Focus positioning method and system based on infrared thermal imaging system

Abstract

The invention belongs to the technical field of focus positioning, and particularly relates to a focus positioning method and system based on an infrared thermal imaging system. According to the method, temperature distribution, edge features, temperature gradient and time domain change information are integrated, and temperature and physiological compensation are combined, so that abnormal temperature information is more real and reliable, the accuracy of focus positioning is improved, the misdiagnosis rate is reduced, individualized data such as the body type of a patient, the blood perfusion rate and the environment temperature are introduced, and the accuracy of focus positioning is improved. The detection result can better adapt to individual differences, the pertinence and reliability of diagnosis are improved, abnormal features can be more accurately extracted by comparing an actual infrared image with a theoretical temperature distribution diagram, meanwhile, the credibility and clinical application value of focus positioning are improved in combination with a clinical pathology database, and the diagnosis accuracy is improved. The generated lesion probability thermodynamic diagram visually displays the core area and the diffusion boundary of the lesion, provides a visual reference basis, and is helpful for finding lesions earlier and formulating a reasonable treatment scheme.

Description

Technical Field

[0001] The present invention belongs to the technical field of lesion localization, and in particular relates to a lesion localization method and system based on an infrared thermal imaging system. Background Art

[0002] Infrared thermal imaging, an imaging method based on infrared radiation signals from an object's surface, is widely used in fields such as medical testing, industrial monitoring, and military reconnaissance. In the medical field, infrared thermal imaging, due to its non-contact, radiation-free, and real-time dynamic monitoring capabilities, has gradually become an effective auxiliary diagnostic tool, demonstrating significant application value in the diagnosis of skin lesions, vascular diseases, muscle injuries, and tumors.

[0003] Lesion localization is a key step in medical diagnosis. Traditional lesion detection and localization mainly rely on imaging methods such as X-rays, ultrasound, magnetic resonance imaging (MRI), and computed tomography (CT). However, these methods have limitations. For example, X-rays and CT may pose a potential threat to patient health when used multiple times. MRI and CT equipment are expensive, the imaging process is complex, and they are limited by space and equipment conditions. Traditional imaging technology is mainly used for static lesion imaging, making it difficult to achieve real-time and dynamic monitoring of lesion changes. Traditional imaging methods have limited adaptability to individual patient differences and cannot accurately reflect temperature changes caused by physiological characteristics or environmental influences.

[0004] Infrared thermal imaging technology measures the temperature distribution of infrared radiation on the human body surface to achieve intuitive observation of the thermal characteristics of the target area. It can detect subtle temperature changes and quickly identify possible abnormal areas, avoiding the radiation risks of traditional imaging methods. It is suitable for testing healthy people and special populations (such as pregnant women and children). It can provide dynamic temperature distribution maps and realize real-time monitoring of lesion changes. It can be carried out in a variety of environments and is suitable for different patients and lesion types.

[0005] Despite this, existing lesion detection methods based on infrared thermal imaging still face numerous challenges in practical application. For example, infrared thermal imaging data is easily affected by ambient temperature, equipment errors, and patient physiological characteristics, which can lead to inaccurate lesion positioning. The lack of efficient abnormal temperature zone detection and feature extraction methods makes it difficult to fully reflect the thermal characteristics of the lesion area. Furthermore, the analysis process is not fully integrated with clinical pathological data, resulting in low diagnostic reliability and difficulty in accurately demarcating the lesion core area and diffusion boundary, which cannot meet the clinical needs for accurate diagnosis and treatment. Summary of the Invention

[0006] The purpose of the present invention is to provide a lesion localization method and system based on an infrared thermal imaging system, which can effectively solve the problems of insufficient accuracy, large individual differences, low detection efficiency, etc. existing in traditional lesion localization methods. It is suitable for the detection and positioning of various lesion types and has broad clinical application prospects.

[0007] The technical solutions adopted by the present invention are as follows:

[0008] A lesion localization method based on an infrared thermal imaging system, comprising:

[0009] Collect infrared thermal radiation data of the target area of ​​the patient, generate a target infrared thermal image temperature distribution map, and obtain target temperature characteristic data of the target infrared thermal image temperature distribution map, wherein the target temperature characteristic data includes temperature distribution characteristic information and temperature change edge characteristic information;

[0010] Acquire temperature characteristic compensation information based on temperature distribution characteristic information and temperature change edge characteristic information;

[0011] Obtain normal tissue infrared thermal imaging data and real rare lesion cases from a multi-center clinical pathology database, generate rare lesion simulation data through a generative adversarial network, and combine it with real rare lesion cases from the multi-center clinical pathology database to construct a dynamically expanded dataset;

[0012] Acquiring ultrasound elastic imaging data and near-infrared spectroscopy data based on the dynamically expanded data set, and acquiring a temperature gradient dynamic compensation value based on the ultrasound elastic imaging data and near-infrared spectroscopy data;

[0013] Acquire the patient's body shape data, blood perfusion rate, and ambient temperature data, and generate physiological compensation information;

[0014] Collect infrared thermal radiation data of the patient's healthy area, and combine temperature characteristic compensation information, temperature gradient dynamic compensation value and physiological compensation information to construct a theoretical infrared thermal image temperature distribution map of the patient's target area;

[0015] Acquire abnormal temperature characteristic information based on the target infrared thermal image temperature distribution map and the theoretical infrared thermal image temperature distribution map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information;

[0016] Generate a lesion probability heat map based on abnormal temperature feature information and combined with a preset clinical pathology database;

[0017] Perform regional segmentation on the lesion probability heat map, determine the lesion core area and diffusion boundary, and output the lesion localization results.

[0018] In a preferred embodiment, the steps of collecting infrared thermal radiation data of a target area of ​​a patient, generating a target infrared thermal image temperature distribution map, and obtaining target temperature characteristic data of the target infrared thermal image temperature distribution map, wherein the target temperature characteristic data includes temperature distribution characteristic information and temperature change edge characteristic information, include:

[0019] Collect infrared thermal radiation data of the patient's target area and generate a target infrared thermal image temperature distribution map;

[0020] Obtaining temperature values ​​of multiple temperature regions in the target infrared thermal imaging temperature distribution map, constructing a temperature distribution matrix based on the temperature values ​​of the multiple temperature regions, and marking the matrix as temperature distribution feature information;

[0021] Construct a plane rectangular coordinate system in the target infrared thermal imaging temperature distribution map;

[0022] A plurality of temperature change edge contour inflection point coordinates corresponding to each temperature region are obtained according to a plane rectangular coordinate system, the plurality of temperature change edge contour inflection point coordinates are summarized, and the summary result is marked as temperature change edge feature information.

[0023] In a preferred embodiment, the step of obtaining temperature characteristic compensation information based on temperature distribution characteristic information and temperature change edge characteristic information includes:

[0024] Obtaining a corresponding temperature distribution matrix according to the temperature distribution characteristic information, and extracting a plurality of temperature distribution vectors corresponding to each temperature region from the temperature distribution matrix;

[0025] Acquire multiple temperature change edge contour inflection point coordinates corresponding to each temperature region according to the temperature change edge feature information;

[0026] A temperature feature compensation value is obtained according to a plurality of temperature distribution vectors corresponding to each temperature region and a plurality of temperature change edge profile inflection point coordinates, and is marked as temperature feature compensation information.

[0027] In a preferred embodiment, the step of obtaining the patient's body shape data, blood perfusion rate, and ambient temperature data and generating physiological compensation information includes:

[0028] Obtain the patient's body shape data, blood perfusion rate and ambient temperature data;

[0029] According to the body shape data, blood perfusion rate and ambient temperature data, a target area tissue matrix, a target area blood perfusion vector and an ambient temperature matrix of the target area of ​​the patient are respectively obtained;

[0030] extracting a target region tissue vector from a target region tissue matrix;

[0031] Extract the ambient temperature vector from the ambient temperature matrix;

[0032] A physiological compensation value is obtained according to the target area tissue vector, the ambient temperature, and the vector target area blood perfusion vector, and is marked as physiological compensation information.

[0033] In a preferred embodiment, the steps of collecting infrared thermal radiation data of a patient's healthy area and constructing a theoretical infrared thermal image temperature distribution map of the patient's target area in combination with temperature characteristic compensation information, temperature gradient dynamic compensation value, and physiological compensation information include:

[0034] Collect infrared thermal radiation data of the patient's healthy area and generate infrared thermal image temperature distribution map of the healthy area;

[0035] Matching the infrared thermal image temperature distribution map of the healthy area with the infrared thermal image temperature distribution map of the corresponding area in the preset clinical pathology database, and marking the matched clinical pathology data and the infrared thermal image temperature distribution map corresponding to the patient's target area as a matched infrared thermal image temperature distribution map;

[0036] Obtaining thermal optimization value based on temperature characteristic compensation information, temperature gradient dynamic compensation value and physiological compensation information;

[0037] Obtaining an optimization table, wherein the optimization table includes multiple thermal optimization intervals and an infrared thermal imaging temperature optimization strategy corresponding to each thermal optimization interval;

[0038] Obtain the corresponding optimization strategy from the optimization table according to the thermal optimization interval corresponding to the thermal optimization value;

[0039] The infrared thermal image temperature distribution map is optimized and matched according to the optimization strategy, and the optimization result is marked as the theoretical infrared thermal image temperature distribution map of the patient's target area.

[0040] In a preferred embodiment, the step of obtaining abnormal temperature characteristic information based on the target infrared thermal image temperature distribution map and the theoretical infrared thermal image temperature distribution map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information, comprises:

[0041] Compare the target infrared thermal image temperature distribution map with the theoretical infrared thermal image temperature distribution map;

[0042] Obtaining the infrared thermal image temperature difference area in the comparison result and generating an abnormal infrared thermal image temperature map;

[0043] Abnormal temperature characteristic information is obtained from the abnormal infrared thermal image temperature map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information.

[0044] In a preferred embodiment, the step of generating a lesion probability heat map based on abnormal temperature characteristic information and in combination with a preset clinical pathology database includes:

[0045] According to the temperature gradient distribution information and the thermal radiation time domain variation information, corresponding multiple temperature gradient vectors and multiple thermal radiation vectors are obtained respectively;

[0046] Obtaining abnormal temperature values ​​according to a plurality of temperature gradient vectors and a plurality of heat radiation vectors;

[0047] Obtain a weight table, where the weight table includes multiple abnormal temperature intervals and an abnormal weight corresponding to each abnormal temperature interval;

[0048] Obtain the corresponding abnormal weight from the weight table according to the abnormal temperature range corresponding to the abnormal temperature value;

[0049] Perform similarity matching between the abnormal temperature feature information and the lesion feature template in the preset clinical pathology database to obtain multiple similarity values, obtain the similarity value that meets the preset conditions, and mark it as the conditional similarity value;

[0050] The lesion probability is obtained based on the abnormal weight and multiple conditional similarity values, and a lesion probability heat map is generated in combination with the preset clinical pathology database.

[0051] In a preferred embodiment, the steps of segmenting the lesion probability heat map, determining the lesion core area and diffusion boundary, and outputting the lesion localization result include:

[0052] Perform regional segmentation on the lesion probability heat map to obtain the lesion core area and diffusion boundary;

[0053] Obtain the temperature of the core area of ​​the lesion and mark it as the core temperature;

[0054] Obtaining a level table, wherein the level table includes multiple core temperature intervals and a temperature level corresponding to each core temperature interval;

[0055] Get the corresponding temperature level from the level table according to the core temperature range corresponding to the core temperature;

[0056] The lesion core area, diffusion boundary and temperature level are summarized, and the lesion localization results are output based on the summary results.

[0057] The present invention also provides a lesion localization system based on an infrared thermal imaging system, which is used in the above-mentioned lesion localization method based on an infrared thermal imaging system, comprising:

[0058] A target temperature module is used to collect infrared thermal radiation data of a target area of ​​a patient, generate a target infrared thermal image temperature distribution map, and obtain target temperature characteristic data of the target infrared thermal image temperature distribution map, wherein the target temperature characteristic data includes temperature distribution characteristic information and temperature change edge characteristic information;

[0059] A temperature compensation module is used to obtain temperature characteristic compensation information based on temperature distribution characteristic information and temperature change edge characteristic information;

[0060] A dynamic expansion module is used to obtain normal tissue infrared thermal imaging data and real rare lesion cases from a multi-center clinical pathology database, generate rare lesion simulation data through a generative adversarial network, and combine it with real rare lesion cases from the multi-center clinical pathology database to construct a dynamically expanded dataset;

[0061] A temperature gradient module is used to obtain ultrasound elastic imaging data and near-infrared spectroscopy data based on a dynamically expanded data set, and to obtain a temperature gradient dynamic compensation value based on the ultrasound elastic imaging data and near-infrared spectroscopy data;

[0062] A physiological compensation module is used to obtain the patient's body shape data, blood perfusion rate and ambient temperature data, and generate physiological compensation information;

[0063] Theoretical temperature module is used to collect infrared thermal radiation data of the patient's healthy area and construct a theoretical infrared thermal image temperature distribution map of the patient's target area by combining temperature feature compensation information, temperature gradient dynamic compensation value and physiological compensation information;

[0064] The abnormal temperature module is used to obtain abnormal temperature characteristic information based on the target infrared thermal image temperature distribution map and the theoretical infrared thermal image temperature distribution map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information;

[0065] The lesion probability module is used to generate a lesion probability heat map based on abnormal temperature characteristic information and a preset clinical pathology database;

[0066] The lesion localization module is used to segment the lesion probability heat map, determine the lesion core area and diffusion boundary, and output the lesion localization results.

[0067] And, a lesion positioning terminal based on an infrared thermal imaging system, comprising:

[0068] one or more processors;

[0069] a storage device having one or more programs stored thereon;

[0070] When one or more programs are executed by one or more processors, the one or more processors implement a lesion localization method based on an infrared thermal imaging system.

[0071] The technical effects achieved by the present invention are:

[0072] The present invention makes abnormal temperature information more real and reliable by integrating temperature distribution, edge features, temperature gradient and time domain change information, and combining temperature and physiological compensation, thereby improving the accuracy of lesion positioning and reducing the misdiagnosis rate. It introduces individualized data such as patient body shape, blood perfusion rate and ambient temperature, so that the detection results can better adapt to individual differences, improve the pertinence and reliability of diagnosis, and use the comparison between actual infrared images and theoretical temperature distribution maps to more accurately extract abnormal features. At the same time, combined with the clinical pathology database, it further improves the credibility and clinical application value of lesion positioning. The generated lesion probability heat map intuitively displays the core area and diffusion boundary of the lesion, providing doctors with an intuitive reference basis, which helps to detect lesions earlier and formulate reasonable treatment plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 is a flow chart of the method provided by the present invention;

[0074] Figure 2 This is a system module diagram provided by the present invention. DETAILED DESCRIPTION

[0075] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0076] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0077] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it constitute a separate or selective embodiment that is mutually exclusive of other embodiments.

[0078] Secondly, the present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention in detail, for the sake of convenience, the schematic diagrams are only examples and should not limit the scope of protection of the present invention.

[0079] Please see the attached Figure 1 As shown, a lesion localization method based on an infrared thermal imaging system is provided, comprising:

[0080] S1. Collect infrared thermal radiation data of a target area of ​​a patient, generate a target infrared thermal image temperature distribution map, and obtain target temperature characteristic data of the target infrared thermal image temperature distribution map, wherein the target temperature characteristic data includes temperature distribution characteristic information and temperature change edge characteristic information;

[0081] S2. Obtaining temperature feature compensation information based on the temperature distribution feature information and the temperature change edge feature information;

[0082] S3. Obtain normal tissue infrared thermal imaging data and real rare lesion cases from a multi-center clinical pathology database, generate rare lesion simulation data through a generative adversarial network, and combine it with the real rare lesion cases in the multi-center clinical pathology database to construct a dynamically expanded dataset;

[0083] S4. Acquire ultrasound elastic imaging data and near-infrared spectroscopy data based on the dynamically expanded data set, and acquire a temperature gradient dynamic compensation value based on the ultrasound elastic imaging data and near-infrared spectroscopy data;

[0084] S5. Acquire the patient's body shape data, blood perfusion rate, and ambient temperature data, and generate physiological compensation information;

[0085] S6. Collect infrared thermal radiation data of the patient's healthy area, and construct a theoretical infrared thermal image temperature distribution map of the patient's target area by combining temperature characteristic compensation information, temperature gradient dynamic compensation value, and physiological compensation information;

[0086] S7. Acquire abnormal temperature characteristic information based on the target infrared thermal image temperature distribution map and the theoretical infrared thermal image temperature distribution map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information;

[0087] S8. Generate a lesion probability heat map based on the abnormal temperature characteristic information and in combination with a preset clinical pathology database;

[0088] S9. Perform regional segmentation on the lesion probability heat map, determine the lesion core area and diffusion boundary, and output the lesion positioning result.

[0089] As in the above steps S1 to S9, infrared radiation data of the target area of ​​the patient is collected by the infrared thermal imaging system to generate a target temperature distribution map, and temperature distribution characteristics and temperature change edge characteristics are extracted. The temperature characteristic compensation information is calculated using the extracted temperature distribution and edge information. At the same time, the patient's body shape data, blood perfusion rate and ambient temperature data are collected to construct physiological compensation information. On the basis of collecting infrared thermal radiation data of the healthy area of ​​the patient, a theoretical infrared thermal image temperature distribution map is generated in combination with temperature characteristic compensation and physiological compensation information. The target temperature distribution map actually collected is compared with the theoretical temperature distribution map to extract abnormal temperature characteristic information, such as temperature gradient distribution and time domain change of thermal radiation. The abnormal temperature characteristic information is used in combination with a preset clinical pathology database to generate a lesion probability heat map. The lesion is determined by performing regional segmentation on the lesion probability heat map. The core area and its diffusion boundary are used to output the final lesion localization result, providing an intuitive basis for clinical diagnosis. By integrating temperature distribution, edge characteristics, temperature gradient and time domain change information, combined with temperature and physiological compensation, abnormal temperature information is made more real and reliable, thereby improving the accuracy of lesion localization and reducing the misdiagnosis rate. The introduction of individualized data such as patient body shape, blood perfusion rate and ambient temperature enables the test results to better adapt to individual differences, improve the pertinence and reliability of diagnosis, and use the comparison between actual infrared images and theoretical temperature distribution maps to more accurately extract abnormal features. At the same time, combined with the clinical pathology database, the credibility and clinical application value of lesion localization are further improved. The generated lesion probability heat map intuitively displays the core area and diffusion boundary of the lesion, providing doctors with an intuitive reference basis, which helps to detect lesions earlier and formulate reasonable treatment plans.

[0090] In a preferred embodiment, the steps of collecting infrared thermal radiation data of a target area of ​​a patient, generating a target infrared thermal image temperature distribution map, and obtaining target temperature characteristic data of the target infrared thermal image temperature distribution map, wherein the target temperature characteristic data includes temperature distribution characteristic information and temperature change edge characteristic information, include:

[0091] S101, collecting infrared thermal radiation data of a target area of ​​a patient and generating a target infrared thermal image temperature distribution map;

[0092] S102, obtaining temperature values ​​of multiple temperature regions in the target infrared thermal imaging temperature distribution map, constructing a temperature distribution matrix based on the temperature values ​​of the multiple temperature regions, and marking it as temperature distribution feature information;

[0093] S103, constructing a plane rectangular coordinate system in the target infrared thermal imaging temperature distribution map;

[0094] S104 , obtaining multiple temperature change edge contour inflection point coordinates corresponding to each temperature region according to a plane rectangular coordinate system, summarizing the multiple temperature change edge contour inflection point coordinates, and marking the summarization result as temperature change edge feature information.

[0095] As in the above steps S101 to S104, infrared thermal radiation data of the target area of ​​the patient is collected by the infrared thermal imaging system, and a temperature distribution map of the target area is generated using this data. In the temperature distribution map, multiple temperature areas are identified, and the temperature values ​​of these areas are extracted. A temperature distribution matrix is ​​constructed using these temperature values, and each element in the matrix corresponds to the temperature value of a temperature area, thereby forming temperature distribution feature information. A plane rectangular coordinate system is constructed on the temperature distribution map. Based on the constructed coordinate system, the edges of each temperature area in the temperature distribution map are processed, and multiple inflection point coordinates on the temperature change edge contour are extracted. These inflection point coordinates are summarized to form temperature change edge feature information, which can comprehensively capture the temperature distribution characteristics and subtle differences in temperature changes in the target area, and help to more accurately identify abnormal temperature areas.

[0096] In a preferred embodiment, the step of obtaining temperature characteristic compensation information based on temperature distribution characteristic information and temperature change edge characteristic information includes:

[0097] S201, obtaining a corresponding temperature distribution matrix according to temperature distribution characteristic information, and extracting a plurality of temperature distribution vectors corresponding to each temperature region from the temperature distribution matrix;

[0098] S202, obtaining the coordinates of a plurality of temperature change edge contour inflection points corresponding to each temperature region according to the temperature change edge feature information;

[0099] S203 : Acquire a temperature feature compensation value according to a plurality of temperature distribution vectors corresponding to each temperature region and a plurality of temperature change edge contour inflection point coordinates, and mark the value as temperature feature compensation information.

[0100] As in steps S201 to S203 above, based on the previously acquired temperature distribution feature information, the corresponding temperature distribution matrix is ​​extracted, and multiple temperature distribution vectors corresponding to each temperature region are extracted from the matrix. At the same time, based on the temperature change edge feature information, the edges of each temperature region are processed, and the coordinates of the inflection points on the multiple temperature change edge contours are extracted. The temperature distribution vectors extracted in each temperature region are combined with the corresponding edge inflection point coordinates to calculate the temperature feature compensation value. The calculation formula of the temperature feature compensation value is: Where B represents the temperature characteristic compensation value, a represents the number of multiple temperature zones, a=1,2,3…b, i represents the number of multiple temperature distribution vectors corresponding to each temperature zone, i=1,2,3…n, W iIt represents the number of the ith temperature distribution vector corresponding to each temperature region, h represents the number of the inflection points of the temperature change edge profile corresponding to each temperature region, h=1,2,3…t, X h It is represented by the x-axis coordinate point of the h-th temperature change edge contour inflection point corresponding to each temperature zone, X h+1 It is represented by the x-axis coordinate point of the h+1th temperature change edge contour inflection point corresponding to each temperature zone, h It is represented by the y-axis coordinate point of the h-th temperature change edge contour inflection point corresponding to each temperature zone, Y h+1 It is represented by the y-axis coordinate point of the h+1th temperature change edge contour inflection point corresponding to each temperature zone. When h is t, t+1 is represented as 1. By separating and quantifying the temperature distribution vector and the edge inflection point coordinates, it is possible to accurately capture the subtle changes in temperature data in the target area, thereby effectively eliminating deviations caused by noise or external interference and enhancing the authenticity and accuracy of temperature data.

[0101] In a preferred embodiment, the steps of obtaining normal tissue infrared thermal imaging data and real rare lesion cases from a multi-center clinical pathology database, generating rare lesion simulation data using a generative adversarial network, and combining the real rare lesion cases from the multi-center clinical pathology database to construct a dynamically expanded dataset include:

[0102] S301. Parallel acquisition of normal tissue infrared thermal image feature matrices and pathologically verified rare lesion case datasets provided by multiple medical centers through a multi-center clinical pathology database interface. The rare lesion case dataset includes the spatiotemporal distribution characteristics of infrared thermal radiation in the lesion area and the corresponding histopathological labels.

[0103] S302. Construct a deep generative adversarial network framework, in which the generator network receives the normal tissue thermal image feature matrix as a reference input, and generates simulated lesion data with lesion location, morphology, and thermal radiation gradient changes by embedding lesion feature control parameters in the latent space vector. The discriminator network adopts a multi-scale feature fusion structure to simultaneously discriminate the spatial distribution rationality of the thermal image data and the temporal evolution law of thermal radiation;

[0104] S303, feature-level fusion of real rare disease case data and simulated disease data output by the generative adversarial network according to a preset ratio to form an initial enhanced data set,

[0105] S304 , establishing a cross-modal association index for each enhanced data sample in the initial enhanced data set, including the corresponding ultrasound elastography feature vector, near-infrared spectral absorption spectrum, and patient physiological parameter set, to form a dynamically expanded data set.

[0106] As in the above steps S301 to S304, two types of key data are obtained in parallel from multiple medical centers using a standardized interface: the infrared thermal imaging feature matrix of normal tissue and the real rare lesion case data verified by pathology, wherein the real rare lesion case data include the distribution characteristics of infrared thermal radiation in space (position, range) and time (evolution, diffusion), and the corresponding tissue pathology diagnosis labels (such as benign and malignant, tissue classification). A deep generative adversarial network (GAN) is designed, and its generator uses normal thermal imaging features as input, implants lesion feature control parameters (such as lesion location, morphological complexity, thermal intensity gradient, etc.) in the latent space, and generates simulated lesion thermal images with pathological characteristics. Its discriminator adopts a multi-scale feature fusion structure to simultaneously judge and optimize the spatial rationality (whether it conforms to the real lesion shape distribution in the thermal image) and temporal evolution rationality (whether it conforms to the real heat diffusion evolution law) of the generated image, and fuses the real lesion data with the simulated lesion data at the feature level. The fusion method is It includes but is not limited to feature vector splicing, post-modal conversion mixing (superposition of thermal map enhancement and real lesion contours) and multi-sample synthesis, controlling the proportion to retain the authenticity of real cases, and injecting sufficiently diverse simulated samples to obtain an initial enhanced data set, and establishing a cross-modal index mapping for each enhanced data sample, including the corresponding patient's ultrasound elastic imaging vector (reflecting tissue elasticity), near-infrared spectral absorption spectrum (reflecting physiological metabolism and tissue characteristics) and other patient physiological parameters (such as body temperature, age, gender, menstrual cycle, etc.), to form a dynamically expanded data set, which not only has thermal imaging data, but also integrates multimodal medical information, which can be used for subsequent multimodal diagnostic model training and joint decision-making reasoning. By generating adversarial networks to supplement simulated data and using them in conjunction with real rare cases, the total amount of data can be effectively expanded and the problem of insufficient training samples can be alleviated. It is especially suitable for unpopular diseases, early minimal lesions, etc. Multimodal indexing (thermal imaging + ultrasound + spectrum + physiological parameters) achieves more comprehensive pathological characterization and improves the clinical acceptability of the model.

[0107] In a preferred embodiment, the steps of acquiring ultrasound elastography data and near-infrared spectroscopy data based on the dynamically expanded data set, and acquiring a temperature gradient dynamic compensation value based on the ultrasound elastography data and near-infrared spectroscopy data, include:

[0108] S401, acquiring an ultrasound elastic imaging vector and a near infrared spectral vector corresponding to the ultrasound elastic imaging data and near infrared spectral data of a target area of ​​a patient based on a dynamically expanded data set;

[0109] S402, obtaining a dynamic weight based on the dynamically expanded data set;

[0110] S403 , obtaining a temperature gradient dynamic compensation value according to the ultrasound elastic imaging vector, the near-infrared spectrum vector, and the dynamic weight.

[0111] As in the above steps S401 to S403, the enhanced data samples that have been marked in the dynamically expanded data set are used to call the ultrasound elastic imaging system and the near-infrared spectroscopy measurement equipment interface respectively to automatically extract and preprocess the two modal data. The ultrasound elastic imaging vector represents the characteristics of the tissue shear wave velocity, elastic modulus, etc., and the near-infrared spectroscopy vector represents the spectral characteristics such as the absorption peak intensity and scattering coefficient in different bands. According to the metadata such as the source center, lesion type, and imaging quality of each sample in the enhanced data set, a dynamic weight is extracted from the pre-set weight data. The dynamic weight reflects the importance of different modal data to temperature gradient compensation. The weight can be based on empirical rules, statistical analysis, or a small-scale regression model. Dynamic adaptive adjustment calculates the dynamic compensation value of the temperature gradient based on the ultrasound elastography vector, near-infrared spectral vector, and dynamic weights to correct deviations in theoretical thermal images caused by differences in tissue mechanical and optical properties. The calculation formula for the dynamic compensation value of the temperature gradient is T = q·f·η, where T represents the dynamic compensation value of the temperature gradient, q represents the ultrasound elastography vector, f represents the near-infrared spectral vector, and η represents the dynamic weight. Combining ultrasound mechanics with spectral chemical characteristics can more comprehensively characterize the physical properties of tissues and more accurately model the thermal diffusion process. The weights are automatically adjusted according to different patients, different lesion types, and imaging conditions to avoid excessive influence of single modal noise or bias on the compensation results.

[0112] In a preferred embodiment, the step of obtaining the patient's body shape data, blood perfusion rate, and ambient temperature data and generating physiological compensation information includes:

[0113] S501, obtaining the patient's body shape data, blood perfusion rate, and ambient temperature data;

[0114] S502, respectively obtaining a target region tissue matrix, a target region blood perfusion vector, and an ambient temperature matrix of a target region of the patient according to the body shape data, the blood perfusion rate, and the ambient temperature data;

[0115] S503, extracting the target region tissue vector from the target region tissue matrix;

[0116] S504, extracting an ambient temperature vector from the ambient temperature matrix;

[0117] S505 : Obtain a physiological compensation value according to the target area tissue vector, the ambient temperature, and the target area blood perfusion vector, and mark it as physiological compensation information.

[0118] As in the above steps S501 to S505, the patient's body shape data, blood perfusion rate and environmental temperature data are collected. The body shape data reflects the tissue structure and morphological characteristics of the patient's target area. The blood perfusion rate indicates the local blood supply and reflects the physiological state of tissue metabolism and temperature regulation. The environmental temperature data reflects the impact of external temperature on the thermal balance of the human body. According to the collected data, the following are constructed: target area tissue matrix, which expresses the tissue information of the patient's target area in matrix form based on the body shape data, reflecting the tissue density, thickness or other structural characteristics in the area; target area blood perfusion vector, which converts the blood perfusion rate data into vector form to accurately describe the changes in blood supply in the target area; environmental temperature matrix: uses the environmental temperature data to construct a matrix to represent the target area affected by the external environment. The distribution of the influence of critical temperature is determined by extracting the key tissue feature vector from the tissue matrix and the ambient temperature vector from the ambient temperature matrix. The target area tissue vector, the ambient temperature vector and the target area blood perfusion vector are fused and calculated to obtain the physiological compensation value. The calculation formula of the physiological compensation value is S=Z·H·L, where S represents the physiological compensation value, Z represents the target area tissue vector, H represents the ambient temperature vector, and L represents the target area blood perfusion vector. It is used to correct the temperature distribution deviation caused by differences in individual tissue structure, local blood flow and ambient temperature during infrared thermal imaging. Finally, the calculated compensation value is marked as physiological compensation information, which can fully reflect the physiological and environmental characteristics of each patient, thereby realizing personalized compensation and improving the accuracy and reliability of detection.

[0119] In a preferred embodiment, the steps of collecting infrared thermal radiation data of a patient's healthy area and constructing a theoretical infrared thermal image temperature distribution map of the patient's target area in combination with temperature characteristic compensation information, temperature gradient dynamic compensation value, and physiological compensation information include:

[0120] S601, collecting infrared thermal radiation data of the patient's healthy area and generating an infrared thermal image temperature distribution map of the healthy area;

[0121] S602, matching the infrared thermal image temperature distribution map of the healthy area with the infrared thermal image temperature distribution map of the corresponding area in the preset clinical pathology database, and marking the matched clinical pathology data and the infrared thermal image temperature distribution map corresponding to the patient's target area as a matched infrared thermal image temperature distribution map;

[0122] S603, obtaining a thermal optimization value based on the temperature characteristic compensation information, the temperature gradient dynamic compensation value, and the physiological compensation information;

[0123] S604: Obtain an optimization table, wherein the optimization table includes multiple thermal optimization intervals and infrared thermal imaging temperature optimization strategies corresponding to each thermal optimization interval;

[0124] S605. Obtain the corresponding optimization strategy from the optimization table according to the thermal optimization interval corresponding to the thermal optimization value;

[0125] S606: Optimize and match the infrared thermal image temperature distribution map according to the optimization strategy, and mark the optimization result as the theoretical infrared thermal image temperature distribution map of the patient's target area.

[0126] As in the above steps S601 to S606, infrared thermal radiation data of the patient's healthy area is collected, and a temperature distribution map of the healthy area is generated. The collected infrared thermal image temperature distribution map of the healthy area is matched with the temperature distribution map of the corresponding area in the preset clinical pathology database. The clinical pathology data obtained by matching is annotated and associated with the temperature distribution map of the patient's target area to form a matched infrared thermal image temperature distribution map, thereby using a large amount of clinical data in the database as a reference, and calculating a thermal optimization value based on the temperature characteristic compensation information, temperature gradient dynamic compensation value and physiological compensation information obtained previously. The calculation formula of the thermal optimization value is Y=B*S*T, where Y represents the thermal optimization value, B represents the temperature characteristic compensation value, S represents the physiological compensation value, and T represents the temperature gradient dynamic compensation value. An optimization table is preset. The table contains multiple thermal optimization intervals and the infrared thermal imaging temperature optimization strategy corresponding to each interval. The optimization table gives different optimization methods and processing strategies according to the thermal optimization values ​​in different ranges to ensure that the most appropriate corrective measures can be taken for different situations. According to the calculated thermal optimization value, the thermal optimization interval in which it is located is determined, and the infrared thermal imaging temperature optimization strategy corresponding to the interval is extracted from the optimization table. The obtained optimization strategy is used to optimize the matching infrared thermal imaging temperature distribution map accordingly. After compensation and optimization processing, a theoretical infrared thermal imaging temperature distribution map marked as the patient's target area is generated. The theoretical map aims to eliminate interference factors in the collected data and provide a temperature distribution benchmark map that is closer to the real physiological state. It can ensure that the constructed theoretical temperature distribution map has a high reference value and effectively compensates for individual differences and errors in data collection.

[0127] In a preferred embodiment, the step of obtaining abnormal temperature characteristic information based on the target infrared thermal image temperature distribution map and the theoretical infrared thermal image temperature distribution map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information, includes:

[0128] S701, comparing the target infrared thermal image temperature distribution map with the theoretical infrared thermal image temperature distribution map;

[0129] S702, obtaining the infrared thermal image temperature difference area in the comparison result, and generating an abnormal infrared thermal image temperature map;

[0130] S703 : Acquire abnormal temperature characteristic information from the abnormal infrared thermal image temperature map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain variation information.

[0131] As in steps S701 to S703 above, the actually collected infrared thermal image temperature distribution map of the target area is compared with the theoretical temperature distribution map. Based on the comparison results, the area with obvious difference in infrared thermal image temperature is automatically located, and an abnormal infrared thermal image temperature map is generated accordingly. The key abnormal temperature zone feature information is extracted from the generated abnormal infrared thermal image temperature map, mainly including: temperature gradient distribution information, which reflects the temperature change amplitude and distribution law in the abnormal area, which is often related to the degree and boundary of the lesion; thermal radiation time domain change information, which describes the temperature change trend of the abnormal area at different time points, can reveal the dynamic characteristics of lesion activity, can quickly discover abnormal areas in temperature distribution, provide intuitive basis for lesion positioning, reduce the risk of misdiagnosis and missed diagnosis, and extract temperature gradient (static feature) and thermal radiation time domain change (dynamic feature) at the same time, so that the description of abnormal temperature zone is more comprehensive.

[0132] In a preferred embodiment, the step of generating a lesion probability heat map based on abnormal temperature characteristic information and in combination with a preset clinical pathology database includes:

[0133] S801, obtaining corresponding multiple temperature gradient vectors and multiple heat radiation vectors according to temperature gradient distribution information and heat radiation time domain variation information;

[0134] S802, obtaining abnormal temperature values ​​according to multiple temperature gradient vectors and multiple thermal radiation vectors;

[0135] S803: Obtain a weight table, where the weight table includes multiple abnormal temperature intervals and an abnormal weight corresponding to each abnormal temperature interval;

[0136] S804. Obtain the corresponding abnormal weight from the weight table according to the abnormal temperature range corresponding to the abnormal temperature value;

[0137] S805: performing similarity matching between the abnormal temperature feature information and the lesion feature template in the preset clinical pathology database to obtain multiple similarity values, obtaining a similarity value that meets a preset condition, and marking it as a conditional similarity value;

[0138] S806: Obtain the lesion probability according to the abnormality weight and the multiple conditional similarity values, and generate a lesion probability heat map in combination with a preset clinical pathology database.

[0139] As in steps S801 to S806 above, the temperature gradient distribution information and the thermal radiation time domain variation information are extracted from the abnormal temperature feature information, and converted into multiple temperature gradient vectors and multiple thermal radiation vectors. Based on the extracted temperature gradient vectors and thermal radiation vectors, a specific abnormal temperature value is obtained through comprehensive calculation. The calculation formula of the abnormal temperature value is: Where F represents the abnormal temperature value, g represents the number of multiple temperature gradient vectors, g = 1, 2, 3…r, T g It is represented as the g-th temperature gradient vector, u is represented as the number of multiple thermal radiation vectors, u=1,2,3…d, Q u It is represented as the u-th thermal radiation vector. According to the preset weight table (which defines multiple abnormal temperature intervals and corresponding abnormal weights), the abnormal temperature values ​​are classified into intervals, and the abnormal weight corresponding to each abnormal temperature value is obtained. The extracted abnormal temperature feature information is matched with the lesion feature template in the preset clinical pathology database for similarity, and the similarity value of each template is calculated. For example, the tongue cosine similarity calculation formula is used for calculation. It will not be overly detailed here. The similarity value corresponding to the template that meets the preset conditions (such as the similarity threshold) is screened out and marked as the conditional similarity value. The lesion probability of each abnormal area is calculated according to the abnormal temperature weight and the conditional similarity value. The higher the probability value, the greater the possibility of the lesion in the area. The calculation formula of the lesion probability is: Where k represents the probability of lesions, α represents the abnormality weight, and j represents the number of multiple conditional similarity values, j = 1, 2, 3…q, D j It is expressed as the j-th conditional similarity value, generating an intuitive lesion probability heat map, marking high-risk lesion areas, and using the partitioning strategy of the weight table to make the diagnostic priority of different abnormal temperature intervals clear, which helps to more accurately assess the severity of the lesion area. The lesion probability heat map presents the possibility distribution of lesions through color gradients, which helps doctors quickly identify high-risk areas and greatly improve diagnostic efficiency.

[0140] In a preferred embodiment, the steps of performing regional segmentation on the lesion probability heat map, determining the lesion core area and diffusion boundary, and outputting the lesion localization result include:

[0141] S901, performing regional segmentation on the lesion probability heat map to obtain the lesion core area and diffusion boundary;

[0142] S902, obtaining the temperature of the core area of ​​the lesion and marking it as the core temperature;

[0143] S903: Obtain a level table, where the level table includes multiple core temperature intervals and a temperature level corresponding to each core temperature interval;

[0144] S904. Obtain the corresponding temperature level from the level table according to the core temperature range corresponding to the core temperature;

[0145] S905: Summarize the lesion core area, diffusion boundary, and temperature level, and output the lesion positioning result based on the summary results.

[0146] As in the above steps S901 to S905, the lesion probability heat map is processed, and the heat map is divided into multiple regions using a region segmentation algorithm (such as threshold segmentation, region growing or edge detection), and the lesion core area and diffusion boundary are extracted. The core area represents the area with the most serious lesion, and the diffusion boundary reflects the influence range of the lesion, which helps to determine the morphology and boundary of the lesion. According to the segmentation result, the highest temperature value is extracted in the lesion core area, and the temperature value is marked as the core temperature. A level table is preset, which contains multiple core temperature intervals and the temperature level corresponding to each interval (such as mild, moderate, severe). According to the extracted Core temperature, determine the temperature range to which it belongs, and obtain the corresponding temperature level from the level table, integrate the regional range of the lesion core area, the morphological characteristics of the diffusion boundary, and the temperature level corresponding to the core temperature to form a detailed lesion localization report. The report contains information such as the core position of the lesion, the diffusion range, and the intensity of the lesion. The final output of the lesion localization result can accurately divide the lesion core area and the diffusion boundary, providing high-precision support for the specific positioning and range determination of the lesion, and can quantify the temperature characteristics of the lesion core area. It can intuitively express the severity of the lesion in the form of temperature levels, providing doctors with a more intuitive reference.

[0147] Please see the attached Figure 2 As shown, the present invention also provides a lesion localization system based on an infrared thermal imaging system, which is used in the above-mentioned lesion localization method based on an infrared thermal imaging system, comprising:

[0148] A target temperature module is used to collect infrared thermal radiation data of a target area of ​​a patient, generate a target infrared thermal image temperature distribution map, and obtain target temperature characteristic data of the target infrared thermal image temperature distribution map, wherein the target temperature characteristic data includes temperature distribution characteristic information and temperature change edge characteristic information;

[0149] A temperature compensation module is used to obtain temperature characteristic compensation information based on temperature distribution characteristic information and temperature change edge characteristic information;

[0150] A dynamic expansion module is used to obtain normal tissue infrared thermal imaging data and real rare lesion cases from a multi-center clinical pathology database, generate rare lesion simulation data through a generative adversarial network, and combine it with real rare lesion cases from the multi-center clinical pathology database to construct a dynamically expanded dataset;

[0151] A temperature gradient module is used to obtain ultrasound elastic imaging data and near-infrared spectroscopy data based on a dynamically expanded data set, and to obtain a temperature gradient dynamic compensation value based on the ultrasound elastic imaging data and near-infrared spectroscopy data;

[0152] A physiological compensation module is used to obtain the patient's body shape data, blood perfusion rate and ambient temperature data, and generate physiological compensation information;

[0153] Theoretical temperature module is used to collect infrared thermal radiation data of the patient's healthy area and construct a theoretical infrared thermal image temperature distribution map of the patient's target area by combining temperature feature compensation information, temperature gradient dynamic compensation value and physiological compensation information;

[0154] The abnormal temperature module is used to obtain abnormal temperature characteristic information based on the target infrared thermal image temperature distribution map and the theoretical infrared thermal image temperature distribution map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information;

[0155] The lesion probability module is used to generate a lesion probability heat map based on abnormal temperature characteristic information and a preset clinical pathology database;

[0156] The lesion localization module is used to segment the lesion probability heat map, determine the lesion core area and diffusion boundary, and output the lesion localization results.

[0157] The target temperature module collects infrared thermal radiation data of the target area of ​​the patient, generates a target infrared thermal image temperature distribution map, and extracts target temperature characteristic data. The target temperature characteristic data includes temperature distribution characteristic information and temperature change edge characteristic information: used to identify the boundary of temperature mutation and reflect the possible lesion distribution area. The temperature compensation module generates temperature characteristic compensation information for error correction based on the target temperature characteristic data. The physiological compensation module generates physiological compensation information by obtaining the patient's body shape data, blood perfusion rate and ambient temperature data to reflect the influence of the patient's physiological factors on the infrared thermal image data. The theoretical temperature module collects infrared thermal radiation data of the patient's healthy area, combines the temperature characteristic compensation information and the physiological compensation information to construct a theoretical infrared thermal image temperature distribution map of the patient's target area. The abnormal temperature module compares the target infrared thermal image temperature distribution map. The layout and theoretical infrared thermal imaging temperature distribution map obtain the characteristic information of abnormal temperature areas, including temperature gradient distribution information and thermal radiation time domain change information. The lesion probability module combines the abnormal temperature characteristic information with the preset clinical pathology database to generate a lesion probability heat map. The lesion localization module performs regional segmentation on the lesion probability heat map, clarifies the lesion core area and diffusion boundary, and finally outputs the lesion localization result. Through multi-layer optimization of target temperature characteristics, physiological compensation and temperature compensation, an accurate theoretical temperature distribution map is constructed to ensure that the identification of abnormal areas is more reliable. Combined with the patient's individual body shape data, blood perfusion rate and ambient temperature, personalized compensation is achieved to adapt to the physiological characteristics of different patients and improve the accuracy of diagnostic results. The generated lesion probability heat map and regional segmentation results can display the lesion distribution and severity in an intuitive way, which is convenient for clinicians to quickly evaluate.

[0158] And, a lesion positioning terminal based on an infrared thermal imaging system, comprising:

[0159] one or more processors;

[0160] a storage device having one or more programs stored thereon;

[0161] When one or more programs are executed by one or more processors, the one or more processors implement a lesion localization method based on an infrared thermal imaging system.

[0162] The foregoing is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art may make various improvements and modifications without departing from the principles of the present invention, and such improvements and modifications are also within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained herein shall, unless otherwise specified or limited, be implemented in accordance with conventional means in the art.

Claims

1. A lesion localization method based on an infrared thermal imaging system, characterized in that: include: Collect infrared thermal radiation data of the target area of ​​the patient, generate a target infrared thermal image temperature distribution map, and obtain target temperature characteristic data of the target infrared thermal image temperature distribution map, wherein the target temperature characteristic data includes temperature distribution characteristic information and temperature change edge characteristic information; Acquire temperature characteristic compensation information based on temperature distribution characteristic information and temperature change edge characteristic information; Obtain normal tissue infrared thermal imaging data and real rare lesion cases from a multi-center clinical pathology database, generate rare lesion simulation data through a generative adversarial network, and combine it with real rare lesion cases from the multi-center clinical pathology database to construct a dynamically expanded dataset; Acquiring ultrasound elastic imaging data and near-infrared spectroscopy data based on the dynamically expanded data set, and acquiring a temperature gradient dynamic compensation value based on the ultrasound elastic imaging data and near-infrared spectroscopy data; Acquire the patient's body shape data, blood perfusion rate, and ambient temperature data, and generate physiological compensation information; Collect infrared thermal radiation data of the patient's healthy area, and combine temperature characteristic compensation information, temperature gradient dynamic compensation value and physiological compensation information to construct a theoretical infrared thermal image temperature distribution map of the patient's target area; Acquire abnormal temperature characteristic information based on the target infrared thermal image temperature distribution map and the theoretical infrared thermal image temperature distribution map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information; Generate a lesion probability heat map based on abnormal temperature feature information and combined with a preset clinical pathology database; Perform regional segmentation on the lesion probability heat map, determine the lesion core area and diffusion boundary, and output the lesion localization results.

2. The lesion localization method based on the infrared thermal imaging system according to claim 1, characterized in that: The steps of collecting infrared thermal radiation data of a target area of ​​a patient, generating a target infrared thermal image temperature distribution map, and obtaining target temperature characteristic data of the target infrared thermal image temperature distribution map, wherein the target temperature characteristic data includes temperature distribution characteristic information and temperature change edge characteristic information, include: Collect infrared thermal radiation data of the patient's target area and generate a target infrared thermal image temperature distribution map; Obtaining temperature values ​​of multiple temperature regions in the target infrared thermal imaging temperature distribution map, constructing a temperature distribution matrix based on the temperature values ​​of the multiple temperature regions, and marking the matrix as temperature distribution feature information; Construct a plane rectangular coordinate system in the target infrared thermal imaging temperature distribution map; A plurality of temperature change edge contour inflection point coordinates corresponding to each temperature region are obtained according to a plane rectangular coordinate system, the plurality of temperature change edge contour inflection point coordinates are summarized, and the summary result is marked as temperature change edge feature information.

3. The lesion localization method based on the infrared thermal imaging system according to claim 1, characterized in that: The step of obtaining temperature characteristic compensation information according to temperature distribution characteristic information and temperature change edge characteristic information includes: Obtaining a corresponding temperature distribution matrix according to the temperature distribution characteristic information, and extracting a plurality of temperature distribution vectors corresponding to each temperature region from the temperature distribution matrix; Acquire multiple temperature change edge contour inflection point coordinates corresponding to each temperature region according to the temperature change edge feature information; A temperature feature compensation value is obtained according to a plurality of temperature distribution vectors corresponding to each temperature region and a plurality of temperature change edge profile inflection point coordinates, and is marked as temperature feature compensation information.

4. The lesion localization method based on the infrared thermal imaging system according to claim 1, characterized in that: The steps of obtaining the patient's body shape data, blood perfusion rate, and ambient temperature data and generating physiological compensation information include: Obtain the patient's body shape data, blood perfusion rate and ambient temperature data; According to the body shape data, blood perfusion rate and ambient temperature data, a target area tissue matrix, a target area blood perfusion vector and an ambient temperature matrix of the target area of ​​the patient are respectively obtained; extracting a target region tissue vector from a target region tissue matrix; Extract the ambient temperature vector from the ambient temperature matrix; A physiological compensation value is obtained according to the target area tissue vector, the ambient temperature, and the vector target area blood perfusion vector, and is marked as physiological compensation information.

5. The lesion localization method based on the infrared thermal imaging system according to claim 1, characterized in that: The steps of collecting infrared thermal radiation data of a patient's healthy area and constructing a theoretical infrared thermal image temperature distribution map of the patient's target area by combining temperature characteristic compensation information, temperature gradient dynamic compensation value, and physiological compensation information include: Collect infrared thermal radiation data of the patient's healthy area and generate infrared thermal image temperature distribution map of the healthy area; Matching the infrared thermal image temperature distribution map of the healthy area with the infrared thermal image temperature distribution map of the corresponding area in the preset clinical pathology database, and marking the matched clinical pathology data and the infrared thermal image temperature distribution map corresponding to the patient's target area as a matched infrared thermal image temperature distribution map; Obtaining thermal optimization value based on temperature characteristic compensation information, temperature gradient dynamic compensation value and physiological compensation information; Obtaining an optimization table, wherein the optimization table includes multiple thermal optimization intervals and an infrared thermal imaging temperature optimization strategy corresponding to each thermal optimization interval; Obtain the corresponding optimization strategy from the optimization table according to the thermal optimization interval corresponding to the thermal optimization value; The infrared thermal image temperature distribution map is optimized and matched according to the optimization strategy, and the optimization result is marked as the theoretical infrared thermal image temperature distribution map of the patient's target area.

6. The lesion localization method based on the infrared thermal imaging system according to claim 1, characterized in that: The step of obtaining abnormal temperature characteristic information according to the target infrared thermal image temperature distribution map and the theoretical infrared thermal image temperature distribution map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information, comprises: Compare the target infrared thermal image temperature distribution map with the theoretical infrared thermal image temperature distribution map; Obtaining the infrared thermal image temperature difference area in the comparison result and generating an abnormal infrared thermal image temperature map; Abnormal temperature characteristic information is obtained from the abnormal infrared thermal image temperature map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information.

7. The lesion localization method based on the infrared thermal imaging system according to claim 1, characterized in that: The steps of generating a lesion probability heat map based on abnormal temperature characteristic information and in combination with a preset clinical pathology database include: According to the temperature gradient distribution information and the thermal radiation time domain variation information, corresponding multiple temperature gradient vectors and multiple thermal radiation vectors are obtained respectively; Obtaining abnormal temperature values ​​according to a plurality of temperature gradient vectors and a plurality of heat radiation vectors; Obtain a weight table, where the weight table includes multiple abnormal temperature intervals and an abnormal weight corresponding to each abnormal temperature interval; Obtain the corresponding abnormal weight from the weight table according to the abnormal temperature range corresponding to the abnormal temperature value; Perform similarity matching between the abnormal temperature feature information and the lesion feature template in the preset clinical pathology database to obtain multiple similarity values, obtain the similarity value that meets the preset conditions, and mark it as the conditional similarity value; The lesion probability is obtained based on the abnormality weight and multiple conditional similarity values, and a lesion probability heat map is generated in combination with the preset clinical pathology database.

8. The lesion localization method based on the infrared thermal imaging system according to claim 1, characterized in that: The steps of segmenting the lesion probability heat map, determining the lesion core area and diffusion boundary, and outputting the lesion localization results include: Perform regional segmentation on the lesion probability heat map to obtain the lesion core area and diffusion boundary; Obtain the temperature of the core area of ​​the lesion and mark it as the core temperature; Obtaining a level table, wherein the level table includes multiple core temperature intervals and a temperature level corresponding to each core temperature interval; Get the corresponding temperature level from the level table according to the core temperature range corresponding to the core temperature; The lesion core area, diffusion boundary and temperature level are summarized, and the lesion localization results are output based on the summary results.

9. A lesion localization system based on an infrared thermal imaging system, applied to the lesion localization method based on an infrared thermal imaging system according to any one of claims 1 to 8, characterized in that: include: A target temperature module is used to collect infrared thermal radiation data of a target area of ​​a patient, generate a target infrared thermal image temperature distribution map, and obtain target temperature characteristic data of the target infrared thermal image temperature distribution map, wherein the target temperature characteristic data includes temperature distribution characteristic information and temperature change edge characteristic information; A temperature compensation module is used to obtain temperature characteristic compensation information based on temperature distribution characteristic information and temperature change edge characteristic information; A dynamic expansion module is used to obtain normal tissue infrared thermal imaging data and real rare lesion cases from a multi-center clinical pathology database, generate rare lesion simulation data through a generative adversarial network, and combine it with real rare lesion cases from the multi-center clinical pathology database to construct a dynamically expanded dataset; A temperature gradient module is used to obtain ultrasound elastic imaging data and near-infrared spectroscopy data based on a dynamically expanded data set, and to obtain a temperature gradient dynamic compensation value based on the ultrasound elastic imaging data and near-infrared spectroscopy data; A physiological compensation module is used to obtain the patient's body shape data, blood perfusion rate and ambient temperature data, and generate physiological compensation information; Theoretical temperature module is used to collect infrared thermal radiation data of the patient's healthy area and construct a theoretical infrared thermal image temperature distribution map of the patient's target area by combining temperature feature compensation information, temperature gradient dynamic compensation value and physiological compensation information; The abnormal temperature module is used to obtain abnormal temperature characteristic information based on the target infrared thermal image temperature distribution map and the theoretical infrared thermal image temperature distribution map, wherein the abnormal temperature zone characteristic information includes temperature gradient distribution information and thermal radiation time domain change information; The lesion probability module is used to generate a lesion probability heat map based on abnormal temperature characteristic information and a preset clinical pathology database; The lesion localization module is used to segment the lesion probability heat map, determine the lesion core area and diffusion boundary, and output the lesion localization results.

10. A lesion positioning terminal based on an infrared thermal imaging system, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When one or more programs are executed by one or more processors, the one or more processors implement the lesion localization method based on the infrared thermal imaging system as described in any one of claims 1 to 8.

Images