progression curve is a standard reference pattern for the evolution of the association strength over time between different characteristic change rate pairs, pre-mined and validated based on a large amount of clinical sample data. Each disease identifier is pre-associated with at least one predetermined association coefficient progression curve.

[0052] Therefore, at least one predetermined correlation progression curve can be determined based on the disease identifier of the target object. Then, the correlation coefficient matrix is ​​matched with the at least one predetermined correlation coefficient progression curve. By quantifying the similarity between the current correlation coefficient matrix and each disease stage, the current disease process position of the patient, i.e., the disease stage, can be inferred.

[0053] The technical solution provided by this invention includes an image dataset comprising multiple image data combinations, each including all image data of the target object at a specific follow-up time. The feature matrix includes at least two feature vectors, each including multiple quantitative feature change rates corresponding to the target follow-up time. These quantitative feature change rates represent the rate of change of a quantitative feature between the target follow-up time and a reference follow-up time, where the reference follow-up time is the previous or first follow-up time, and the target follow-up time is a time other than the first follow-up time. The correlation coefficient matrix includes the correlation coefficients of each feature change rate pair in the feature matrix. Therefore, it at least includes the synergistic changes of different biological processes at the same follow-up time point and the autocorrelation of the same quantitative feature at different time points. This encodes a high-dimensional spatiotemporal dependency network of disease evolution, enabling a paradigm shift from single-indicator threshold judgment to multi-indicator relationship modeling. The target model includes at least one predetermined correlation coefficient progression curve corresponding to the target object's disease identifier. Therefore, based on the correlation coefficient matrix and the target model, the current disease stage of the target object can be accurately determined, providing accurate data support for doctors to make an overall judgment on the disease progression of the target object.

[0054] Based on the aforementioned example, the image dataset includes image data in at least two modalities, which are distributed across a combination of image data from multiple follow-up times.

[0055] Specifically, a modality can be understood as a specific type of data source or representation. Images of different modalities correspond to different imaging principles. For example, ultrasound images, CT images, MRI images, and PET images all have different imaging principles, and therefore different image modalities.

[0056] Therefore, each image dataset may include medical images from one image modality, such as only ultrasound images; or it may include medical images from at least two modalities, such as both ultrasound and CT images. Since the image dataset includes image data from all follow-up times to be analyzed, it includes image data from at least two modalities. Correspondingly, the correlation matrix corresponding to this image dataset also includes cross-time and cross-modal correlation coefficients, which characterize the correlation between changes in different quantitative features at different follow-up times. These changing correlations, or dynamic correlations, can transform how different biological changes interact during disease evolution into calculable, comparable, and verifiable quantitative indicators, thereby achieving a paradigm shift from single-point interpretation in existing technologies to systemic understanding, and from empirical medicine to precision medicine. Examples include the correlation between changes in tumor volume and metabolic levels, and the correlation between changes in cerebral blood perfusion and nerve fiber repair.

[0057] Based on the aforementioned embodiments, "S120, determining the feature matrix according to the image data set" can be refined as follows: taking the predetermined image data in the latest image data combination as a reference, for each other image data in the image data set, determine the registration algorithm corresponding to the current modality combination, and register the image data to the predetermined image data based on the registration algorithm. The current modality combination is a combination of the modality identifier of the predetermined image data and the modality identifier of the image data; and determine the feature matrix according to the image registration result of the image data set.

[0058] For example, the predetermined image data in the latest image data set is a CT image. If the image data set also includes an ultrasound image, the ultrasound image is registered to the CT image based on the registration algorithm corresponding to the ultrasound and CT images.

[0059] This embodiment implements multimodal registration of image datasets based on a hybrid registration strategy, which includes rigid and non-rigid registration algorithms. Rigid registration algorithms are used to correct for differences in body position, translation, and rotation, and are suitable for tissues with minimal deformation, such as bones and lung nodules. Non-rigid registration is used to address misalignments caused by respiratory motion and tissue deformation, and is suitable for soft tissues such as the liver and brain. Multimodal registration establishes spatial correspondences by extracting multimodal image features, such as CT bony structures and MRI soft lesion features, achieving precise alignment of image data from different modalities.

[0060] Figure 2 This is another schematic flowchart illustrating the assisted diagnostic method based on historical image data provided in this embodiment of the invention. The technical solution of this embodiment can be combined with other embodiments; for the same or related parts, they can be described in conjunction with the descriptions of other embodiments, and will not be repeated here. Figure 2 As shown, the method in this embodiment may specifically include:

[0061] S210. Obtain the image data set of the target object. The image data set includes multiple image data combinations, and the image data combinations include all image data of the target object during a follow-up period.

[0062] S220. Determine a feature matrix based on the image data set. The feature matrix includes at least two feature vectors. Each feature vector includes multiple quantitative feature change rates at the corresponding target follow-up time. The quantitative feature change rate is the rate of change of the quantitative feature between the corresponding target follow-up time and the reference follow-up time. The reference follow-up time is the previous follow-up time or the first follow-up time, and the target follow-up time is a time other than the first follow-up time.

[0063] S230. Based on the pre-trained correlation model, determine the correlation coefficient matrix corresponding to the feature matrix. The pre-trained correlation model is used to determine the nonlinear dependency weights of each feature change rate pair in the feature matrix, and the normalized result of all nonlinear dependency weights is output as the correlation coefficient matrix.

[0064] S2401. Determine at least one predetermined correlation coefficient progression curve corresponding to the disease identifier of the target object.

[0065] S2402. Generate and display an analysis report based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve, wherein the analysis report includes at least one of the following: a key image comparison map, a predetermined feature quantification table, a disease progression speed curve, diagnostic information, and a difference map corresponding to the key image comparison map, as well as a stage identifier corresponding to the disease stage.

[0066] Key image comparison images include regions of interest from images taken at various follow-up times. For example, tomographic CT images of the target tumor at multiple follow-up times. Key image comparison images are preferably displayed in chronological order of follow-up time.

[0067] The difference map is a result corresponding to the key image comparison map, which includes the change of the region of interest at each target follow-up time relative to the baseline follow-up time, or the change of the region of interest at each target follow-up time compared to the previous follow-up time.

[0068] The predetermined feature quantification table includes information on the change of at least one predetermined quantitative feature from baseline time to the latest follow-up time. The predetermined quantitative features included in the predetermined feature quantification table may optionally be associated with disease identifiers.

[0069] Disease progression curves are used to illustrate the rate of disease progression in a target individual at each stage, helping clinicians to grasp the overall disease progression. For example, they can differentiate between "physiological fluctuations," such as short-term changes in inflammation and edema; "pathological progression," such as continuous tumor growth; and determine the recovery stage, such as progression / stabilization / slowing down / complete recovery.

[0070] Diagnostic information includes descriptive information about the target individual's current condition, such as good disease control or disease progression faster than expected.

[0071] The analysis report may also include clinical recommendations, such as adjusting the treatment plan or shortening the follow-up period.

[0072] This embodiment presents users with a variety of clinical diagnostic information by displaying analysis reports, making it easier for users to understand the overall disease progression of the target subject.

[0073] Based on the foregoing embodiments, the analysis report is updated in response to modifications to the stage identifier and / or diagnostic information.

[0074] Specifically, if the user determines that there is an error in the stage identifier and / or diagnostic information in the analysis report, they can enter the accurate stage identifier and / or diagnostic information in the interactive interface; the processor updates the analysis report based on the received stage identifier and / or diagnostic information.

[0075] Modifying analysis reports through interactive methods not only ensures the accuracy of the reports by basing them on user judgments but also improves the user experience.

[0076] Based on the aforementioned embodiments, the step "determine the stage identifier of the current disease stage of the target object according to the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve" in S240 is refined to input the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve into the pre-trained analysis model, determine and display the analysis report.

[0077] This embodiment analyzes the correlation coefficient matrix and the progress curve of at least one predetermined correlation coefficient using a pre-trained analysis model. By determining the matching degree between the corresponding correlation coefficient in the correlation coefficient matrix and the progress curve of each predetermined correlation coefficient, the stage identifier represented by each predetermined correlation coefficient is determined. Based on the disease stage identifier represented by each predetermined correlation coefficient, the stage identifier of the target object's current disease stage and the confidence level of the stage identifier are determined.

[0078] Because the pre-trained analysis model can learn many non-linear relationships based on data during the training process, an accurate analysis report can be determined based on the pre-trained analysis model.

[0079] At this point, in response to modifications to the stage identifiers and / or diagnostic information, the analysis report and the pre-trained analysis model are updated. Automatic updates to the pre-trained analysis model improve its generalization ability and the accuracy of the analysis reports it generates.

[0080] Figure 3A This is a schematic diagram of the structure of an auxiliary diagnostic device based on historical image data provided in an embodiment of the present invention. The auxiliary diagnostic device based on historical image data can be used in terminal devices, as shown below. Figure 3A As shown, the auxiliary diagnostic device based on historical image data includes:

[0081] The acquisition module 310 is used to acquire an image data set of a target object, the image data set including multiple image data combinations, and the image data combination including all image data of the target object during a follow-up time.

[0082] The feature matrix determination module 320 is used to determine a feature matrix based on the image data set. The feature matrix includes at least two feature vectors. Each feature vector includes multiple quantitative feature change rates at a corresponding target follow-up time. The quantitative feature change rate is the rate of change of a quantitative feature between the corresponding target follow-up time and a reference follow-up time. The reference follow-up time is the previous follow-up time or the first follow-up time. The target follow-up time is a time other than the first follow-up time.

[0083] The coefficient matrix determination module 330 is used to determine the correlation coefficient matrix corresponding to the feature matrix based on the pre-trained correlation model. The pre-trained correlation model is used to determine the nonlinear dependency weights of each feature change rate pair in the feature matrix and output the normalized result of all the nonlinear dependency weights as the correlation coefficient matrix.

[0084] The stage identifier determination module 340 is used to determine at least one predetermined correlation coefficient progress curve corresponding to the disease identifier of the target object, and determine the stage identifier of the current disease stage of the target object based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progress curve.

[0085] The technical solution provided by this invention includes an image dataset comprising multiple image data combinations, each including all image data of the target object at a specific follow-up time. The feature matrix includes at least two feature vectors, each including multiple quantitative feature change rates corresponding to the target follow-up time. These quantitative feature change rates represent the rate of change of a quantitative feature between the target follow-up time and a reference follow-up time, where the reference follow-up time is the previous or first follow-up time, and the target follow-up time is a time other than the first follow-up time. The correlation coefficient matrix includes the correlation coefficients of each feature change rate pair in the feature matrix. Therefore, it at least includes the synergistic changes of different biological processes at the same follow-up time point and the autocorrelation of the same quantitative feature at different time points. This encodes a high-dimensional spatiotemporal dependency network of disease evolution, enabling a paradigm shift from single-indicator threshold judgment to multi-indicator relationship modeling. The target model includes at least one predetermined correlation coefficient progression curve corresponding to the target object's disease identifier. Therefore, based on the correlation coefficient matrix and the target model, the current disease stage of the target object can be accurately determined, providing accurate data support for doctors to make an overall judgment on the disease progression of the target object.

[0086] In one embodiment, the feature matrix includes at least two of the following: rate of change of structural features, rate of change of metabolic features, and rate of change of functional features.

[0087] In one embodiment, the image dataset includes image data in at least two modalities, which are distributed across a combination of image data from multiple follow-up times.

[0088] In one embodiment, the stage identifier determination module 340 is used for:

[0089] An analysis report is generated and displayed based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progress curve;

[0090] The analysis report includes at least one of the following: a key image comparison map, a predetermined feature quantification table, a disease progression rate curve, diagnostic information, and a difference map corresponding to the key image comparison map, as well as a stage identifier corresponding to the disease stage.

[0091] In one embodiment, the stage identifier determination module 340 is used for:

[0092] Input the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve into the pre-trained analysis model, and determine and display the analysis report.

[0093] In one embodiment, such as Figure 3B As shown, the device also includes an interactive model 350, which is used for:

[0094] In response to modifications to the stage identifier and / or the diagnostic information, the analysis report and the pre-trained analysis model are updated.

[0095] In one embodiment, the feature matrix determination module 320 is used for:

[0096] Based on the predetermined image data in the latest image data set, for each other image data in the image data set, a registration algorithm corresponding to the current modality combination is determined, and the image data is registered to the predetermined image data based on the registration algorithm. The current modality combination is a combination of the modality identifier of the predetermined image data and the modality identifier of the image data.

[0097] The feature matrix is ​​determined based on the image registration results of the image dataset.

[0098] Based on the aforementioned embodiments, the device is associated with a database and also includes an image retrieval module. The database is used to store follow-up images, and each follow-up image carries a patient identifier, acquisition time, image modality identifier, and disease type identifier.

[0099] The image retrieval module receives retrieval information and queries the image data set corresponding to the retrieval information. The retrieval information includes at least a patient identifier. In one embodiment, the detection information also includes a time period and / or a disease type identifier. The patient identifier includes at least one of name, ID number, and date of birth. The image data set is arranged in chronological order of examination time, or in a predetermined modal order, or grouped in a predetermined modal order, with each group containing images arranged in chronological order of examination time.

[0100] In one embodiment, if a query result is missing historical image data, the user can manually upload that historical image data to the database. This eliminates the possibility that the database lacks image data for any patient due to communication problems.

[0101] The auxiliary diagnostic device based on historical image data provided in the embodiments of the present invention can execute the auxiliary diagnostic method based on historical image data provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method execution.

[0102] It is worth noting that the various units and modules included in the above-mentioned auxiliary diagnostic device based on historical image data are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the protection scope of the embodiments of the present invention.

[0103] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0104] like Figure 4As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0105] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0106] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as assisted diagnosis based on historical image data.

[0107] In some embodiments, the auxiliary diagnosis based on historical image data can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via read-only memory (ROM) 12 and / or communication unit 19. When the computer program is loaded into random access memory (RAM) 13 and executed by processor 11, one or more steps of the auxiliary diagnosis based on historical image data described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform auxiliary diagnosis based on historical image data by any other suitable means (e.g., by means of firmware).

[0108] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0109] Computer programs for implementing the present invention based on historical image data-assisted diagnosis can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are performed. The computer programs can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0110] This invention provides a computer-readable storage medium storing computer instructions for causing a processor to perform an auxiliary diagnosis based on historical image data, including:

[0111] Obtain an image data set of the target object, the image data set including multiple image data combinations, the image data combination including all image data of the target object at one follow-up time;

[0112] A feature matrix is ​​determined based on the image data set. The feature matrix includes at least two feature vectors. Each feature vector includes multiple quantitative feature change rates at a corresponding target follow-up time. The quantitative feature change rate is the rate of change of a quantitative feature between the corresponding target follow-up time and a reference follow-up time. The reference follow-up time is the previous follow-up time or the first follow-up time. The target follow-up time is a time other than the first follow-up time.

[0113] The correlation coefficient matrix corresponding to the feature matrix is ​​determined based on the pre-trained correlation model. The pre-trained correlation model is used to determine the nonlinear dependency weights of each feature rate of change pair in the feature matrix, and the normalized result of all the nonlinear dependency weights is output as the correlation coefficient matrix.

[0114] Determine at least one predetermined correlation coefficient progression curve corresponding to the disease identifier of the target object, and determine the stage identifier of the current disease stage of the target object based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve.

[0115] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0116] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0117] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0118] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0119] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication unit 19, or installed from storage unit 18, or installed from ROM 12. When the computer program is executed by processor 11, it performs the functions defined in the methods of the embodiments of the present invention.

[0120] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the auxiliary diagnostic method based on historical image data according to any embodiment of the invention.

[0121] In implementing the computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0122] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0123] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. An auxiliary diagnostic method based on historical image data, characterized in that, The method includes: Obtain an image data set of the target object, the image data set including multiple image data combinations, the image data combination including all image data of the target object at one follow-up time; A feature matrix is ​​determined based on the image data set. The feature matrix includes at least two feature vectors. Each feature vector includes multiple quantitative feature change rates at a corresponding target follow-up time. The quantitative feature change rate is the rate of change of a quantitative feature between the corresponding target follow-up time and a reference follow-up time. The reference follow-up time is the previous follow-up time or the first follow-up time. The target follow-up time is a time other than the first follow-up time. The correlation coefficient matrix corresponding to the feature matrix is ​​determined based on the pre-trained correlation model. The pre-trained correlation model is used to determine the nonlinear dependency weights of each feature rate of change pair in the feature matrix, and the normalized result of all the nonlinear dependency weights is output as the correlation coefficient matrix. The feature rate of change pair includes two quantitative feature rates of change. Determine at least one predetermined correlation coefficient progression curve corresponding to the disease identifier of the target object, and determine the stage identifier of the current disease stage of the target object based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve.

2. The method according to claim 1, characterized in that, The feature matrix includes at least two of the following: the rate of change of structural features, the rate of change of metabolic features, and the rate of change of functional features.

3. The method according to claim 1, characterized in that, The image data set includes image data in at least two modalities, which are distributed across a combination of image data from multiple follow-up times.

4. The method according to claim 1, characterized in that, The step of determining the stage identifier of the current disease stage of the target object based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve includes: An analysis report is generated and displayed based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progress curve; The analysis report includes at least one of the following: a key image comparison map, a predetermined feature quantification table, a disease progression rate curve, diagnostic information, and a difference map corresponding to the key image comparison map, as well as a stage identifier corresponding to the disease stage.

5. The method according to claim 4, characterized in that, The step of generating and displaying an analysis report based on the correlation coefficient matrix and the at least one predetermined correlation coefficient progress curve includes: Input the correlation coefficient matrix and the at least one predetermined correlation coefficient progression curve into the pre-trained analysis model, and determine and display the analysis report.

6. The method according to claim 5, characterized in that, Also includes: In response to modifications to the stage identifier and / or the diagnostic information, the analysis report and the pre-trained analysis model are updated.

7. The method according to claim 1, characterized in that, Determining the feature matrix based on the image data set includes: Based on the predetermined image data in the latest image data set, for each other image data in the image data set, a registration algorithm corresponding to the current modality combination is determined, and the image data is registered to the predetermined image data based on the registration algorithm. The current modality combination is a combination of the modality identifier of the predetermined image data and the modality identifier of the image data. The feature matrix is ​​determined based on the image registration results of the image dataset.

8. An electronic device, characterized in that, The electronic device includes: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the assisted diagnostic method based on historical image data as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the auxiliary diagnostic method based on historical image data as described in any one of claims 1-7.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the auxiliary diagnostic method based on historical image data according to any one of claims 1-7.

Images